1 of 122 celltrion, inc., exhibit 1006 - microsoft · pdf file·cer principles &...

1 of 122 Celltrion, Inc., Exhibit 1006

EDITED BY

Vincent T. De Vita, Jr., MD Benno C. Schmidt Chair in Clinical Oncology, Attending Physician and Member, Program of Molecular Pharmacology and Therapeutics, Memorial Sloan-Kettering Cancer Center, Professor of Medicine, Cornell University Medical College, New York, New York

Samuel Hellman, MD Dean, Division of the Biological Sciences and The Pritzker School of Medicine, Vice President for the Medical Center, The University of Chicago, Chicago, Illinois

Steven A. Rosenberg, MD; PhD Chief of Surgery, National Cancer Institute, Professor of Surgery, Uniformed Services University of the Health Sciences School of Medicine, Bethesda, Maryland

214 Contributors


·cER Principles & Practice of Oncology

4th Edition

]. B. LIPPINCOTT COMPANY Philadelphia


Project Editor: Dina K. Rubin Indexer: Sandra King Design Coordinator: Doug Smock Production Manager: Caren Erlichman Production Coordinator: Sharon McCarthy Compositor: Tapsco Incorporated Printer/Binder: Courier Book Company/Westford Color Insert Printer: Village Craftsmen/Princeton Polychrome Press

4th Edition

Copyright© 1993, by J.B. Lippincott Company. Copyright© 1989, 1985, 1982 by J.B. Lippincott Company. All rights reserved. No part of this book may be used or reproduced in any manner whatsoever without written permission except for brief quotations embodied in critical articles and reviews. Printed in the United States of America. For information write J. B. Lippincott Company, 227 East Washington Square, Philadelphia, Pennsylvania 19106.

654321

Library of Ccingress Cataloging in Publications Data

Cancer: principles and practice of oncology I( edited by] Vincent T. De Vita, Jr., Samuel Hellman, Steven A. Rosenberg; 214 contributors.-4th ed.

p. cm. Includes bibliographical references. Includes index. ISBN 0-397-51214-7 (one-vol. ed.) ISBN 0-397-51321-6 (two-vol. set) ISBN 0-397-51322-4 (vol. 1) ISBN 0-397-51323-2 (vol. 2) ISSN 0892-0567 . 1. Cancer. 2. Oncology. I. DeVita, Vincent T., Jr. II. Hellman, Samuel.

III. Rosenberg, Steven A.

The authors and publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any change in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new or infrequently employed drug.


CHAPTER 4

Cancer: Principles & Practice of Oncology, Fo11rt/1 Edition, edited by Vincent T. De Vita, Jr., Samuel Hellman, Steven A. Rosenberg. J.B. Lippincott Co., Philadelphia© 1993.

Renato Baserga

Principles of Molecular Cell Biology of Cancer: The Cell Cycle

Tissues a:nd organs consist of populations of cells held together by intercellular substance that is secreted (e.g., collagen). The growth of tissues and organs may occur by an increase in the number of cells, an increase in the size of cells, or both.' In animals, growth in number usually predominates over growth in size, although some growth in size also occurs during normal and abnormal growth. In Homo sapiens, growth in cell number is by far the most important component irt development. An adult grows to an average of 1015 cells from a single fertilized egg. Although there is a threefold to fourfold increase in the size of cells from newborn to adult humans, most growth is due to an increase in cell number. After humans reach maturity, the number of cells remains essentially constant. However, even in adults, cell division continues at a brisk rate. Approximately I 012 cells die each day and must be replaced. Most of the cells that die come from tissues and organs like the gastrointestinal tract, skin, and bone marrow.

In the adult animal, the number of cells that are produced equals the number of cells that die. This simple equation is fundamental to our understanding of normal and abnormal growth. If the number of cells that are produced exceeds the number of cells that die in a given period of time, there is growth. If fewer cells are produced than die, the organ or tissue shrinks, producing atrophy, a negative form of growth that often occurs with disuse or old age.

THE CELL CYCLE

In every population of cells, there are three subpopulations (Fig. 4-1). The first group is cycling cells that continuously proliferate, going from one mitosis to the next one. The second is composed of terminally differentiated cells that irrevocably 60

leave the growth cycle and are destined to die without dividing again. A third subpopulation of nondividing cells are not cycling and do not divide but can reenter the cell cycle if an appropriate stimulus is applied (G0 cells).' Fig. 4-1 also shows that cycling cells go through four different phases that are defined as G,, S phase, G2, and mitosis.

G0 or quiescent cells are normally present in the living animal. In the adult liver, most of the cells are in G0 . However, if two thirds of the liver are surgically removed, the remaining liver cells quickly resume the cell cycle, proliferate, and restore the liver roughly to its original size. There are several other types of G0 cells in the body. One that is of particular interest to medical oncologists is the stem cell of the bone marrow. These stem cells are capable of reproducing themselves and of producing all the different lineages of hemopoietic cells from lymphocytes to erythrocytes to megakaryo- .. cytes. Most of the bone marrow stem cells are in G0 , a fortunate occurrence because these cells are often protected from chemotherapeutic agents used to treat leukemia or metastatic cancer. The bone marrow depletion caused by chemotherapeutic agents stimulates the protected stem cells to reenter the cell cycle and eventually repopulate the bone marrow. The G0 cells in cell populations help to optimize our therapeutic approaches. In tissue cultures, the G0 state is achieved by restricting the availability of certain growth factors.

TUMOR GROWTH

Figure 4-1 reveals that any population of cells can grow in number by any one of three mechanisms: shortening the length of the cell cycle, resulting in more cells being produced

\ l


s

.. ,., ' '

on-dividing cell

FIGURE 4-1. In the cell cycle, continuously dividing cells go from one mitosis (M) to the next, passing through G1, S (DNA synthesis phase), and G2 . Some cells leave the cycle temporarily, entering a G0 state fron1 which they can be rescued by appropriate mitogenic stimuli. Other cells leave the cycle permanently, entering terminal differentiation.

per unit time; decreasing the rate of cell death; and moving G0 cells into the cell cycle, again resulting in more cells produced per unit time. All three mechanisms operate in normal and abnormal growth. In tumors, there is sometimes a shortening of the cell cycle compared with normal cells, or there is an increase in the growth fraction (i.e., fraction of proliferating cells) or a decrease in the rate of cell loss. 1

In most tumors, all three mechanisms are important in determining the aggressiveness of the tumor, which is best characterized by its doubling time.2 Steel summarized data from various sources on the doubling time of human tumors and their metastases.' They range from as little as 1 7 days for Ewing sarcoma to more than 600 days for certain adenocarcinomas of the Colon and rectum. However, the fastest growing tumor is probably Burkitt's lymphoma, for which the clinical doubling time is 38 to 116 hours, with a mean of 66 hours, or less than 3 days. The doubling time is the real measure of aggressiveness of a tumor, providing a better index of aggressiveness than the mitotic index or the degree of ana- · plasia that are so dear to traditional pathologists.

GENE EXPRESSION AND CELL PROLIFERATION

Many genes are growth regulated, which means that there are genes whose steady-state mRNA levels increase when quiescent fibroblasts, lymphocytes, or other types of cells are stimulated to proliferate by growth factors or by any other mechanism. Several protooncogenes, including c-MYC, c-FOS, and c-MYB, are growth regulated, and because protooncogenes are in some way involved in the regulation of cellular proliferation, it has been assumed that growth-regulated genes are also growth regulatory and that they control cell proliferation. This is not true for three reasons. First, the expression of some growth-regulated genes and protooncogenes (e.g., c-

Gene Expression and Cell Proliferation 61

FOS) is also induced in situations in which cell proliferation does not occur or is inhibited. Second, some exquisitely growth-regulated genes do not have any growth regulatory function. The best example is the thymidine kinase gene, which is growth regulated and yet is a dispensable gene that can be eliminated without affecting the growth of cells in culture or of certain animals, like the squirrel.

The third reason is somewhat more complicated. Cellular proliferation and the cell cycle are complex processes that require the correct expression of several genes and the appropriate modifications of many gene products. The function of gene products can be inhibited by the use of appropriate antibodies, or antisense RNA, or by exposing the cells to antisense oligodeoxynucleotides. Targeting these inhibitors to a variety of cellular RNAs or proteins like those of c-RAS, cMYC, c-FOS, CDC2, c-MYB, and PCNA results in inhibition of cellular proliferation. This is not surprising. The CDC2 and PCNA gene products, for instance, are both proteins necessary for DNA replication. If the production of these proteins is inhibited, DNA synthesis is inhibited, and cell proliferation will grind to a halt. Does that make PCNA and CDC2 growth regulatory? The answer is yes or maybe, depending on the definition of growth regulatory. The products of the PCNA and CDC2 genes are necessary for cellular proliferation, and in this context, they are growth regulatory. However, under this broad definition, many other gene products and cellular components are growth regulatory, including ATP and essential amino acids. There is nothing wrong in calling so many genes and molecules growth regulatory, except that by doing so the heuristic content of the word becomes almost trivial.

If an· antibody or an antisense RNA or an antisense oligodeoxynucleotide targeted to an appropriate gene product blocks cell cycle progression, the targeted gene is said to be required for cellular proliferation. The term growth regulatory should be reserved only for those genes that actually regulate the extent of cellular proliferation, such as the ability to induce cellular proliferation in quiescent cells. Unfortunately, if we use this much stricter definition, cellular growth regulatory genes have not yet been identified. There are some viral gene products, such as the SV40 T antigen or the adenovirus ElA protein, that can cause quiescent mammalian cells to enter DNA synthesis and divide without addition of growth factors, but there is no single cellular gene that microinjected or transfected into quiescent cells causes them to divide.

There is substantial evidence that the molecular biology and the biochemistry of chromosomal replication and mitosis are essentially similar in all animal cells, but the expression of some genes important in cell proliferation is specific to cell types, and in the case of common genes required for cellular proliferation, the regulation of expressio.n may differ from one cell type to another. For instance, the expression of c-MYB is necessary for the entry of hemopoietic cells into S phase, but c-MYB is not expressed in most fibroblasts and epithelial cells, in which its function is probably taken over by similar but different genes. Hemopoietic cells use some growth factor receptors that are different from the growth factor receptors used by fibroblasts or epithelial cells. In other cases, the genes are the same but the way in which they are regulated differs. Differences among different cell types should not be underestimated, because they are probably the basis· of some differential responses to chemotherapeutic agents.


62 The Cell Cycle

MOLECULAR BIOLOGY OF THE CELL CYCLE

The three fundamental components of the cell cycle are chromosomal replication (i.e., S phase), doubling of the other cellular components (i.e., doubling in size), and mitosis. Growth in size occurs throughout the cell cycle at a steady pace, but S phase and mitosis occur in discrete periods. 1 The all-or-none characteristic of S phase and mitosis and the fact that cellular proliferation is triggered by subtle changes in the environment (e.g., small variations in the concentration of certain growth factors) define the cell cycle as a perfect example of a nonlinear chemical clock.

THE S PHASE

The replication of chromosomal DNA in mammalian cells is not completely understood, but considerable progress has been made by studying the replication of SV40 DNA, which, being a simple system, is more amenable to analysis. Several investigators have established in vitro systems capable of supporting the complete replication of plasmid DNA containing the SV 40 origin of replication. Except for the requirement for the SV40 large T antigen, all other proteins involved in the replication reaction are of cellular origin. It is reasonable,

·therefore, to assume that the mechanism of SV40 DNA replication is fund:imentally similar to that of cellular DNA replication.5

Several cellular proteins are necessary for SV40 DNA replication in vitro: DNA polymerase a-primase complex, topoisomerases I and II, replication protein A (RP-A), protein phosphatase 2A, PCNA, replication factor C (RF-C), and DNA polymerase 6. The model proposed for SV40 DNA replication by Kelly and coworkers is illustrated in Figure 4-2.5 RP-A (also known as replication factor-A or RF-A) is amultisubunit cellular protein, which is required in step 2 and also functions as an auxiliary protein for DNA polymerases a and 6. It is phosphorylated at the G,-S boundary and dephosphorylated at mitosis, providing a link between DNA replication and cell cycle-regulated protein phosphorylation. 6

Cellular extracts from G1 cells cannot support DNA replication in vitro, but extracts from S phase cells can. Roberts and coworkers have shown that the activation of DNA replication in S-phase extracts reflects the appearance at the G1-

S boundary of an essential cellular replication factor, called RF-S. 7 The human RF-S contains the homolog- of Schizosaccharomyces pombe p34cdc2 kinase, which is also required for mitosis and whose mechanism of action is described in that section. D'Urso and colleagues ". . . suggest that accumulation of a cyclin during G, leads to the activation of the p34 kinase at the start of S phase. DNA synthesis would begin when phosphorylation of specific replication proteins by the p34cdcz kinase promoted the formation of functional initiation complexes at chromosomal replication origins. "7

The DNA building blocks, the deoxynucleotides, are also necessary for DNA replication, and the cell niust synthesize them. All enzymes necessary for the synthesis of deoxynucleotides (e.g., dihydrofolate reductase, thymidylate synthase, and so forth) are also necessary for entry into S and cellular proliferation. Chromosomal proteins (e.g., histones and nonhistones) must also be synthesized to protect the integrity of

Recognition -,,,,,,,,,,:x,,,,,,,,,,,,

(PP2A) i • . ,.,

/ RP-A I

To po t

Unwinding I I

\ ~ Priming

: :: 8 : :4, :;n-:C: 1a:1::111!8DI::: ::::::

Pola

' Elongation

' RF·C

FIGURE 4-2. In DNA replication, the first step is binding of Tantigen to the· origin of replication, which is facilitated by ATP. In the second step, T antigen catalyzes the unwinding of the duplex in a reaction that requires ATP and replication factor-A (RP-A) and is stimulated by phosphatase 2A. In the third step, short DNA chains are synthesiied by the DNA polymerase a-primase complex, which is the' only activity capable of starting de novo DNA synthesis. Elongation of DNA chains requires the topoisomerases, DNA polymerase 6, RF-C, and PCNA. (Weinberg DH, Collins KL, Simancek Reconstitution of si.rnian virus 40 DNA replication with purified proteins. Proc Natl Acad Sci USA !990;87:6769-6771)

chromosomal structure and funciion. Not surprisingly, the inhibition of the synthesis of these proteins leads to a complete cessation of DNA replication.' The DNA synthesis genes and genes coding for enzymes required for the synthesis of the DNA building blocks are extremely sensitive to low concentrations of cycloheximide (an inhibitor for protein synthesis), indicating that their transcriptional activation depends on prior de nova protein synthesis.

MITOSIS

In eukaryotic cells, mitosis is initiated by the activation of the p34"'" protein kinase, which in mammalian cells is the homolog of the yeast START gene, designated as CDC2 in S. pombe and CDC28 in S. cerwisiae. Given its central role, it is important to know how the p34cdcz is activated and how the p34"''2 kinase induces the biochemical and morphologic changes of mitosis. Two excellent reviews of this subject were written by Moreno and Nurse• and by Lewin.'

M phase is characterized by the activation of a kinase, known under various names (e.g., maturation-promoting fac-

I I

I )


tor, histone HI kinase), whose catalytic subunit is the CDC2 gene product of S. pombe or its homolog in mammalian cells. The catalytic subunit is the CDC2 protein or its equivalent, and the regulatory subunits belong to the category of cyclins, which are a heterogenous group of proteins involved also in the G1 to S transition. The activation of p34c:dc:2 for mitosis is a complex process, diagrammed in Figure 4-3.

After the p34"''2 kinase is active, it phosphorylates substrates. The embarrassingly heterogenous list of possible targets includes histone HI, lamins, nucleolin, myosin light chain, pp60'", SW I5 and T antigen, and probably others.9

To these, Moreno and Nurse add RNA polymerase II, cyclin B. EFI/l, and EFl')'. from which they deduce an amino acid consensus sequence for phosphorylation by. p34: S/T-P-X-Z. where X is a polar amino acid and Z is usually a basic amino acid.8 Notice that the phosphorylated residue can be serine or threonine. Moreno and Nurse give an attractive explanation for the presence of so many substrates; they implicated the many structural and functional changes that occur during mitosis.' Thus, phosphorylation of lamins (a component of the nuclear envelope) may alter coiled-coil interactions in Jamin dimers and promote the dissolution of the nuclear membrane that characterizes mitosis. Histone HI phosphorylation may change nucleosome packing leading to chromosome condensation, a feature that can be observed at the microscope. Phosphorylation of pp60'" may account for the cytoskeletal reorganization occurring at mitosis, and phosphorylation of

p34 threonine phosphorylation

p34 tyrosine phosphorylation

p34/cyclin association

p34 tyrosi~e dephosphorouo

Cyclin phosphoryl ion

Cyclin l degradation

t.ITOSIS

FIGURE 4-3. Mitotic activation of p34cdc2. Activation of p34cdc2 involves threonine 167 phosphorylation and tyrosine 15 dephosphorylation of p34 and its association with a phosphorylated cyclin, which is proteolytically degraded, leaving the activated p34 kinase. After mitosis, cyclins are synthesized de nova, and the cyclin-p34 association leads to an inactive complex, in which the p34 subunit is phosphorylated on tyrosine 15, and the cycle repeats itself. (Lewin B. Driving the cell cycle: M phase kinase, its partners, and substrates. Cell !990;61,549--051)

Molecular Biology of the Cell Cycle 63

nucleolar :Proteins may provide the basis of nucleolus disassembly. Phosphorylation of RNA polymerase II and other DNA-binding proteins may be responsible for the inWbition of transcription during mitosis.

Although some of these explanations are still speculative, a picture of mitosis is clearly emerging, with the p34 "'"' kinase solidly in. the center. On one side of it are the changes necessary for its activation, and on the other side, the structural and functional changes that p34"'"' can cause by the phosphorylation of appropriate substrates. For a pathologist who started his scientific career looking at mitotic figures, it is thrilling to know that the complicated morphologic changes of mitosis are written in the language of molecular biology.

At least two protooncogenes, c-MOS and c-RAS, play an important role in the control of mitosis. 10 The mechanism of action of these two protooncogenes in mitosis is not yet known, but they seem to activate the p34-cyclin complex.

CONTROL OF CELLULAR PROLIFERATION: G0 AND G1

Some of the molecules involved in chromosomal replication and mitosis have been identified. They provide targets for inhibiting cellular proliferation. Antisense strategies or antibodies to these targets consistently arrest the cell cycle. However, if we adopt the strict definition of control of cell proliferation, these are not the genes or the gene products that decide whether the cell should divide. In the case of chromosomal replication, many of the genes coding for the proteins of the DNA-synthesizing apparatus are almost synchronously· activated at the G,-S boundary. Something regulates their expression, and it is this something that decides whether t.he cell should enter S phase, because without the propucts of these genes, there is no DNA synthesis. What is the delicate balance of positive and negative stimuli that results in the activation of the G1-S boundary genes? Is there a candidate gene (or geries) that could be considered the activator of all these other genes?

There is a very promising candidate in budding yeast. DNA synthesis genes are periodically expressed in the yeast cell cycle. The upstream promoter sequences of all these genes have in common a palindromic hexamer element, ACGCGT,. which Lowndes and associates" and Gordon and Campbell have shown capable of conferring cell cycle regulation of expression to heterologous genes.12 Lowndes and colleagues also showed that a protein (which they called DSC!) binds to this hexamer, and that this protein is cell cycle regulated." These are the same genes (e.g., DNA polymerase genes, primase genes, PCNA, CDC2) that are induced at the G1-S boundary in mammalian cells. In the immediate 5' flanking sequences of the PCNA and DNA polymerase ex genes of mammals, there are two repeats of a ACGCGG hexamer, whose homology to the yeast hexamer is striking.

The next step is to identify candidate genes for the DSC! protein and for those proteins that may cooperate with or regulate the expression of the mammalian counterpart of DSC!. Looking for candidates among growth-regulated genes is like looking for a needle in a haystack. The number of genes whose mRNA levels are growth regulated is very Iarge.13

-15 Even allowing some overlap, it is clear that many

genes may be required for cell proliferation but do not nee-


64 The Cell Cycle

essarily control it. Some of them (e.g., c-FOS) are induced even in conditions that have nothing to do with cellular proliferation. However, we can make a distinction between early growth-regulated genes and late growth-regulated genes. As a rule, early growth-regulated genes (e.g., c-FOS, c-MYC, c-JUN) are insensitive to cyclohexiniide even at high concentrations, indicating that de nova protein synthesis is not required for their activation. The expression of late growthregulated genes (e.g., thymidine kinase gene, DNA polymerase a gene, PCNA) is extremely sensitive to cycloheximide, even at concentrations that decrease cellular protein synthesis by only 30% to 40%.15

Because of the large number of genes whose expression increases in cells stimulated to proliferate, only a few genes, whose role in the control of cell proliferation is supported by substantial evidence are considered here. C-MYC Is one of these genes. 16 Circumstantial evidence includes the fact that it is the cellular equivalent of a retroviral transforming gene, that it is induced by platelet-derived growth factor (PDGF) and other growth factors (i.e., it is growth regulated), and that antisense oligodeoxynucleotides to c-MYC RNA inhibit cell proliferation. It is often rearranged or amplified in certain tumors, and if overexpressed in fibroblasts, it can abrogate the requirement for PDGF. 17·

18 C-MYC can mimic all the sets of events that are induced by PDGF, one of two growth factors that are necessary and sufficient for the growth of fibroblasts. Another crucial gene is that for IGF-1; the IGF-I requirement can be abrogated by a constitutively expressed c-MYB, which suggested that c-MYC and c-MYB are the two genes required for the activation of the DNA synthesis genes.19 It turns out, however, that c-MYB activates the genes for IGF-I and the IGF-I receptor, providing an example of autocrine stimulation. But this also raises again the question: how does the activation of the IGF-I receptor regulate the expression of the DNA synthesis genes?

THE ROLE OF CYCLINS AND OF THE IGF-1 RECEPTOR

C-MYC and c-MYB certainly play a role in cell cycie progression, but it is obvious that several other oncogene products are involved in or can modify growth regulation. Hunter lists 51 protooncogene products. 20

That oncogene products that are growth factors or growth factor receptors can modify the control of cellular proliferation makes sense. For instance, ERBB, in its truncated form as vERB, encodes a permanently activated epidermal growth factor (EGF) receptor, constantly transmitting a mitogenic signal. With transcription factors, the explanation is not so simple, but it still is reasonable. Transcription factors may activate cell cycle genes or genes whose products can neutralize the products of antiproliferative genes. A good example is c-MYB. It can activate transcription from the IGF-I and the IGF-I receptor genes, leading to autocrine stimulation and back to growth factors and their receptors. 19·

21 Other oncogenes are involved in signal transduction, which is only a step removed from activated receptors. C-MOS is involved in the phosphorylation of cyclin B, which is required for mitosis and meiosis, and ppeoc·src may also be involved in mitosis.

It is still not clear how some of the protoonccigenes alter growth regulation. That they are required is acceptable, but

why would they stimulate cell division when overexpressed, amplified, or mutated? And why are there so many of them? The problem of too many oncogenes is the same as that of too many antioncogenes. If there is a single controlling event, how can so many different gene products bypass it? If multiple events are required, why are c-MYB and c-MYC enough to stiniulate cell proliferation?

There are cyclins involved in the G2 to M transition. In humans, two mitosis-associated cyclins have been cloned and sequenced: cyclin A and cyclin B. Cyclin A is identical to p60, a protein that associates with p34 cdc2 in interphase cells and with adenovirus El A in transformed cells. 22 G1 cyclins have liniited but significant sequence similarities to mitotic cyclins. G1 cyclins have been well characterized in yeast. Wittenberg and Reed said ". . . the mdst notable property of G, cyclins is that their accumulation or activation appears to be rate limiting for the execution of START."23 START is the CDC2-CDC28 regulated step. These proteins are encoded by three genes, CLNJ, CLN2, and CLN3, which are redundant. Any two of the genes can be deleted without affecting cell growth, but the function of the remaining gene is essential. A human cDNA related to CLN-type cyclins was isolated by Xiong and colleagues and called cyclin Dl.24 It is overexpressed in glioblastoma cell lines and is identical to PRAD 1, a gene overexpressed in parathyroid tumors and possibly related to the BCLJ oncogene.25 Two mouse cyclin-like genes, regulated by CSF-1 during the G, phase of the macrophage cell cycle have also been identifieq and cloned." The role of cyclins in the control of the G1-S transition is a key question in future investigation, and the first problem to be solved is finding at what point of the cell cycle they act.

Pardee's restriction point can be identified with the activation of the IGF-1 receptor." The IGF-1 receptor and its ligand play a crucial role in determining the extent of cellular proliferation in a variety of cell types.28 After the number of IGF-1 binding sites on the cellular membrane reaches a certain level, the cell can respond to IGF-1 and enter S phase. The number of IGF-1 receptors is regulated by many agents, including growth factors like PDGF and EGF, protooncogenes like c-MYC and c-MYB, and DNA oncogenes like the SV40 T antigen. Some of these agents also increase the expression of the ligand, IGF-1, creating an autocrine mechanism that drives the cell into S phase.

The activation of the IGF-1 receptor undoubtedly constitutes an essential step in cell cycle progression in fibroblasts, hemopoietic and epithelial cells, and lung cancer and breast cancer cells. After a sufficient number of IGF-1 receptors is activated, the cell progresses to S phase without further need of growth factors or unique copy gene expression. 27 In terms of enabling cell proliferation, the closest thing to the SV 40 large Tantigen or the adenovirus ElA is the activated IGF-1 receptor. Other growth factors (e.g., PDGF, EGF, mitogenic stimuli of hemopoietic cells) require a functional IGF-1 receptor to stimulate cell proliferation, but an activated, overexpressed IGF-1 receptor does not need the activation of other receptors, such as the PDGF or EGF receptors.28

•29 The IGF-

1 receptor is upstream of the onset of S phase and precedes and regulates the increase in mRNA levels of the DNA synthesis genes like the DNA polymerase a gene, thymidine kinase gene, PCNA, and CDC2. 30 It is downstream from c-MYC, other early growth-regulated genes, and c-MYB.


t

AB. c. D. E.

Molecular Biology of the Cell Cycle 65

- -FIGURE 4-4. Cell cycle regulation of gene expression. The time of gene expression during the cell cycle (including G0) is represented by the heavy lines. (A) Early growth-regulated genes (i.e., a set of transcripts necessary for the G0 to G1 transition). Prototypes are c-MYC and c-FOS, but many others are known. (B) G1 genes provide a set of transcripts necessary for cell cycle progression during G1 • Prototypes are ornithine decarboxylase, C-MYB and the mostly unknown genes complementing G1-specific temperature-sensitive mutants of the cell cycle. (C) Genes at the restriction point (i.e., genes necessary to pass through the restriction point). Prototypes are IGFl and IGFl receptor and the genes immediately controlled by the activation of the IGFl receptor. It is not clear yet where G1 cyclins fit, whether before or after the IGF! receptor. (D) DNA synthesis genes provide all the genes required for chromosomal replication. Prototypes are DNA polymerases a and 6, CDC2, and PCNA. Antioncogenes, like the products of the p53 and ~b genes, could act at this point or between the activation of the IGFl receptor and the transcription of the DNA synthe.sis genes. (E) Genes required for the S to M transition. Prototypes are p34"d"2 and associated cyclins A and B, and C-MOS.

The uncertain role of G1 cyclins is emphasized in Figure 4-4, a schematic r~presentation of gene expression during cell cycle progression in animal cells. Only a few representative key genes are given. Each of those areas is a target for therapeutic manipulations.

TUMOR SUPPRESSOR GENES

One way to slow cellular proliferation is to reduce the expression of genes whose products stimulate cell division. Another alternative is to increase the expression of genes or to increase the activity of gene products that suppress cellular proliferation. The latter alternative has gained ascendancy with the discovery of antioncogenes or tumor suppressor genes. These have been reviewed by Bishop31 and by Marshall. 32

The concept of tumor suppressor genes stems from a 30-year-old observation that hybrid cells generated by the fusion of tumorigenic and nontumorigenic cells are nontumorigenic, unless certain chromosomes are lost from the hybrids, which generate tumorigenic variants. The identification of tumor suppressor genes has depended on the study of certain human tumors, especially retinoblastoma and Wilms' tumor. Chromosomal deletions in these conditions have allowed isolation of genes that have an antiproliferative effect; the number of such genes have been rapidly expanding, and Table 4-1 gives a list of tumor suppressor genes and other factors that may exert an inhibitory effect on cellular proliferation.

We are concerned here only with how these genes may regulate cell cycle progression. The products of the RB 1 and p53 genes bind to viral oncogene proteins, such as the EIA transforming protein of adenovirus or the SV40 large T antigen.33·34 These oncogenes may act by neutral~zing the inhibitory effect of RBI or p53 proteins. In cell cycle terms, researchers are looking for a cellular equivalent of EIA or T,

and the search is on for cellular proteins that bind to the pl05-RB protein or to the p53 protein. Cyclin A has been shown· to bind to EIA, and it seems that the CDC2 kinase is part of the EIA-cyclin A complex.35 The NFJ antioncogene may exercise negative control over the RAS proteins.

There is ambiguity about how these genes may inhibit cellular proliferation. Bishop31 and Marshall32 point out that the RBI and p53 gern1 products could be directly involved in DNA

TABLE 4-1. Genes With Antiproliferative Action

Inhibitory Gene or Factor

RBl

p53

WTl DCC NFl FAP MENl Prohibitin

GAS genes

C/EBP

Statin

Remarks

Retinoblastoma gene; also deleted in osteosarcomas and other tumors

P53 protein; binds oncogene products; deleted in some tumors

Wilms tumor Carcinoma of the colon Neurofibromatosis type I Carcinoma of the colon. Multiple endocrine neoplasias A cDNA clone capable of inhibiting cell

proliferation Growth-arrested genes, overexpressed in G0

cells CAA T-binding protein; inhibits cell

proliferation Overexpressed in growth-arrested and

senescent cells

(Modified from Bishop JM. Molecular themes in oncogenesis. Cell 1991;64,235-248)


66 The Cell Cycle

synthesis or indirectly as transcription factors. Evidence has been presented for both functions; it is possible that they may have a dual function. After all, the NFl transcription factor, a CCAA T binding protein, is also necessary for adenovirus DNA replication.36 Another transcription factor, the CCAATenhancer binding protein (C/EBP), is also an inhibitor of cell proliferation, indicating that regulation of transcription may play both a positive and negative role in cell division. 37

Growth in cell number is regulated by the length of the cell cycle, the growth fraction, and the rate of cell death. The wild-type p53 induces apoptosis; a mutant p53 loses the property of regulating cell death.38 Another candidate for regulating the rate of cell death is BCL2, which prevents apoptosis.39

There are other attractive candidates as negative regulators of cell proliferation, like the growth-arrest-specific (GAS) genes, whose mRNAs are maximally expressed in G0 cells, and prohibitin, whose mRNA inhibits entry into S phase. 40

•41

It is difficult to determine the extent of the role of antioncogenes in the cell cycle. Marshall, in his review, limited himself to comment on the peculiar fact that there are already multiple suppressor genes. 32 An unbiased perspective must await the discovery and definition of other antioncogenes and their products.

REFERENCES I. Baserga R. The biology of cell reproduction. Cambridge, MA: Harvard University Press,

1985. 2. Bresciani F, Paoluzi R, Benassi M. Cell kinetics and growth of squamous cell carcinomas

in man. Cancer Res. 1974;34:2405-2415. 3. Steel GG. Growth kinetics of cum ors. Oxford: Clarendon Press, 1977. • 4. Baserga R. The cell cycle: My1hs and realities. Cancer Res 1990;50:6769-6771. 5. Weinberg DH, Collins KL,Simancek P. Reconsdtution ofsirn:ian virus 40 DNA replication

with purified proteins. Proc Natl Acad Sci USA 1990;87:8692-8696. 6. Din S, Brlll SJ, Fainnan MSP, et al, Cell cycle-regulated phosphorylation of DNA rep

lication factor A from human and yeast cells. Genes Dev 1990;4:968-977. 7. D'Urso G, Marracclno RL, Marshak DR. et al. Cell cycle control of DNA replication by

a homologue from human cells of the p34..sa protein kinase. Science 1990;250:786-791.

8. Moreno S, Nurse P. Substrates for p34~: In vivo veritas? Ceil 1990;61:549-551. 9. Lewin B. Driving the cell cycle: M phase kinase, its partners, and substrates. Cell 1990;61:

743-752. 10. Daar I. Nebreda AR, Yew N, Sass P, Paules R. Santos E, WJgler M, Vande Woude GF.

The RAS oncoprotein and M·phase activity. Science 1991;253:74-76. 11. Lowndes NF, Johnson AL, Johnston LH. Coordination of expression of DNA synthesis

genes in budding yeast by a cell cycle regulated transcription factor. Nature 1991;350: 247-250.

12. Gordon CB, Campbell JL. A cell cycle-responsive transcriptional control element and a negative control element in the gene encoding DNA polymerase a: In Saccharomyces cerevisiae. Proc Natl Acad Sci USA 1991 ;88:6058-6062.

13. Almendral JM, Sommer D. MacDonald-Bravo H, et al. Complexity of the early genetic response to growth factors in mouse fibroblasts. Mo! Cell Biol 1988;8:2140-2148.

14. Lau LF, Nathans D. Identification of a set of genes expressed during the Ga-G1 transition of cultured mouse cells. EMBOJ 1985;4:3145-3151.

IS. Hofbauer R, Denhardt D. Cell cycle-regulated and proliferation stimulus-responsive genes. Eukaryotlc Gene Expression 1991;1:247-300.

16. Studzinski GP. Oncogenes, growth and the cell cycle: An overview. Cell Tissue Ktnet 1989;22:405-424.

17. Armelin HAM, Armelin MCS, Kelly K. et al. Functional role for c-MYC in miiogenic response to platelet.derived growth factor. Nature 1984;310:655-660.

18. Kaczmarek L, Hyland JK, Watt R, et al. Micro-injected c-MYC as a competence factor. Science 1985;228:1313-1315.

19. Travali S, Reiss K, Ferber A, et al. Constitutively expressed c-MYB abroga_tes the requirement for insulin-like growth factor I in 3T3 fibroblasts. Mo! Cell Biol 1991;2:731-736.

20. HunterT. Cooperation between oncogenes. Cell 1991;64:249-270. 21. Reiss K, Ferber A, Travali S, et al. The protooncogene c-MYB increases the expression

of insulin-like growth factor I and insulin-like growth factor I receptor messenger RNAs by a transcriptional mechanism. Cancer Res 1991;51:5997-6000,

22. Pines], HunterT. Human cyclinA is adenovirus EIA associated protein 60 and behaves dlfferently from cyclin B. Nature 1990;346:760-763.

23. Wittenberg C, Reed SL. Control of gene expression and the yeast cell cycle. Crlt Rev Eukaryotic Gene Expression 1991;1:189-205.

24. Xiong X, Connolly T, Fulcher B, et al. Human D-type cyclin. Cell 1991 ;65:691-()99. 25. Motokura T, Bloom T, Kim HG, et al. A novel cyclln encoded by a BLC-J-linked candidate

oncogene. Nature 1991;350:512-515. 26. Matsushlme H, Roussel MF, Ashmun RA et al. Colony-stimulating factor 1 regulates

novel cyclins during the G1 phase of the cell cycle. Cell 1991;65:701-713. 27. Pardee AB. G1 events and regula!lon of cell proliferation. Science 1989;246:603-608. 28. Baserga R. The IGF-1 receptor as the res!riction point of the cell cycle. Ann NY Acad

Sci (in press). 29. Pletrzkowski Z, Lammers R, Carpenter G, et al. Constitutive expression of IGF-1 and

IGF-1 receptor abrogates all requirements for exogenous growth factors. Cell Growth Diff 1992;3:199-205.

30. Sunnacz E, Nugent P, Pietrzkowksl Z, Baserga R. The role of the IGF-1 receptor in the regulation'ofCDC2 mRNA in fibroblasts. Exp Cell Res 1992;199:275-278.

31. Bishop JM. Molecular themes In oncogenesis. Cell 1991;64:235-248. 32. ~arshall CJ. Tumor suppressor genes. Cell 1991:64:313-326. 33. Whyte P, BuchkoVich K. Horowltz JM, et al. Association between an oncogene and an

antloncogene: The adenovirus EIA proteins bind to the retlnoblastoma gene product. Nature 1988;334:124-129.

34. De Caprio JA, Ludlow JW, Figge J, et aJ. SV40 large tumor antigen fonns a specific complex with the product of the retinoblastoma susceptibility gene. Cell 1988:54:275-283.

35. Pines], HunterT. Human cyclin A is adenovirus EI A-associated protein p60 and behaves differently from cyclln B. Nature 1990;346:760-763.

36. Jones KA, KadonagaJT, Rosenfeld PJ, et aJ. A cellular DNA-binding protein that activates eukaryotic transcription and DNA replication. Cell 1987;48:79-89.

37. Umek RM, Friedman AD, McKnight SL CCAAT-enhancerblnding protein: A component of a differentiation switch. Science 1991 ;251:288-292.

38. Yonish-Rouach E, Resnitzky D, Loten J, et al. Wild-type p53 induces apoptosis ofmyeloid leukaemic cells that is inhibited by lnterleukin-6. Nature 1991 ;352:345-347.

39. Williams GT. Programmed cell death: Apaptosls and oncogenesis. Cell 1991;65:1097-1098.

40. Ciccarelli C, Philipson L, Sorrentino V, Regulation of expression of growth anest·speclflc genes In mouse fibroblasts. Mo! Cell Biol 199~10:1525-1529,

41. Nuell MJ, Stewart DA, Walker L, et al. Prohlliitln, an evolutionarily conserved intracellular protein that blocks DNA synthesis Jn normal fibroblasts and HeLa cells. Mo! Celi BJol 1991;2:1372-1381.

\ 11 of 122 Celltrion, Inc., Exhibit 1006

CHAPTER tfi

Cancer: Pri11ciples & Practice of Oncology, Fourth Edition, edited by Vincent T. DeVita, Jr., Samuel Hellman, Steven A. Rosenberg. J.B. Lippincott Co., Philadelphia© 1993.

Vincent T. DeVita, Jr

Principles of Chemotherapy

HISTORY

Cancer chemotherapy had its roots in the work of Paul Ehrlich, who coined the word chemotherapy. Ehrlich's use of rodent models of infectious diseases to develop antibiotics led George Clowes, at Roswell Park Memorial Institute in Buffalo, New York, in the early 1900s, to develop inbred rodent lines that could carry transplanted tumors.' These types of models served as the testing ground for potential cancer chemotherapeutic agents and only recently have been effectively supplemented by human cells grown in culture. Alky lating agents, the first modern chemotherapeutic agents, were a product of the secret war gas program in both world wars. An explosion in Bari Harbor during World War 112·' and the exposure of seamen to mustard gas led to the observation that alkylating agents caused marrow and lymphoid hypoplasia, which led to their use in humans with hematopoietic neoplasms such as Hodgkin's disease and lymphocytic lymphomas, first attempted at Yale-New Haven Medical Center in 1943. Because of the secret nature of the gas warfare program, this work was not published until 1946. 1

-' The demonstration of dramatic regressions in advanced lymphomas with chemicals caused much excitement and later much disappointment, because the tumors invariably grew back. After Farber's observation on the effects of folic acid on leukemic cell growth in children with lymphoblastic leukemia and the development of the folic acid antagonists as cancer drugs, the chemotherapy of cancer began in earnest. The cure of childhood leukemias and Hodgkin's disease with combination chemotherapy in the 1960s proved that human cancers, even in their advanced stages, could be cured by drugs, and the application to the chemotherapy of solid tumors began.

276

CHEMOTHERAPY AS PART OF .THE INITIAL TREATMENT OF CANCER

There are four ways chemotherapy is generally used': as an induction treatment for advanced disease, as an adjunct to the local methods of treatment, as the primary treatment for patients who present with localized cancer, and by direct installation into sanctuaries or by site-directed perfusion of specific regions of the body most affected by the cancer.

The term induction chemotherapy has been used to describe the drug therapy given as the primary treatment for patients who present with advanced cancer for which no alternative treatment exists. 6 Development of new treatments is based on the effectiveness of the cancer drugs in rodent models. Combinations of drugs are fashioned based on the effectiveness, the level of cross-resistance, and the limiting toxicity of the available drugs when used alone in similar patient populations. Patients who fail after one drug treatment and require further chemotherapy pose a particularly difficult treatment problem because of the volume of tumor, their poor general health, and drug resistance. Induction chemotherapy in these patients is referred to as salvage treatment.

Adjuvant chemotherapy denotes the use of systemic treatment after the primary tumor has been controlled by an al- · ternative method, such as surgery and radiotherapy. The selection of an adjuvant treatment program for a particular patient is based on response rates in separate groups of patients with advanced cancers of the same histologic type. The determination of a population of patients as suitable for adjuvant treatment is based on available data on their average


Clinical Endpoints in Evaluating Response to Chemotherapy 277

risk of recurrence after local treatment alone and on disease variables known to influence prognosis adversely.

Primary chemotherapy denotes the use of chemotherapy as the initial treatment for patients who present with localized cancer for which there is an alternative, but less than completely effective, treatment. This approach also has been called neoadjuvant chemotherapy, but the term primary chemotherapy is more accurate.7

•8 For chemotherapy to be used as the

primary treatment of a partially curable, localized cancer, there must be considerable evidence for the effectiveness of the drug program against advanced disease of the same type.

CLINICAL ENDPOINTS IN EVALUATING RESPONSE TO CHEMOTHERAPY

INDUCTION CHEMOTHERAPY

In induction chemotherapy for advanced cancer, it is possible to determine the response to drugs on a case-by-case basis. The partial response rate is usually defined as the fraction of patients that demonstrates at least a 50% reduction in measurable tumor maSs; such responses usually are not of much value clinically, because they usually are brief and offset by the drug toxicity associated with continuous treatment. However, partial responses are useful in the evaluation of new drugs, or new drug programs, to.·determine whether the particular experimental approach is worth.pursuing further. The most important indicator of the effectiveness of chemotherapy is the complete response rate-it is the prerecjuisite for cure. When new programs consistently produce more than an occasional complete remission, they have invariably later proved of practical value in medical practice. The qualitative and quantitative differences in the clinical value between a complete and a partial response are such that complete responses should always be reported separately. The most important indicator of the quality of a complete remission is the relapsefree survival from the time all treatment is discontinued. This criteria is the only clinical counterpart of the quantifiable cytoreductive effect of drugs in rodent systems. A current trend among many clinical investigators is the use of "freedom from progression" of patients who have attained complete and partial responses, measured as a combined group.8 This method is said to be an indicator of the practical potential of a new treatment, but it obscures the value of a relapse-free survival of complete responders as the major determinant of the quality of remission and the potential for cure. Other endpoints, such as median response duration and median survival, are also of little practical value until treatment results have been refined so that the complete response rate is higher than 50%.

ADJUV ANT CHEMOTHERAPY

There was great excitement concomitant with the use of chemotherapy as an adjunct to local treatments. The promise was great, because tumor volume is at a minimum when adjuvant therapy is initiated, and it was assumed that treatment with drugs at this stage would produce a much higher cure rate or that treatment intensity could be reduced and side effects thereby diminished, without loss of therapeutic effectiveness. Both assumptions have little scientific basis. Failure

to appreciate the problems surrounding the assessment of the response of a group of patients to adjuvant chemotherapy is the source of some of the current disillusionment with the positive, but less than dramatic, results achieved with adjuvant chemotherapy in common tumors, such as breast and colorectal cancers.9

·10

It should be remembered that the major indicator of effectiveness of a chemotherapy program-the complete remission rate-is lost in the adjuvant setting, because the primary tumor has already been removed. In the clinic, treatment is selected for individual patients based on response rates in an entirely different population of patients with advanced disease with the same histologic type. Relapse-free survival remains the major endpoint, but the micrometastases in the population of adjuvant-treated patients consist of a mixture of tumor cells, some of which can be expected to be sensitive to chemotherapy, and others resistant. The relapse-free survival in the adjuvant setting, therefore, measures time to regrowth of cells unresponsive, partially responsive, or very sensitive to chemotherapy and is the equivalent of the duration of reffiission of complete and partial responders and the interval of regrowth in patients who would have been classified as nonresponders. In this sense, it is similar to the use of freedom from progression in patients with advanced disease. Attempts to use in vitro assays of drug sensitivity from the biopsy material of primary tumors to overcome the shortcomings of the absence of an indicator of individual response have not proved practical.

PRIMA.RY CHEMOTHERAPY

The unique feature of using chemotherapy in patients with localized tumor, before or instead of purely local treatments, is the preservation of the presenting tumor mass as a biologic marker of responsiveness to the drugs. As with induction chemotherapy for patients with advanced cancer, it is possible to determine, on a case-by-case basis, the potential effectiveness of a new treatment program. By definition, the presenting tumor mass is also the largest aggregate of tumor in the body and, historically, the oldest, and it is therefore the aggregate mass of tumor cells most likely to contain one or more resistant cell lines. 11 Being the largest mass of cells, it is also the mass with the lEiast favorable cell kinetics. It is reasonable to assume, then, that whatever the effect of chemotherapy the physician sees on the primary tumor, a similar or greater effect is occurring fairly unifor1nly in micrometastatic deposits. A poor response of the primary tumor to chemotherapy indicates a grOup of patients for which alternative methods of treatment should be used quickly. Another feature of primary chemotherapy is the ability to delineate partial responders with varying degrees of prognosis, as determined by the state of the residual tumor mass after an initial good bu~ partial response. Removal of residual tumor and histologic examination of the tissue allow determination of the viability of the remain_ing tumor mass. The response duration of complete and partial responders must be catalogued separately.

The most important issue facing investigators of primary chemotherapy is whether an effective primary chemotherapy treatment, pursued flexibly and intensively to complete remission, plus two or more additional cycles of treatment, will define a significant fraction of patients whose disease is cured


278 Principles of Chemotherapy

by chemotherapy alone, without the addition of alternative treatments. In carefully selected patients with some stages of the commonest tumors for which there is less than satisfactory standard treatment, such studies are ethically and theoretically sound and are being pursued. Such an approach could result in briefer, less morbid, and more effective treatment programs.5

The use of chemotherapy as the primary treatment is reviewed, when appropriate, in each of the disease-oriented chapters. Table 16-1 lists tumors in which primary chemotherapy for localized forms of the cancer in question have already been incorporated in clinic protocols (first and second categories) and in which current clinical trials show considerable progress (third category).12

-17

SPECIAL USES OF CHEMOTHERAPY.

Special uses of chemotherapy include (1) the installation of drugs into the spinal fluid, directly through a lumbar puncture needle or into an implanted Ommaya reservoir, to treat meningeal leukemia and lymphoma, and into the pleural or the pericardia} space to control effusions; (2) splenic infusion to control spleen size; (3) hepatic artery infusion to treat hepatic metastases selectively; (4) carotid artery infusion to treat head and neck cancers and brain tumors; and (5) intraperitoneal installation of drugs using dialysis techniques. These uses are discussed throughout this book in r.elation to specific cancers. In all instances, the rationale for directed chemotherapy is based on achieving a higher concentration over time (C X T) against the target tumor tissue while sparing normal tissue. The usefulness of intracerebrospinal fluid and intrapleural administration of drugs is already established. Hepatic infusion

TABLE 16-1. Primary Chemotherapy

Neoplasms in Which Chemotherapy Is the Primary Therapeutic Modality Localized diffuse large cell lymphoma Burkitt's lymphoma Childhood Hodgkin's disease Wilms' tumor Embryonal rhabdomyosarcoma Small cell lung cancer

Neoplasms in Which Primary Chemotherapy Can Allow Less Mutilating Surgery Anal carcinoma Bladder carcinoma Breast cancer Laryngeal cancer Osteogenic sarcoma Soft tissue sarcomas

Neoplasms in Which Clinical Trials Indicate an Expanding Role for Primary Chemotherapy in the Future Non-small-cell lung cancer Breast cancer Esophageal cancer Nasopharyngeal cancer Other cancers of the head and neck region

of chemotherapy has been simplified and improved enough by the development of technology for the infusion of drugs that a reevaluation of this approach isjustifled (see Chap. 61, section 3). It is now possible to ineasure the·active principles of cancer drugs and their targets within the biologic range, and drugs can be infused in timing with the body's circadian rhythm (see Chap. 69, section 7).

The intraperitoneal administration of drugs to treat ovarian cancer, a disease that kills almost exclusively by local effects in the abdomen, is now being investigated, because it allows a wide distribution of antitumor drugs in the smallest interstices of the abdominal cavity, and because a higher C X T at the tumor is achieved (see Chaps. 18 and 39) ... ·" The concentration of drug available in the peritoneal cavity for some drugs with this "belly bath" technique far exceeds the plasma level achievable with systemic administration. The effects are particularly marked for drugs such as 5-fluorouracil, which is metabolized in the liver and excreted by the kidney, and doxorubicin and cisplatin, which, because of their molecular size, diffuse more slowly across the peritoneal membrane. A similar approach is being explored with abdominal installation of photoaffinity compounds, with subsequent exposure to laser light sources (see Chap. 69, section 7).

Drugs can also be encompassed in lipid bilayer droplets called liposomes. 20-22 The surface characteristics of liposomes can be altered to direct their delivery to specific organ sites or into resistant cell lines. Labile liposomes thiit dissolve at temperatures of 41 °C can deposit drugs selectively in preheated areas. 21 A disadvantage of liposomes, however, is their failure to leave the vascular system except in the sinusoids of the liver and the spleen; liposome encapsulation of drugs for targeted delivery has been of limited value.22

PRINCIPLES GOVERNING THE USE OF COMBINATION CHEMOTHERAPY

With some exceptions (e.g., choriocarcinoma and Burkitt's lymphoma), single drugs do not cure cancer. It became apparent in the 1960s that drug combinations are necessary to produce durable clinical responses. In the early years of chemotherapy, drug combinations were developed based on known biochemical actions of available anticancer drugs rather than on their clinical effectiveness. These programs were largely ineffective.23

-27 The era of effective combination

chemotherapy began when an array of active drugs from different classes became available for use in combination in the treatment· of leukemias and lymphomas. Combination chemotherapy has now been extended to the treatment of most other malignancies, as described throughout this text.

Combination chemotherapy accomplishes three important objectives not possible with single-agent treatment. It·provides maximal cell kill within the range of toxicity tolerated by the host for each drug; it provides a broader range of coverage of resistant cell lines in a heterogenous tumor population; and it prevents or slows the development of new resistant lines.

The following principles have been useful in the selection of drugs in the most effective drug combinations, and they guide the development of new programs:

1. Only drugs known to be partially effective when used alone should be selected for use in combination. If avail-

' I .j


Impact of the Goldie-Coldman Hypothesis on Design of Clinical Trials Using Combination Chemotherapy 279

able, drugs that produce some fraction of complete remission are preferred to those that produce only partial responses.

2. When several drugs of a class are available and are equally effective, a drug should be selected on the basis of toxicity that does not overlap with the toxicity of other drugs to be used in the combination. Although such selection leads to a wider range of side effects and more general discomfort for the patient, it minimizes the risk of a lethal effect caused by multiple insults to the same organ system by different drugs and allows dose intensity to be maximized.

3. Drugs should be used in their optimal dose and schedule. 4. Drug combinations should be given at consistent inter

vals. The treatment-free interval between cycles should be the shortest possible time necessary for recovery of the most sensitive normal target tissue, which is usually the bone marrow.

Omission of a drug from a combination may allow overgrowth by a cell line sensitive to that drug alone and resistant

· to other drugs in the combination. Also, arbitrarily reducing the dose of an effective drug to add other less effective drugs may reduce the dose of the most effective agent below the threshold of effectiveness and destroy the capacity of the combination to cure that particular patient.

Bone marro)V ha5 a storage compartment that can supply mature cells to the peripheral blood for 8 to 10 days after the stem cell pool has been damaged by cytotoxic drugs. Events measured in the peripheral blood are usually a week behirtd events occurring in the bone marrow. In previously untreated patients, leukopenia and thrombocytopenia are discernible on the ninth or tenth day after initial dosing. Nadir blood counts are noted between days 14 to 18, with recovery apparent by day 21 and usually complete by day 28. Prior treatment with drugs or x-radiation may alter this sequence by depleting the stem cell pool, shortening the time to the appearance of leukopenia artd thrombocytopenia, and prolonging the recovery time. Curiously, when the second half of a combination given in the clinic on a day l, day 8 schedule is omitted, leukopenia and thrombocytopenia comparable with that seen with the full combination usually occur, suggesting that the second set of doses does not cause an equal increment in bone marrow suppression, possibly because the stem cell com'partment is still in a quiescent state. This result also. suggests that, in most instances, the day-8 doses can be given safely, even if leukopenia and thrombocytopenia have already become evident. The interval of greatest importance in the clinic is the duration of the nadir level of leukocytes and platelets. The highest risk of infection or bleeding occurs with granulocyte counts lower than 500/dl and platelet counts lower than 20,000/dl. If this nadir lasts only 4 to 7 days, it is tolerated by most patients without supplemental support. Increasing doses of most anticancer drugs, within the range of the maximally tolerated dose, usually does not ablate the marrow or even prolong the time to recovery. Repeated dosing during the phase of early recovery of the marrow (days 16-2.1), however, may cause severer toxicity in the second treatment cycle in patien~s whose marrow is not the source of, or involved with, the tumor.

These types of data led to the familiar 2-week interval be-

tween cycles of the most effective drug combinations (new cycles begin on day 28 after the first dose) to accommodate the recovery time of human bone marrow. Although this treatment schedule is suitable for some tumors, the regrowth characteristics of others, such as diffuse large cell lymphomas, Burkitt's lymphoma, and leukemia, often permit the tumor volume to return to pretreatment levels in the interval required for bone marrow recovery, and other approaches to cycling drug combinations are now being explored. One such approach has been to use non-marrow-toxic chemotherapeutic agents, cycled with marrow-toxic agents, to permit the bone marrow to recover despite continuous treatment. This method has been useful in patients with rapidly growing diffuse large cell lymphomas. It is limited by the sensitivity of the tumor in question to the available·non-marrow-toxic agents. The availability of colony-stimulating factors as supportive tools (see Chap. 62, section 2) is altering the design of clinical trials as well. Colony-stimulating factors have been coupled with cytotoxic combination chemotherapy, and the nadir leukopenia can usuallY be avoided or ameliorated.28

-32

IMPACT OF THE GOLDIE-COLDMAN HYPOTHESIS ON DESIGN OF CLINICAL TRIALS USING COMBINATION CHEMOTHERAPY

In 1943, Luria and Delbruck described a principle in bacterial genetics important to our understanding of the development of spontaneous resistance to cancer chemotherapy that led to a recorisideration of how chemotherapy was used in clinical trials.33 They observed that the bacterium Escherichia coli developed resistance to bacterial viruses (bacteriophage), not by surviving exposure, but by expanding clones 9f bacteria that had spontaneously mutated to a type inherently resistant to phage infection. In 1979, Goldie and Coldman applied this · principle to the development of resistance by cancer cells to anticancer drugs without prior exposure to these drugs. 34 They proposed that the nonrandom cytogenic ch~nges now known to be associated with most human cancers probably were tightly associated with the development of the capacity to resist the action of certain types of anticancer drugs.35 They developed a mathematical model that predicted that tumor cells mutate to drug resistance at a rate intrinsic to the genetic instability of a particular tumor and that these events would begin to occur at population sizes between 103 and 106 tumor cells (1000-1 million cells), much lower than the mass of cells considered to be clinically detectable (109, or 1 billion cells). The probability that a given tumor will contain resistant clones when a patient is newly diagnosed would be a function of the mutation rate and the size of the tumor. If the mutation rate is as infrequent as 10-a, a tumor composfd of 109 cells (a I-cm mass) would still be certain to have at least one drugresistant clone; however, the absolute number of resistant cells in a tumor composed of 109 cells would be relatively small. Therefore, in the clinic, such tumors should initially respond to treatment with a complete or partial remission, but then recur as the resistance clone(s) expanded to repopulate the mass or masses. Such a pattern is seen with the use of chemotherapy in many drug-responsive cancers in the clinic.

The Goldie-Coldman hypothesis, therefore, predicts that



resistance should be a problem even with small tumor burdens, and the maximal chance for cure occurs when all available

· effective drugs are given simultaneously.34 Because this would involve using 8 to 12 drugs simultaneously, this approach has not been tested in the clinic because of the fear that the use of more than five cytotoxic drugs, at full doses, would not be possible. The alternative approach, using two programs of equally effective, noncross-resistant drug combinations in alternating cycles, has been under evaluation for more than 10 years. Unfortunately, many studies purporting to test the Goldie-Coldman hypothesis have been poorly designed. In many instances, inadequate testing has been done to determine whether the alternate combination is truly noncrossresistant and as equally effective as the primary treatment (in most instances, it is not), 1 which it must be to fulfill the hypothesis as described by Goldie and Coldman. Second, dosing is rarely controlled, so that doses of essential drugs are modified downward, a priori, without testing the impact of such dose reductions on outcome. A more recent approach is the use of half of the drugs of each of the effective combination on days I and 8, respectively (so-called hybrid combinations). This approach is being tried in patients with Hodgkin's disease and diffuse large cell lymphomas. The use of alternating cycles of combination chemotherapy has not yet proved to be significantly more effective than are full doses of a single effective combination program.

In 1986, Day reanalyzed the Goldie-Coldman hypothesis and relaxed the requirement for symmetry in the model. 36•

37

Although his model verified the basic tenets of the GoldieColdman hypothesis, it suggested a different approach to sequencing combinations: In many instances, the sequential use of combinations should out-perform alternating cycles, because no two combinations are likely to be strictly noncrossresistant or have equal cell killing capacity, the symmetry assumed by Goldie and Coldman. Day formulated "the worstdrug rule," which refers to any strategy using more or earlier doses of a treatment that is the least effective of two or more available options. 38 The worst-drug rule has interesting implications. First, it is a nonintuitive approach. If two treatments are available, treatments A and B, and Bis known to be better, a physician is more likely to use B first. However, cells that are resistant to the best treatment, B, must be eliminated by the weaker program, A, and because it is the weaker program, one cannot wait too long to use it or the overgrowth of the population resistant to B will place the physician and patient in a situation difficult to overcome. The model predicts that, if six cycles of A and B are planned, using the weaker program-A-first performs better. There have been clinical examples in which sequential therapies have outperformed alternating cyclical use of the same programs if the dose intensity of the two regimens is carefully controlled. 39

No rigid schedule can accommodate all the variables assumed to be important for maximum effectiveness of combination chemotherapy and the requirements of the patients in the practice of medical oncology. Physicians must often adjust doses at intervals to administer drugs safely. The certainty that the therapeutic effect of a drug or drug combination can be lost if the dose or schedule is altered should temper these judgments. Reductions in dose rates also often result in only minimal decreases in toxicity but major reductions in the capacity to attain a complete remission in patients with

drug-responsive tumors. 38 The physician and the patient must consider the risk of dying from cancer prematurely compared with the transient benefits of reducing the acute side effects of treatment. Adhering to the standard sliding scale for dose adjustments, usually published with most new treatments, is the most appropriate way to make the necessary adjustments without compromising long-term outcome. In addition to providing guidelines for dose reduction, these sliding scales provide consistency between patients, and between studies, by preserving the intervals between cycles and the integrity of each ·drug combination. These alternatives and their potential impact on the quality and quantity of life should be made clear to patients as part of the informed consent process if they are to share intelligently in decisions about dose modifications made by their physicians. 40

RESPONSE TO CHEMOTHERAPY IS AFFECTED BY THE BIOLOGY OF TUMOR GROWTH

In the early 1960s Skipper and colleagues identified the guiding principles of present-day chemotherapy, using the rodent leukemia L12IO as a modeJ.41-43 Applying these principles to the drug treatment of human cancers required an understanding of the differences not only between the growth characteristics of this rodent leukemia and of human cancers but also in growth rates of normal target tissues in mice and humans. For example, Ll2IO leukemia is a rapidly growing tumor with a high percentage of cells synthesizing DNA, as measured by the uptake of tritiated thymidine (the labeling index). Because Ll2IO leukemia has a growth fraction of 100% (i.e., all of its cells are actively progressing through the cell cycle), its life cycle is consistent and prepictable.44

The relation between cell number and survival in L!2IO leukemia is linear, as shown in Figure 16-1. The time to death of animals bearing Ll2IO leukemia is the interval required to achieve a population size of about 109 (I billion) cells. With a growth fraction of 100% and a doubling time of 12 hours, 109 cells accumulate by 19 days after the injection of a single cell, by I 0 days after the injection of 105 cells, and by 5 days after the administration of IO' cells. Skipper and associates postulated that the increase in host lifespan after cytotoxic chemotherapy of L12IO leukemia was largely due to the cytocidal effect of treatment on the tumor cell popu-· lation. In these early elegant mouse experiments, they calculated the residual number of cells after treatment by extrapolating back from the duration of prolongation of life after a single treatment. An increase of 2 days in life would be equivalent to a 90% destruction of tumor cells (a I-log kill), or a reduction in the cell number from 106 to I05• A 99.999% destruction of tumor cells, a number that s"eems enormous to most clinicians: represents only a 5-log kill and does not cure animals unless the initial inoculum is small, perhaps 104 cells or fewer. If multiple treatments are given, the net tumor cell kill per treatment is the sum of the surviving cells plus the regrowth of the tumor cell population before the next treatment.

The killing effects of cancer drugs in this model tumor follow log-kill kinetics, that is, if a particular dose of an individual drug kills three logs of cells and reduces tumor burden from

r '

u 16 of 122 Celltrion, Inc., Exhibit 1006

Biochemical Resistance to Chemotherapy Is .the Major Impediment to Successful Treatment 281

10'

10'

• MEANS OF 193 EXPTS.

•MEANS OF 6-12 EXPTS.

A SINGLE EXPTS.

-=- 106

~ ::;; :> _J

:> () 0 ;;!:; u. 0 w to! (/)

105

10'

10'

102

101

.. 1., 1-.....L.-.l---.+-..i.46:..o/.c:c•_,S'-'U"-R'-'V-'-IV:...0:..iR.:.S::_._-'t,.__, 0 2 4 6 8 10 12 14 16 18 20

AVG. LIFE SPAN (DAYS)

FIGURE 16-1. Relation between size of tumor cell inoculation and time to death of the host in L1210 leukemia in CDF1 mice.

1010 to 107 cells, the same dose used at a tumor burden of 105

cells reduces the tumor mass to 102 • The cell kill, therefore, is proportional, regardless of tumor burden. This model fits the response of Ll210 murine leukemia to chemotherapy. When treatment failed in the experiments of Skipper and colleagues, it was because the .initial tumor burden was too high to allow the delivery of curative doses of chemotherapy to eradicate the last leukemia cell. The cardinal rule of chemotherapy-the invariable inverse relation between cell number and curability-was established in this model and applies to all others. Skipper and colleagues proceeded to show that with an understanding of these basic facts, this rodent leukemia could be cured by specifically designed doses and schedules tied to tumor volume and growth characteris.tics.41

Although murine leukemias seemed to follow exponential kinetics, available data suggested that most human tumors did not appear to grow exponentially. For example, the concept of log kill would have predicted that some large tumors in the

clinic should have Qeen more sensitive to treatment than has been experienced. In toto, the available data in human cancers supPort a Gompertzian model of tumor growth and regression. The critical distinction between Gompertzian and exponential growth is that in Gompertzian kinetics, the growth fraction of the tumor is not constant but decreases exponentially with time (exponential growth is matched by exponential retardation of growth). The growth fraction peaks when the tumor is about 37o/o of its maximum size. In a Gompertzian model, when a patient with advanced cancer is treated, the tumor mass.is larger, its growth fraction is low, and the fraction of cells killed is therefore small. An important feature of Gompertzian growth is that response to chemotherapy depends on where the tumor is in its growth curve. Gompertzian-gfowing tumors will respond to cytotoxic drugs in a Gompertzian fashion. Therefore, predictions can also be made about the behavior of small tumors, such as tumor burdens that might be present after primary surgical therapy. When the tumor is clinic:i.lly undetectable, its growth fraction would be at its largest and, although the numerical reduction in cell number is small, the f!actional cell kill from a ''known to be effectiVE!'' therapeutic dose of chemotherapy would be higher than it would be later in the tumor course. This observation initially was used to justify dose reductions at lower tumor volumes, which may account for sOme of the dis:ippointment in the outcome of adjuvant studies in breast cancer, because there is no scientific justification for such dose reductions if residual cancer cells survive the treatment. The Gompertzian growth model is important for another reason: It impacts on the patterns of regrowth of residual tumor cells. In breast cancer, Norton.has analyzed data from multiple adjuvant studies and also the only available studies of untreated patients who presented with localized cancers who were followed until death.37

•38.45

•46 He found that in all instances a Gompertzian

growth model precisely fitted the growth curves of these tumors. Jn adjuvant situations, the model showed that relapsefree survival and survival curves cannot discriminate between residual cell populations of I or I million cells, because the regrowth of residual cell populations will be faster at smaller residual volumes than it will be at larger residual volumes, producing identical endpoints sometimes at 5 years after diagnosis. Therefore, the effect of dose alterations cannot be differentiated using standard assay systems of effectiveness. Experimental observations imply that there are kinetic reasons for the failure of chemotherapy to cure large tumors and for the inadequate assessm~nt of the effectiveness of drug adjuvant treatment. The data imply that short of total eradication of micrometastases (cure), varying residual volumes produce similar 5-year relapse-free survival and obscure the . deleterious effect of dose reductions. This information has been useful in the design of new adjuvant treatment protocols in breast cancer.

BIOCHEMICAL RESISTANCE TO CHEMOTHERAPY IS THE MAJOR IMPEDIMENT TO SUCCESSFUL TREATMENT ·

The cancer cell presents a variable and moving target for anticancer drugs. The interrelation of pharmacokinetics and



tumor and normal target cell kinetics is the fulcrum of clinical cancer chemotherapy. The therapeutic and toxic effects of chemotherapeutic agents are related to the time the active principle is exposed in an effective concentration to its target (Fig. 16-2). The same degree ofcytotoxicity can be achieved, on differ:ent schedules, from the same concentration of drug multiplied by the time of exposure (C X T). This relation obtains across different species when the drugs are metabolized and excreted in a similar fashion. This principle has made it possible to translate doses of drugs devised in animals to humans for early clinical testing.47

•48 A given concentration

of drug multiplied by the time of exposure to its target (C X T) genentlly is equally cytotoxic in populations of cells with equivalent growth characteristics and sensitivity to the agent(s) in question.

When the active principles of an anticancer drug reach their target, however, another obstacle to the capacity to kill the cancer cell appears: specific and permanent biochemical resistance to the drug. Resistance to drugs occurs de nova in cancer cells (intrinsic resistance) or is concomitant to exposure to drugs (acquired resistance). 48- 57

Many specific mechanisms of drug resistance have been revealed whereby cancer cells demonstrate the ability to circumvent a well-defined pathway of attack by an individual cytotoxic agent. These mechanisms are discussed in detail in Chapter 18 and are summarized in Figure 16-2.49 Mechanisms of primary drug resistance include decreased uptake caused by changes in drug-specific transport mechanisms, de-

DNA Damage and Chemosensitivity

Treatment DNA Damage Direct•

UV radiation

X-radiation Bleomycin

Anti metabolites

Alkylators

Nitrosoureas

Mitomycin C

Cis-platinum

Anthracyclines

Aminoacridines

Ellipticines Epipodophyllotoxins

Camptothecin

Repair

Glycosylases

(AP) Endonucleases

Li gases

Exonucleases

Topoisomerases

Polymerases

FIGURE 16-2. Schematic illustration of DNA damage and· repair mechanisms in cells exposed to cytotoxic modalities. •Reversal/removal. AP, apurinic (apyrimidinic) site. (Epstein, RA. Drug-induced DNA damage and.turrior chemosensitivity. J Clin Qncol 1990;8:2062-2084)

creased activation of prodrugs, alteration in target enzymes of the drug, alterations in cellular metabolism and repair mechanisms, and increased inactivation of drugs. Gene amplification of an enzyme target has been documented to occur in a tumor as a result of exposure to the drug,57 with the attendant development of chromosomal homogenous staining regions or double-minute chromosomes representing an increased copy number of the target gene.

Some tumors do not respond to chemotherapeutic agents, however, even when diagnosed with apparently minimal tumor volume, which suggests that they are inherently resistant to drugs or are made up mostly of clones that have mutated to resistance and have become the dominant cell line in the population due to cell loss. As tumor masses grow, there is considerable cell loss from shedding of cells, for example, into the lumen of the bowel, which can amount to 90% of the total tumor volume. In such a setting, a tumor 1 cm in size and consisting of IO' cells, although appearing to be an early tumor, may have experienced as many as 1200 doublings rather than an estimated 32 doublings, if cell loss was not a factor, to reach that size to compensate for cell loss. According to the Goldie-Coldman somatic mutation hypothesis, such a kinetic history, together with the expected genetic instability, could be associated with a high probability that the entire mass consists of resistant cell lines.

It now appears that intrinsic resistance at the clinical level may be related to the expression of generic defense mechanisms in cells, whereas resistance that is related to individual mechanisms of action occtirs later because of the selection after exposure to the drug or drugs in question. One such generic resistance is that referred to to as multidrug resistance (MDR). When malignant cell lines are made resistant to a single natural product chemotherapeutic agent, by stepwise incubation in increasing amounts of drug, some such lines, curiously, are found to be resistant to structurally unrelated cytotoxic compounds. This phenomenon of broad resistance wa's termed pleiotropic drug resistance, or MDR. 58 Cell lines that display the MDR phenotype are generally resistant to natural product cytotoxic agents, such as the anthracyclines, vinca alkaloids, epipodophyllotoxins, and actinomycin D. Because all these agents are believed to have different mechanisms of action, investigation of MDR has focused on the cell's basic defense mechanism against toxic agents that humans are exposed to naturally in the environment.

MDR has been shown to be associated with decreased intracellular drug accumulation and the presence of a 170-kd plasma membrane-associated glycoprotein (P-glycoprotein) that is not detectable in most parenteral drug-sensitive lines. 59

•60 P-glycoprotein content has been directly correlated

with the degree of the decrease in intracellular accumulation of the toxins and the drug resistance exhibited by the cell. 61 ·62 These observations suggest that the appearance of P-glycoprotein is associated with resistance perhaps by regulating the transport of toxins out of the cell. Most, but not all, cell lines with the MDR phenotype that have since been established show increased expression of.the gene encoding P-glycoprotein, the MDR gene. 03- 10 Recently, the gene for cystic fibrosis has been cloned and shown to produce a protein with marked homology to the P-glycoprotein, and additional data suggest that impaired drug influx may be a separate phenomenon. A great deal of evidence suggests that the P-

r


glycoprotein is an energy-dependent drug efflux pump, whose primary function is to extrude chloride ions. P-glycoprotein also binds photoaffinity analogs of the natural product vinblastine, a reaction that is competitively inhibited by unlabeled vinblastine and by anthracyclines. as-72 Several agents, including the calcium channel blocker verapamil, quinidine, and nifedipine, can bind to the P-glycoprotein and compete with the vinblastine analogs for binding with the P-glycoprotein. Full-length cDNA sequences encoding the mouse and human P-glycoprotein gene have been isolated, and their nucleotide sequences determined. The deduced amino acid sequence of this protein shows structural siffiilarities to a wellcharacterized bacterial membrane transport protein. 12- 74 Pglycoprotein RNA expression has been found in high levels in normal adrenal gland and kidney tissue and in moderate . levels in liver and colon tissue. 75 Recently, P-glycoprotein has been shown to be expressed on CD-34+ bone marrow stem cells, but not in more mature cells of the hematopoietic system, and in the endothelial lining of blood vessels. Both observations have important implications for the design of treatment protocols.8 ·

76·82 Because colon, kidney, and liver

tissues are exposed to naturally occurring environmental toxins, the role of the P-glycoprotein in health may be one of protecting by facilitating efflux of these toxins or to serve as an alternative ion transport mechanism.

The range of expression of P-glycoprotein by human tumors is under intensive st'udy.77

-79

•82

-84 In general, P-glycoprotein

is highly expressed in tumors intrin.sically resistant to chemotherapy and poorly expressed in drug-responsive tumors, unless they are sampled after relapse. These data provide avenues to pursue clinically. Protocols can be designed to use agents not affected by the pump in resistant tumors, or natural product anticancer drugs can be given early in the course of the disease, before MDR is expressed in most cells. 38 Agents that block the pump mechanism also can be used simultaneously with natural product anticancer drugs. These approaches are reviewed in reference 38 and in Chapter 69, section 6.

Another type of multidrug resistance often follows on the heels of that associated with the P-glycoprotein expression and is associated with the topoisomerase enzymes.85-

88 Topoisomerases are necessary for DNA replication, and they catalyze changes in the secondary and tertiary structures of DNA to relax DNA tension during transcription and cell division. Topoisomerase II appears to be the enzyme that is the target of antineoplastic drugs that act as DNA-intercalating agents, such as etoposide and the anthracyclines. The only known class of topoisomerase I-targeting drugs-the camptothecins-has shown unprecedented activity against human cancers in animal models and is currently undergoing clinical trials. Topoisomerases, therefore, may represent the final common pathway of cytotoxicity of several different classes of antineoplastic agents. Drugs that act through interaction with the topoisomerases do so by preventing the religation of DNA through the formation of cleavable complexes. Resistance may occur through alterations in production and function of the enzymes. For example, an etoposide-resistant· Chinese hamster ovary cell line that was cross-resistant to the structurally dissimilar agents m-AMSA, mitoxantrone, and the anthracycline doxorubicin demonstrated altered topoisomerase II activity. 89 In addition, alteration in the topoisomerase

Concept of Dose Intensity 283

I-like activity was found in Chinese hamster cells selected for resistance to elipticine and cross-resistance to m-AMSA and etoposide. 90

As a result of these data, there is considerable excitement about the prospect of improving the effectiveness of chemotherapy by circumventing both types of MDR, by preventing the development of resistance or by interfering with the mechanisms themselves.

CONCEPT OF DOSE INTENSITY

For drug-sensitive cancers in favorable kinetic circumstances, the factor limiting the capacity to cure is proper dosing. The dose-response curve in biologic systems is usually sigmoidal in shape, with a threshold, a lag phase, a linear phase, and a plateau phase. For radiation therapy and chemotherapy, it is the difference between the dose-response curves of normal and tumor tissue that must be exploited during treatment. In experimental models, the dose-response curve is usually steep in the linear phase. Almost without exception in rodents bearing transplantable tumors, a reduction of doses in the linear phase of the dose-response curve results in a loss of the capacity to cure the tumor before there is a diminution in the response rate. That is, complete remissions will continue to be observed, but with dose reduction as small as 20%, the last few residual cells may not be eliminated, and relapse is inevitable. There is an extremely important lesson in these animal data for clinicians who, in their daily practice, judge the adequacy of their therapy by measuring the response rate of visible or palpable tumor masses and only much later are able to evaluate the treatment by survival results. This point is illustrated in Table 16-2, which summarizes data from numerous experiments conducted by Skipper and colleagues at the Southern Research Institute using the transplantable and palpable Ridgway osteosarcoma tumor model.91 •92 Reduction in the average dose intensity of the two-drug combination of L-phenylalanine mustard (L-PAM) and cyclophosphamide causes a marked decrease in the cure rate before a significant reduction in the complete remission rate occurs. On the average, a dose reduction of approximately 20% leads to a Joss of 50% of the cure rate. The converse is also true. In high-growth-

TABLE 16-2. Ridgway Osteogenic Sarcoma: Response to Different Dose Intensity of Two-Drug Combination of Cyclophosphamide and L-PAM

RD/

CPA L-PAM Average %CR % Cures

0.38 0.82 0.60 100 60 0.75 0.18 0.47 100 44 0.25 0.55 0.44 100 10 0.50 0.12 0.31 10 0 0.17 0.36 0.27 0 0

RDI, relative dose intensity; CPA, cyclophosphamide; L-PAM, L-phenylalanine; CR, complete response. Tumors weighed 2-3 g. (Modified from Skipper HE. Booklet No. 5. Birmingham: Southern

, Research Institute, 1986)



TABLE 16-3. Sample Calculations: Dose Intensity, Relative Dose Intensity, and Average Relative Dose Intensity

Calculation of Dose Intensity Test Schedule Cyclophosphrunide 80 mg/m2/d (continuously)

Calculation of Relative Dose Intensity Standard Cyclophosphamide 80 mg/m2/d (continuously)

Test Schedule Cyclophosphamide 100 mg/m2/d (d 1-14, q 28 d)

Relative Dose Intensity Dose Intensity

560 mg/m2/wk

560 mg/m2/wk

350 mg/m2/wk 350/560 ~ 0.62

Calculation of Average Relative Dose Intensity Standard* Cyclophosphamide 2 mg/kg/d Methotrexate 0.7 mg/kg/wk 5-Fluorouracil 12 mg/kg/wk

Test Regimen Cyclophosphamide 100 mg/m'/d (l-14) Methotrexate 40 rng/m2/d 1, 8 5-Fluorouracil 600 mg/m2/d l, 8 Repeat cycles every 28 d

560 mg/m2/wk 28 mg/m2/wk

· 480 mg/m2/wk

350 mg/m'/wk 20 mg/m2/wk

300 mg/m2/wk

350/560 ~ 0.62 20/28 ~ 0.71

300/480 ~ 0.62 Average 0.65

•Assume standard regimen to be CMF. To convert mg/kg to mg/m2, multiply by 40.

(Hryniuk WM: The importance of dose intensity in the outcome of chemother.apy. In: DeVita VT, Hellmaii S, Rosenberg SA, eds. Important advances in oncology. Philadelphia: JB Lippincott, 1988: 121-142)

fraction tumors, a twofold increase in dose often leads to a tenfold increase (1 log) in tumor cell kill. Although animal models are not the perfect analog for human cancers, the invariable nature of these data indicates that the general principle is transferable to the clinic and is ignored at great peril. Because anticancer drugs' are toxic, it is often appealing to avoid acute, but not life-threatening, toxicity by diminishing the dose or increas~ng the intervals between cycles of treatment. This kind of ad hoc adjustment of dosing is probably the main reason for treatment failure in patients with drugsensitive human tumors undergoing their first chemotherapy treatment.

It has been difficult to compare the impact of different dosing practices in treatment programs. Hryniuk and colleagues analyzed treatment outcome in a number of different tumors as a function of what they have termed dose intensity. 93- 9s They defined dose intensity as the amount of drug delivered per unit of time, expressed as mg/m2/wk, regardless of the schedule or route of administration. Relative dose intensity (RD!) is the amount of drug delivered per unit of time relative to an arbitrarily chosen standard single drug, or, for a combination regirnen, the decimal fraction of the ratio of the average dose intensity of all drugs of the test regimen compared with the standard regimen. A sample calculation of the RD! for a commonly used regimen, the cyclophosphamide, metho!rexate, 5 .. f!uorouracil (CMF) combination for breast cancer, is provided in Table 16-3.94 To calculate the average RDI for a regimen containing fewer drugs than the standard regimen, a dose intensity of zero is assigned to the missing drug(s),

and the average RD! of the test regimen is divided by the total number of drugs in the standard.9' The dose intensity of various programs is compared over whatever time frame the treatment programs are administered. Calculations can· be made of intended dose intensity, the dose intensity as described in the treatment protocol, or actual or received dose intensity. Received dose intensity calculations are more useful data because they reflect the impact of dose reductions and necessary treatment delays imposed in actual practice.

Because calculations are made based on the amount of drugs given per week, regardless of schedule, treatment delays are given equal weight with dose reductions. Calculations of the dose intensity, therefore, require the assumption that differences in scheduling does not influence treatment outcome. Although this first appears to be heretical, close scrutiny of all available data in humans and rodents shows that scheduling . influences outcome mostly by affecting·toxicity, in this way allowing higher doses to be administered over the same time frame. An example can be found in the use of methotrexate in rodents and humans. Daily administration of low doses of methotrexate is extremely toxic and severely limits the dose and duration of therapy with this drug. A twice-weekly schedule, which is much more effective in rodents' and humans, allows much higher doses to be delivered for longer durations, because this schedule is associated with less toxicity. The dose intensity of the twice-weekly schedule, therefore, is much greater than that of the daily oral schedule, if the calculatidn is based on mg/m2/wk of delivered drug. In practice, the impact of scheduling on the calculation of dose intensity can be

r


neutralized by comparing programs in which drugs with toxicities affected by scheduling, such as the antimetabolites, are given on similar schedules.

Calculation of an average RDI of a drug combination also assumes that each drug has an equal efficacy against the tumor in question. However, the impact of any single drug or combinations of two or three drugs in a multidrug combination can be assessed separately. This measurement has been done by Hryniuk and colleagues tci show the greater impact of cisplatin in a drug combination for ovarian cancer94

-98 and by

others99-

102 to show the importance of adequate doses of alkylating agents and vinca alkaloids in lymphoma treatment. This kind of analysis can help identify the most effective drug in a combination and is important because such data can help avoid adjustments that radically alter the effectiveness of a program. Also, by identifying the most important drugs, protocols can be developed that emphasize the dose intensity of these agents compared with other less effective drugs.36 Negative alterations in dose intensity of the most effective drug in a combination of drugs has great impact, as illustrated in Table 16-4, which displays the effects of the two-drug combination of L-PAM and the antimetabolite 6-mercaptopurine (6-MP) against the Ridgway osteogenic sarcoma model. In this instance, L-PAM is the more effective drug. The relation of average dose intensity of the two drugs to outcome is erratic, but the relation of the dose intensity of L-PAM to outcome is linear. Decreases in the dose intensity of L-PAM reduce the effect of the combination, even if the dose of 6-MP is increased to compensate for these reductions. In fact, any decrease in the dose intensity of L-PAM below 55% of the optimal singledose schedule results in a loss of the capacity of this combination to cure animals, regardless of the dose of 6-MP.

TABLE 16-4. Ridgway Osteogenic Sarcoma: Effect of Varying Dose Intensity of More Effective Drug, L-PAM

Relative Dose Intensity

Ratio Observed

L-PAM 6-MP (L-PAM/6-MP) Average %CR % Cures

0.82 0.49 1.7 0.66 100 60 0.73 1.3 0.56 1.0 90 50 0.55 1.0 0.55 0.78 90 20 0.55 0.33 1.7 0.44 80 20 0.36 0.67 0.54 0.52 56 0 0.36 0.21 1.7 0.29 30 0 0.27 1.5 0.18 0.89 70 0 0.24 0.44 0.57 0.35 0 0 0.24 0.15 1.6 0.20 0 0 0.18 1.0 0.18 0.59 0 0 0.12 0.67 0.18 0.50 0 0 0.08 0.44 0.18 0.26 0 0

L-PAM, L-phenylalanine mustard; 6-MP, mercaptopurine. Tumors weighed.2-3 g. Varying the dose intensity of L-PAM has a greater impact on outcome than can be overcome by increasing the dose of 6-MP. (Skipper HE. Booklet No. 4. Birmingham: Southern Research Institute, 1986)

Concept of Dose Intensity 285

To judge adequately the dosing of a particular protocol, data on total dose of each drug used and cumulative doses of each drug are necessary. Collection of such data is not part of routine practice, and reports are not generally available in the literature. To assess the impact of dosing schedules in practice and in clinical trials, such data should be required before papers are accepted for publication.

A positive relation between dose intensity and response rate has been demonstrated in advanced ovarian, breast, and colon cancers and in the lymphomas. 92.93.95.98,100-102

Calculations of the impact of dose intensity on outcome are particularly in1portant in estimating the value and exploring some of the pitfalls of adjuvant chemotherapy. The steep doseresponse curve for anticancer drugs indicates that dose reductions in adjuvant drug treatment programs are likely to be associated with significantly less therapeutic effect. Dose reduction, however, has been the norm in the design of adjuvant trials. An example is the standard CMF regimen for breast cancer referred to in Table 16-3. The model for the regimen was published in 197 4 by Canellas and associates. 100

It produced an impressive complete remission rate of approximately 30%, but its toxicity was considerable. As a result, when it was advanced for use in a cooperative group setting, first for advanced disease" and later for adjuvant trials by Bonadonna and colleagues, 1°

1 its doses were arbitrarily reduced without pretesting the impact of such reductions on outcome. In ·addition, further reduction was made, a priori, for patients older than 60 years of age, with the assumption that such reductions would be required because of age. When the effec;:t of these reductions is related to outcome, there is a strong suggestion of a negative impact.99

·102 In premeno

pausal women the differences in relapse-free survival at the high and low doses of CMF are statistically significant. The most important point, however, is that the average dose intensity of CMF as used in clinical trials and in the community is probably only half the dose intensity of the original program. These dose reductions exceed the levels that animal models predict would lead to a loss in the capacity to cure.

Another example of the potential impact dose intensity can have on the design of clinical trials has been provided by Hryniuk, as shown in Figure 16-3.93 The dose intensity of 5-:fluorouracil is plotted against the response rate for advanced colorectal cancer in panel A. Points indicated by the asterisks are from a single study in which response was reported for actual delivered doses at three different levels. ' 03 The steep nature of the dose-response curves should be noted. Panel B of Figure 16-3 plots the same three points from the single study but adds the doses used in four published adjuvant stud- . ies. 103- 10.8 The doses in all of these studies are well below the level that most investigators would consider the threshold for producing useful responses in advanced colorectal cancer. The effect of dose intensity on the capacity to cure advanced Hodgkin's disease and diffuse large cell lymphomas is also striking and is described in detail in Chapters 51 and 52.

Increasing the dose intensity can. be a useful way to improve the effect of certain drugs or combinations of drugs, but it is not useful in all clinical circumstances. Large tumor burdens tend to shift the dose-response curve to the right. At the low end of the curability curve (i.e., in the presence of the highest tumor burdens), increasing the dose intensity often leads to unacceptable toxicity and may not produce more impressive



50

40

~30 <!... * a: ooo/ a. + a: 20

VASAG1+1 t * u

·v 10

N.Y.GP * 0

400 800 1200 1800 2400

Dose Intensity of 5-FU (mglm'lwk) A

50

40

• $ 30 1;.: a: a. + •• a: 20 ;:· u

•• 10 :·

>!< .

0 400 800 1200 1800 2400

Dose Intensity of 5-FU (mg/m 2/wk)

B FIGURE 16-3. (A) Response rate at various intended dose intensity of 5-ftuorouracil in advanced colorectal cancer. Each point represents results from one arm of a randomized trial. Asterishs indicate results of three doses from a single study, and solid circles indicate received dose intensity. (B) Dose interisities of 5-fluorouracil used in four adjuvant studies of colorectal cancer superimposed on the dose-response line for advanced disease shown in panel A. Asterishs represent received dose intensity from single study (see text). (Hryniuk WM. The importance of dose intensity in the outcome of chemotherapy. In: DeVita VT Jr, Hellman S, Rosenberg SA, eds. Important Advances in Oncology 1988. Philadelphia' JB Lippincott, !988,J25)

treatment outcomes, because the dose-response curve is flat. In addition, regimens that are already curing nearly 100% of a subset of patients, such as the combination of platinum, vinblastine. and bleomycin in low-burden testicular cancer and the nitrogen mustard, vincristine, procarbazine, prednisone (MOPP) chemotherapy program in stage IIIA Hodgkin's disease, cannot be expected to be improved on by augmenting . dose intensity. However, for most drugs and most tumors, there appears to be a threshold dose that produces responses, and the remarkable success of high-dose chemotherapy programs with bone marrow transplantation support in refractory lymphomas, breast cancer, childhood sarcomas, and neuro-

blastomas suggests that maximizing dose intensity can improve the chances of cure in drug-res_ponsive tumors.

IN VITRO TESTS TO SELECT CHEMOTHERAPEUTIC AGENTS FOR INDIVIDUALIZED TREATMENT

Short-term assays are not useful in determining the primary treatment for patients for whom a known and effective treatment exists. The assays are also of minimal value for the remainder ofnewly diagnosed patients and for those with drugsensitive tumors who fail the first trial of chemotherapy. The basic problem is that the pool of available active drugs is too small for most cancers to make those assays useful beyond clinical judgment. They can, however, be of some use to avoid patient exposure to the toxicity of drugs that are unlikely to be effective, but the tests are too cumbersome and expensive for routine practice. No convincing reports in the literature have indicated that short-term assays provide additional benefi\ beyond what the clinician can provide by. using good judgment and a knowledge of the effectiveness of the limited number of available single agents .

CANCER DRUG DEVELOPMENT

The steps in the development of anticancer drugs are shown in Figure 16-4. The most important step in the drug selection process is mass screening, the mechanism used to narrow the universe of chemicals potentially useful for the treatment of human cancers to a manageable number of high-priority drugs for clinical testing. 105

-109 The mass screening effort is con

ducted by the National Cancer Institute (NCI) drug development program.

Many currently available anticancer drugs active against human leukemias and lymphomas were identified and developed as a result of the NCI system. The input to this type of screening program reached its maximum of 40,000 compounds screened per year in 1975. In 1975, a major change was made in the NCI screening program because of the avail-

ACQUISITION

t SCREENING

t PRODUCTION AND FORMULATION

t TOXICOLOGY

t PHASE I CLINICAL TRIALS

t PHASE II CLINICAL TRIALS

t PHASE llHV CLINICAL. TRIALS

t GENERAL MEDICAL PRACTICE

FIGURE 16-4. Steps in cancer drug development.


ability of new rodent models. More rational selection of compounds was coupled with a panel of transplantable rodent tumor screens designed to mcitch the histologic type of common visceral cancers. These rodent solid tumor screens were matched to human tumor cell lines of the same type grown in nude mice. This panel posed the question of the clinical specificity of the preclinical models. The panel of tumors was changed periodically to pose additional questions to the screening process. 110 Taxol, a new 3.gent of considerable interest, was not active in traditional screens but was selected by the new screening panel. Later, human tumOrs grown in soft agar and under the renal capsule were also introduced into the screening program to further test the hypothesis that the use of human tissue in short-term assays could better select compounds more active in the clinic than could simpler rodent tumor models. Problems with the technical details of these in vitro system.s led to their discontinuation. 110

-113

As it became possible to maintain human tumor cells in defined media, the screening program was again changed by. developing disease-oriented panels of human tumor cell lines grown in defined media.113

·115 The initial selection of cell lines

for this screening panel was based on several considerations, including the use of representatives of major histologic subtypes, the use of multiple cell lines for each tumor type, and the use of cell lines that retain appropriate features of the tumor of origin. The cell lines currently in use include lung, ovarian, and renal cancer; malignant melanoma; brain tumors; and leukemia. Because of the interest in the phenomenon of MDR and the likelihood that it is one of the factors limiting the effectiveness of chemotherapy, a human breast cancer line and an MDR variant of this line, selected for resistance to doxorubicin, are included, along with a murine leukemia and a comparable doxorubicin-induced MDR variant of this murine leukemia. These cell lines provide the potential for identifying new agents with particular activity against MOR cell populations.

A key element in screening strategy is to maintain the capacity for high-volume screening. The most commonly used approach is a colorimetric growth inhibition assay that is based on the metabolic reduction of the tetrazolium salt formazan inside viable cells. Under appropriate conditions, a linear relation is obtained between viable cell number and formazan optical density, measured using a standard enzyme immunosorbent assay plate reader. 116 Automation of this assay has made it possible to main.tain an adequate volume of in vitro screening (10,000 compounds per year) at less expense. Preliminary analysis of screening results indicates that individual . cell lines show characteristic degrees of in vitro chemosensitivity to individual test compounds with known patterns of clinical activity. Ease of automation of the colorimetric assay and the stability of the cell lines have largely overcome the technical problems associated with clonogenic or subrenal capsular assays. 113

•117 The central goal of the in vitro-based,

disease-oriented screening program is to identify new antitumor drug candidates that would not have been disco~ered by the previously available screening program. Clinical testing of these new leads, as with previous versions of preclinical screening, will ultimately be the only way to establish or disprove the validity of the new screen for identifying new drugs active against the common refractory human solid tumors.

In the early days of scree.ning, the acquisition of agents was

Cancer Drug Development 287

purely random. Random acquisition of chemicals for screenihg was associated with two major problems: repetitious screening of compounds already tested and screening of analogs of drugs already known to be active rather than the identification of new structures. Modern molecular biology techniques present an unusual opportunity to select materials, defined at the molecular level, that might prove useful in the inhibition of vital cell functions.

Inherent in all screening systems is the tenet that biologic activity in some preclinical system must be demonstrated before human testing is performed. No currently marketed, useful chemotherapeutic agent is devoid of such preclinical antitumor effect.

Toxicology testing has evolved during the last decade from complicated testing in rodents, dogs, and monkeys to a less expensive and simpler system that relies on toxicity testing primarily in· mice. Large amounts of data accumulated since the beginning of anticancer drug development have allowed

· comparisons to be made across species with respect to common toxicity of chemicals. These data have shown that there is no safety advantage in using larger animal species instead of rodents. In tbe current system, implemented in 1980, the dose-response curve of a new drug is first developed in mice. The lethal dose (LD) in 10% of animals is determined, and the reproducible lethal dose in 10% of tested animals (LD10)

is used as the basis for establishing the initial dose in Clinical trials. Usually, 10% of the LD10 dose in rodents is selected for the initial human dose; this dose is first tested for toxicity in dogs, before uSe in humans, 'to minimize the risks associated with ad:rµinistering an unknown compound to humans. Although the correlation of toxic effects on rapidly dividing normal tissue between rodents, dogs, monkeys, and humans is good, correlation of other toxic effects is not as consistent. 118

Therefore, routine pathologic examination of rodent tissue is not always performed before clinical testing.

All drugs should be given with reference to body weight or surface area. The preferable reference point is body surface area, because better cross-species comparisons can be made, and because calculations based on body surface area allow . doses to be determined for adults and children without further

TABLE 16-5. Representative Surface Area to Weight Ratios (km) of Various Species' ·

Surface Area Body Weight Surface to Weight

Species (kg) Area Ratio (km)

Mouse 0.02 0.0066 3.0 Rat 0.15 0.025 5.9 Monkey 3 0.24 12 Dog 8 0.40 20 Human

Child 20 0.80 25 Adult 60 1.6 37

• To express a mg/kg dose in any given species as the equivalent mg/ m2 dose, multiply the dose by the appropriate km. In the adult human, for example, 100 mg/kg is equivalent to 100 mg/kg x 37 kg/m2

~ 3700 mglm'.



TABLE 16-6. Equivalent Surface Area Dosage Conversion Factors*

Mouse, Rat, Monkey, Dog, Human, 20 g 150 g 3 kg 8 kg 60 kg

Mouse ~ ' l h ' Rat 2 I ' t t • Monkey 4 2 I ' !

' ' Dog 6 4 • ~ ' Human 12 7 3 2

"' This table shows approximate factors for converting doses expressed in terms of mg/kg from one species to an equivalent surface area dose expressed in the same terms mg/kg in the other species. For example, given a dose of 50 mg/kg in the mouse, what is the appropriate dose ii:i the human, assuming equivalency on the basis of mg/ m2?

50 mg/kg X iz = 4.1 mg/kg

adjustment. The assumptions leading to the dose conversion factors have been described in detail by Freireich and coworkers48 and are shown in Table 16-5, which is useful in converting doses in milligram per kilogram to the comparable milligram per square meter dose. Table 16-6 shows the procedure for conversion of a milligram per kilogram dose in rodents, monkeys, or dogs to the equivalent dose in hum.ans.

EARLY CLINICAL TRIALS OF ANTITUMOR AGENTS

Antitumor agents pass through four phases of clinic~! testing before they are accepted for general medical practice, marketed, or discarded (see Fig. 16-4).'"-"' The average time from discovery of an effective antitumor agent to marketing of that agent is quite long, in the range of 10 to 12 years, and is expensive, costing $40 million to $80 million.

Although the main purpose of phase I trials is to identify a maximally tolerated dose (MTD) in one of several schedules suggested by the preclinical data, patients are entered into phase I trials with therapeutic intent. For most of the effective anticancer drugs, some therapeutic effect was often seen even in phase I trials. Because a limited number of patients with a variety of diseases are treated in phase I trials, and because doses may be less than the ultimate the~apeutic range in a fraction of the patients, the absence of any positive effect in a phase I trial is not a sufficient reason to discontinue testing of a drug. The only reason not to proceed to a phase II study is prohibitive toxicity in phase I trials. Escalation of doses in phase I trials is usually done by a modified Fibonacci system: Doses are first doubled and then increased at decreasing increments of 66%, 50%, and 33% in succeeding groups of patients (usually three at a time) untillimiting toxicity is observed.119 Attempts have been made to rationalize and accelerate dose escalation by the systematic use of preclinical pharmacologic data. 124 This approach has relied on the assumption that the elimination rate of a drug determines its time concentration curve (C XT), and it further assumes that for agents showing no major differences in target cell sensitivity, schedule dependence, or toxicity between mice and

humans, the C X T at the mouse LD'° and the human MTD should be similar. These assumptions lead naturally to a simple algorithm for escalating doses by targeting the human C X T in a phase I trial to the mouse C X T at LD10 •

121 The steps are as follows:

1. Determine the mouse LD,, (part of the routine preclinical toxicology testing discussed earlier).

2. Determine the mouse C X T at LD,, . 3. Begin human testing at a safe starting dose (currently

one tenth of the mouse-equivalent LD,,). 4. Determine the human C X T at the starting dose in

phase I. 5. Escalate doses in subsequent patients based on how close

the C X T at the starting dose is to the target C X T.

Preliminary studies have suggested that applying this procedure may save 20% to 50% of the esc:i.lation steps for many agents. This approach is now being tested prospectively in the NCI phase I testing program.

The definition of a dose as maximally tolerated depends on how much toxicity the patient and physician are willing and able to tolerate. It has been amply demonstrated that for several drugs, such as cyclophosphamide, thiotepa, l ,3-bis(2-chlorethyl)-1-nitrosourea, and etoposide, the MTD, as determined from toxic effects other than bone marrow suppression, is three to ten times higher than the conventional MTD determined by granulocytopenia. The fact that the response rates . are commonly a function of dose gives a strong impetus to further trials exploring the upper end of the dose-response curve. As a result, an alternative approach to phase I testing is under consideration, that is, to redefine the MTD as the dose at which unacceptable non-marrow-related toxicity supervenes, despite deployment of all the modem aspects of care. This approach should be greatly facilitated by the availability of colony-stimulating factors, if they succeed in eliminating bone marrow suppression as the rate-limiting step in early tes'ting. It seems prudent to delay decisions on escalation· of new agents past the conventionally determined MTD until more information about their clinical characteristics is available.

The purpose of phase II studies is to develop estimates of the response rate to a particular drug of patients with specified tumor types. Phase II studies determine activity rather than clinical usefulness and answer a biologic and clinical question. The outcome of the phase II trial is a decisive point in the development of a drug. These results determine whether a new treatment should be pursued.

When a drug enters phase II testing in individual diseases, it should be tested in a patient group with easily, evaluated endpoints of responsiveness, provided it is ethically permissible to do so. This ambition is best fulfilled by enrolling patients with advanced cancer, but who have a maximal performance status, a minimal heterogeneity of metastatic sites, and a minimal amount of prior chemotherapy .125 This means that for tumors sensitive to chemotherapy·, patients who have failed no more than one prior regimen are ideal for study. For the less sensitive epithelial cancers, in many instances previously untreated patients can and should be entered into phase II studies. In view of the poor record of most single agents in heavily pretreated patients with advanced disease, such a strategy seems sensible, because for patients with advanced


drug-resistant cancer, the likelihood of toxicity is much greater than is the likelihood of therapeutic benefit.

The number of patients accrued to phase II trials should be appropriate for the scientific goals of the study. In ideal circumstances, a drug that produces no antitumor effect in 14 patients with the same tumor type, particularly if the heterogeneity of the distribution of metastases is minimized, has a greater than 95% chance of being ineffective against that tumor and could reasonably be eliminated from further studies against that specific cancer. One or two responses, however, increase the chance of efficacy sufficiently to dictate an expansion of the trial to 30 or more patients, so as not to miss a drug with a response rate in the 20o/o range. In general, partial response rates higher than 20% place the agent in a category of potential clinical usefulness to be determined in further studies. Response rates in the range of 5o/o to 10% are consistent with observer variation in phase II trials. Response rates lower than than 20% can be meaningful, however, if the quality of the response is good. For example, a few. complete remissions, even if the overall frequency of complete response is low, should lead to a decision to proceed with further testing in that disease, because complete disappearance of disease, however infrequent, is an important sign of a potentially effective new treatment. Because multiple doses and schedules may be tested, a phase II trial for each drug, schedule, and tumor type is required before a drug can be disqualified from further clinical use. Given all these confounding variables, a complete phase II trial often requires 600 or more patients.

At the completion of a phase II trial, a decision is made to. proceed with or discard the agent. This decision is based on a lack of efficacy or an excessive or intolerable toxicity, against the background of the observed therapeutic effect. Because it is not· possible to test each new agent against every tumor type, the potential for discarding agents that might be useful in rare tumors is significant. The early testing results of cisplatin are particularly instructive and illustrative of the problems faced by industry in cancer drug development. Cisplatin showed little activity against the common tumors in its early testing, because it was tested in heavily pretreated patients. Because its use was associated with considerable toxicity, it was almost discarded. Incidental testing in patients with testicular cancer, who are not generally part of major phase II studies, however, revealed interesting activity, and cisplatin was quickly advanced to use in combination with other drugs to treat advanced testicular cancer with curative intent. It ~as proposed for marketing on the basis of its usefulness in testicular cancer, but few data on its single-agent activity in this disease were available to the Food and Drug Administration in its appraisal of the new drug application, and marketing was delayed almost 3 years, despite nearly uniform enthusiasm for the drug in the oncology community. Only subsequent to widespread aftermarketing testing of cisplatin in other tumors did its broader effectiveness become apparent. This drug has now proved to be not only the mainstay of curative treatment of advanced testicular cani;:er but an important part of the therapy in lung and bladder cancer, head and neck cancer, ovarian cancer, and other common tumors. The U.S. Government recognized its role in delaying the marketing of this important drug by extending its patent life by 3 years.

If a drug is found effective in phase II trials, phases III and

Overcoming the Limitations of Cancer. Treatment 289

IV testing establishes its place in the therapeutic armamentarium. These clinical trials usually require large numbers of patients and are difficult to perform. The issue of randomized versus historical controls in phases III and IV trials is an important one and is discussed in detail in Chapter 19.

OVERCOMING THE LIMITATIONS OF CANCER TREATMENT

Since it has become apparent that chemotherapy could cure advanced cancers of some types, the question of why it could not cure more cancers, particularly those of the more common histologic types, has plagued investigators. In general, the limitations of cancer treatment have been difficult to overcome until the recent migration of molecular biology methodology to the clinic. The major limitations of cancer treatment are (1) the inability to determine precisely which cancers have metastasized at the time of diagnosis, (2) the inability to detect minimal residual disease after apparently successful treatment, (3) the.inability to escalate doses of effective anticancer drugs to the high end of the dose-response curve, (4) the hurdle presented by the unexpected expression of MDR, and (5) the inability to measure the moment-to-moment impact of treatment on cancer cells.

The issue of determining which cancers have metastasized has been addressed indirectly by using panels of prognostic factors, which has proven too crude for practical use in most cancers. It,is now being addressed on the molecular level by probing· for expression of genes that normally control cell migration; this subject is reviewed in Chapter 8. Molecular diagnosis has advanced to the stage in which the ability to detect minimal residual disease has improved from crude morphologic methods and other tests that detect approximately I cancer cell in 20 or 100 normal cells to the ability to detect one malignarit cell in 1 million normal cells, using tests such as the polymerase chain reaction. These approaches are reviewed in Chapter 6. We are nQw in the age in which the availability of colony-stimulating factors (see Chap. 62, section 2) and autologous bone marrow transplantation (see Chap. 62, section !) are effectively allowing significant dose escalation, and the problems presented by the expression of MDR are in focus and attackable at the clinical level (see Chap. 69, section 6). Although scientists and practitioners are not yet able, at the clinical level, to monitor the impact of cancer treatment on cancer cells moment by moment, techniques such as magnetic resonance spectroscopy and positron emission tomography are available that offer promise. The capacity to measure instantaneously the impact of treatment should radically alter schedules of drug administration to whatever is required to kill cancer cells and no more.

Chemotherapy can cure about 15% of patients with advanced cancers, many of which occur in young patients. The impact of these successes, measured in person-years of life saved, is, however, disproportionately great because the younger average age at diagnosis results in a highly significant salvage of person-years of productive life. Systemic treatment also bears a special burden, because of fear of toxicity in the minds of physicians and patients, beyond that associated with surgery or radiation therapy, and because the effects of systemic thera~y cannot be limited precisely to the region in-



TABLE 16-7. Evolution of Cancer Treatn:ient*

RSR Radiation (%) Date Surgery Therapy Systemic Therapy

(±0) 1894 Radical mastectomy X·Rays discovered

20 1920 Antibiotics 250 kv units Transplantable rodent tumors 1946 Supportive care Nitrogen mustard in

lymphomas 1955 Radical surgery Choriocarcinoma

33 1957 Micrometastases Cobalt units 1961 Linear accelerator Drug cures of leukemia and

advanced Hodgkin's disease

36 1970 Resection of metastases Radiosensitizers Adjuvant therapy Immunotherapy Hybridoma technology

Particle therapy MDR 41 1980 Conservative surgery Neutron generators Biologics

Reconstructive surgery Trea_tment planning Mo ABS with CT scans Dose intensity

ABMT 49 1985 Tailoring procedures to Hyperthermia Primary chemotherapy

other treatments Overcoming drug resistance Conformal RRx Biochemotherapy

1990 Detect capacity to Attacking the signaling metastasize system

Antisense compounds Monitor response to treatment Determine residual disease

*Illustrates the complexity of integrating advances in each field .over a long time span. RSR, relative survival rates from NCI Surveillance, Epidemiology, and End Results (SEER) Program; MDR, multidrug resistance; CT, computed tomography; ABMT, autologous bone marrow transplant; RRx, radiation therapy; MoABS, monoclonal antibodies (also see chapters on principles of surgical oncology [14], radiation therapy [15]. and biologic therapy [17]).

volved by tumor. Even though more sophisticated techniques of delivering systemic therapy to target organs have been developed, systemic toxicity is always a concomitant of systemic treatment. The same is true of biologicals. The promise of diminished side effects with the use of biological materials because they were natural products has not been fulfilled. A general principle is th.at all chemicals, natural or xenobiotic, when used in pharmacologic doses, produce significant side effects.

Patients cured of cancer by any modality generally find the to"xicity associated with the treatment a justifiable experience. For those patients with advanced unresponsive cancers, however, who have the burden of progressive tumor and impending death, the risks and side effects of treatment should be carefully balanced against the potential benefits. Patients who are offered systemic therapy with a potential for cure should be treated aggressively; those for whom palliation is the only choice should not be overly burdened with extremely toxic treatments that may prolong life minimally and in an uncomfortable fashion. In practice, this means that patients with metastatic cancer for which no known effective systemic treatment exists often are best advised to avoid treatment with standard anticancer drugs or to take part in one of the many clinical trials testing new treatments.

To expand the benefits of chemotherapy. more use of anticancer drugs before and after surgery and radiotherapy is

required when tumor burden is minimal, kinetic features of cell growth are favorable, and drug resistance is less likely to be present or easier to overcome. Because patients sometimes feel free of tumor after surgery, however, they may be less willing to accept the additional trauma of chemotherapy. unless the benefits are carefully explained and accurately and honestly balanced against the chances of recurrence.

Cancer treatment has progressed considerably, despite the frustrations. Nonetheless, the separation of the specialties remains a significant problem; it is often difficult to integrate newer approaches in each field into the practice of medicine. Each field has grown separately.' and yet the solutions to clinical problems require the careful integration of components of each specialty.36 The evolution of the three approaches to cancer treatment are summarized in Table 16-7, which illustrates the complexity of integrating fields that have evolved slowly over eight decades, a time span that considerably exeeds the lifespan of clinical investigators, the change agents in cancer medicine.

REFERENCES l. Marshall EK Jr. Historical perspectives in chemotherapy. In: Gofdin A. Hawking IF.

eds. Advances in chemotherapy. Vol I. New York: Academic Press, 1964:1-8. 2. Hersh SM. Chemical and biological warfare: America's hidden arsenal. New York:

Bobbs Merrill, 1968.


3. Alexander SF. Final report of Bari mustard casualties. Allied Force Headquarters, Office of the Surgeon, APO 512, June 20, 1944.

4. DeVita VT. The evolution of therapeutic research in cancer. N Engl J Med 1978;298: 907-910.

5. De Vita VT. On the value of response criteria in therapeutic research. Le Bulletin Du Cancer, Colloqulm INSERM-John Libbey Series, Proceedings of the 2nd International Congress on Neo-Adjuvant Chemotherapy 1988:75(9):863-869.

6. Holland JF. Induction chemotherapy: An old term for an old concept. Jn: Neoadjuvam chemotherapy. Colloque INSERM 1986;137:45-47.

7. Frei A Ill, Clark JR, Miller D. The concept of neoadjuvant chemotherapy. In: SaJmon· SE, ed. Adjuvant therapy of cancer V. Orlando: Grune & Stratton, 1987:67.

8. Muggia FM. Primary chemotherapy: Concepts and issues. In: Primary chemotherapy ln cancer medicine. New York: Alan R Liss, 1985:377-383.

9. De Vita VT. The relationship between tumor mass and resistance to treatment of cancer. Cancer 1983;51:1209-1220.

10. Skipper HE: Critical variables in the design of combination chemotherapy regimens to be used alone or in adjuvantsetting. Jn: Neoadjuvantchemotherapy. Colloque JNSERM 1986;137:11-12.

l 1. Goldie JH. Scientific basis for adjuvam and primary (neoadjuvant) chemotherapy. Semln Oncol 1987;14:1-7.

12. Bonadonna G. Veronesi U, Brambilla C, et al. Primary chemotherapy to avoid mastectomy in tumors with diameters of three centimeters or more. JNCI 1990;82:1539-154 l.

13: Jacobs C. Adjuvant and neoadjuvant treatment of head and neck cancers. Se min On col 1991;18:504-514.

14. Kelsen DP, Bains M, Bun M. Neoadjuvant chemotherapy and surgery of cancer of the esophagus. Semln Surg Oncol 1990;6(5):268-273.

15. Herr HN, Whitmore NF. Morse MJ et al. Neoadjuvant chemotherapy in invasive bladder cancer: The evolving role of surgery. J Urol 1990;144: 1083-1088.

16. Jacquillat C, Weil M, Baillet F, et al. Results ofneoadjuvant chemotherapy and radiation therapy in the breast conserving treatment of250 patients with all stages ofinfiltratiVe breast cancer. Cancer i 990;66:119-129.

17. DeVlta VT. Prlmary chemotherapy can avoid mastectomy but there is more to lt than that. JNCI [Editorial] 1990:82(19):1522-1523.

18. Dedrick RL, Myers CE, Bungay PM, et al. Pharmacokinetic rationale for peritoneal drug administration in treatment of ovarian cancer. Cancer Treat Rep 1978;62:1-I l.

19. Jones RB, Myers CE, Guarlno AM, et al. High volume intraperitoneal chemotherapy (''belly bath'') for ovarian cancer: Pharmacologic basis and early results. Cancer Chem other Pharmacol 1978;1:161-166.

20. Papahadjopoulos D, Poste G, Vail WJ, et al. Use oflipid vesicles as carriers to introduce actinomycin D into resistant tumor cells. Cancer Res 1976;36:2988-3012.

21. Weinstein JM, Magin RL, Cysyk RL, et al. Treatment of solid Ll2IO murine tumors with local hypenhermia and temperature-sensitive liposomes contai,ning methotrexate. Cancer Res 1980;40:1388-1396.

22. Weinstein JN. Llposomes as drug carriers in cancer therapy. Cancer Treat Rep 1984;68: 127-135.

23. Nathanson L, Hall TC, Schilling AC, et al. Concurrent combination chemotherapy of human solid tumors: Experience with three-drug regimen and review of the literature. Cancer Res 1969:29:419-425.

24. Potter VR. Sequential !?locking of metabolic pathways in vivo. Proc Soc Exp Biol Med 1951;76:41-46.

25. Elion GB, Singer S, Hitchings GH. Antagonists of nucleic acid derivatives: VIII. Synergism ln combinations ofbiochemically related antimetabolites. J Biol Chem 1954;208:477-488.

26. Sartorelli AC. Approaches to the· combination chemotherapy of transplantable neoplasms. Prog Ext Tumor Res l 965;6:228-288.

27. DeVita VT, Schein PS. The use of drugs In combination for the treatment of cancer: Rationale and results. N Eng[J Med 1973;288:998-1006.

28. Morstyn, Campbell L, Souza LM, et al. Effect of granulocyte colony stimulating factor on neutropena induced by cytotoxin chemotherapy. Lancet 1988;1:667-672.

29. Antman KS, Griffin JD, Elias A, et al. Effect of recombinant human granulocytemacrophage colony-stimulating factor on chemotherapy induced myclosuppression. N Engl J Med 1989;319:593-598.

30. Gabrilove J, Jakubowski A, Scher H, et al. Granulocyte colony-stimulating factor reduces neutropenia and associated morbidity of chemotherapy for transitional cell carcinoma of the urothelium. N Engl J Med 1988;318:1414-1422.

31. Ganser A, Volkers B, Ottman OG, et al. Recombinant human granu!ocyte-macrophage colony-stimulating factor in patients with myelodysplasia syndrome: Toxicity, pharmacokinetics and hematological effects. J Clin Oncol 1989;7:629-637.

32. Champlin RE, Nimer SD, Ireland P, et al. Treatment of refractory aplastic anemia with recombinant human granulocyte·macrophage-colony-stimulating factor. Blood 1989;73:694-699.

33. Luria SE. Delbruck M. Mutations of bacteria from virus sensitivity to virus resistance. Genetics I 943;28:491-51 l.

34. GoldieJH. Cold man AJ, A mathematic model for relating the drug sensitivity of tumors to the spontaneous mutation rate. Ganeer Treat Rep 1979;63: 1727-1733.

35. Yunis J. The chromosomal basis of human neoplasia. Science 1983;221:227-236. 36 . Day RS. Treatment sequencing, asymmetry and uncertainty: Protocol strategies for

combina1ion chemotherapy. Cancer Res 1986:46:3876-3885.

37. Norton L, Day R. Potential innovations in scheduling)n cancer chemotherapy. In: DeVita VT, Hellman S, Rosenberg SA, eds. Important advances in oncology 1991. Philadelphia: JB Llpplncott, 1991:57-73.

38. DeVita VT. The influence of information on drug resistance on protocol design: The

References 291

Harry Kaplan Memorial Lecture given at the Fourth International Conference on Malignant Lymphoma,June6-9, 1990, Lugano, Switze'rland. Ann Oncol 1991;2:93-106.

39. Buzzoni R, Bonadonna G, Valagussa P, et al. Adjuvant chemotherapy with doxorubicin plus cyclophosphamide, methotrexate, and flourouracll in the treatment ofresectable breast cancer wlth more than three positive axillary nodes. J Cl in Oncol 1991;9:2134-2140.

40. DeVita VT. Only if you believe in magic. In: Jones SE, Salmon SE, eds. Adjuvant therapy of cancer, 4th ed. Orlando: Grune & Stratton, 1984:3-16.

41. Skipper HE, Schabel FM Jr, Wilcox WS. Experimental evaluation of potential antlcancer agents: XII. On the criterla and kinetics associated with "curability" of experimental leukemia. Cancer Chemother Rep 1964;35: 1-11 l.

42. Skipper HE. Reasons for success and failure Jn treatment of murine leukemias with the drugs now employed in treating human leukemias. In: Cancer chemotherapy. Vol L Ann Arbor, Ml: University Microfilms International, 1978:1-166.

43. Skipper HE, Schabel FM Jr, Mellet LB, et al. Implications of biochemical, cytoklnetJc, pharmacologic, and toxicologic relationships in the design of optimal therapeutic schedules. Cancer Chemother Rep 1950:54:431-450.

44. Yankee RA, De Vita VT, PerryS. The cell cycle ofleukemiaLl210 cells in vivo. Cancer Res 1968;27:2381-2385.

45. Norton LA. A Gompenziail model of human breast cancer growth. Cancer Res 1988;48: 7067-7071.

46. Bloom H, Rlchardson M. Harris B. Natural history of untreated breast cancer (1804-1933): Comparison of treated and untreated cases according to hlstologic grade of malignancy. Med J 1962;2:213.

47. Schabel FM Jr. Simpson-Herren L. Some variables In experimental tumor systems which complicate interpretation of data from in vivo kinetic and pharmacologic studies with anticancer drugs. Antibiot Che mother 1978;23: 113-127.

48. Freirelch EJ. Gehan EA, Rall DP, ct al. Quantitative comparison of toxicity of anticancer agents in mouse, rat, dog, monkey and man. Cancer Chemother Rep 1966;50:219-244. -

49. Epstein RA. Drug-induced DNA damage and tumor chemosensitivity. J Clin Oncol 1990;8:2062-2084.

50. Brockman RW. Circumvention of resistance: Pharmaco!ogic basis of cancer chemotherapy. In: Proceedings of the 27th Annual Symposium on Fundamental Research. Baltimore: Williams & Wilkins, 1975:691-711.

SL DeVita VT Jr, Young RC, Canellos GP. Combination versus single agent chemotherapy: Review of the basis of selection of drug treatment of cancer. Cancer 1975;35:98.

52. Hutchinson OJ. Schmid FA. Cross-resistance and collateral sensitivity. In: Mihich E, ed. Drug resistance and selectivity: Biochemical and cellular basis. New York: Academic Press, 1973:73-126.

53. Hutchinson DJ. Cross-resistance and collateral sensitivity studies in cancer chemotherapy. In: Haddow A, Weinhouse S, eds. Advance6 in cancer research. Vol 7. New Yt;irk: Academic Press, 1963:235-350.

54. Klein M. A mechanism for the development of resistance to streptomycin and penicillin. J Bactcriol 1947;53:463-467.

55. Furth J, Kahn MC. The transmission of leukemia of mice with a single cell. Am J Cancer 1937;31:276-282.

56. Goldin A, Venditti JM, Humphries SR, et al. Influences of the concentration of leukemic inocu!um on the effectiveness of trcatmen1. Science 1956;123:840.

57. Ozols RJ. Cowan KH. New aspects of clinical drug resistance: The role of gene amp!lficatlon and the reversal of drug refractory cancer. In: DeVita VT Jr, Hellman S, Rosenberg SA, eds. Important advances in oncology 1986. Philadelphia: JB Lippincott, 1986:129-157.

58. Biedler JL, Riehm H. Cellular resistance to actinomycin Din Chinese hamster cells in vitro: Cross-resistance, radioautographlc and cytogenlc studies. Cancer Res 1970;30: ll74-ll84.

59. Juliano RL, Ung V. A surface glycoprotein modulating drug permeability In Chinese hamster ovary cell mutants. Biochem Biophys Acta 1976;455:152-162.

60. Bech-Hanson NT. Till JE, Ling V. Pleiotropic phenotype of colchiclne-resistant CHO cells: Cross-resistance and collateral sensitivity. J Cell Physlol 1976;88:23-32.

61. Ling V, Thompson LH. Reduced permeability in CHO cells as a mechanism of resistance to colch!cine. J Cell Physiol 1973;83: I 03-116.

62. Kartner N, Riordan JR, Ling V. Cell surface P-glycoprotein as associated with multidrug resistance in mammalian cell lines. Science 1983;221:1285-1288.

63. Fojo AT, Whang-Peng J, Gottesman MM, et al. Amplification of DNA sequences in human multidrug resistant KB carcinoma cells. Proc Natl Acad Sci USA 1985;82: 7661-7665.

64. Gros P, Croop J, Roninson I, et al. Isolation and characterization of DNA sequences ampllfied Jn multldrug-reslstant hamster cells. Proc Natl Acad Sci USA l 986;83:337-341.

65. Fairchild CR, Ivy SP, Kao-Shan CS, et al. Isolation of amplified DNA sequences associated with plciotroplc drug resistance from human breast cancer cells. Cancer Res 1987;47: 5141-5148.

66. Scotto KW, Biedler JL, Melera PW. Amplification and expression of genes associated with multidrug resistance in mammalian cells. Science 1986;232:751-755.

67. Roninson IB, Abelson HT, Housman DE, et al. Amplification of specific DNA sequepces correlates with multidrug resistance in Chinese hamster cells. Nature 1984;309:2070-2076.

68. Corwell MM, Safa AR, Felsted RL, et al. Membrane vesicles from multidrug-resistant human cancer cells contain a specific 150 to 170-kDa protein detected by photoaffinity labelling. Proc Natl Acad Sci USA 1986;83:3847-3850.

69. Safa AR, Glover Cl, Meyers CB, et al. Vinblastine photoaffinity labeling of a high molecular weight surface membrane glycoprotein specific for multidrug-resistant cells. J Biol Chem 1986;261:6137-6140.



70. Cornwell MM, Pastan I. Gottesman MM. Certain calcium channel blockers bind specifically to multidruk-resistant human KB carcinoma membrane vesicles and inhibit drug binding to P-g!ycoprotein. J Biol Chem 1987;262:2166-2170.

· 71. Gros P, CroopJ, Housman D. Mammalian multldrug-reslstance gene: Complete cDNA sequence indicates strong homology to bacterlal transport proteins. Cell 1986;47:371-374.

72. Chen C-J, Chin JE. Ueda K, et al. Internal duplication and homology to bacterial transport proteins in the mdrt (P-glycoprotein) gene from multidrug-resistant human cells. Cell 1986;47:381-389.

73. Gerlach JH, Endicott JA, Juranka PF, et al. Homology between P-glycoprotein and a bacterial haemolysin transport protein suggests a model for multidrug resistance. Nature 1986;324:485-489.

74. Past en JH. Gottesman MM. Molecular biology of multidrug resistance in human cells. In: DeVita VT, Hellman S, Rosenberg SA. eds. Important advances in oncology 1988. Philadelphia: JB Lippincott, 1988:3-16.

75. Fojo AT, Ueda K. Slamon DJ, et al. Expression of a multidrug-resistant gene in human tumors and tissues. Proc Natl Acad Sci USA 1987;84:265-269.

76. Chaudhary PM, Roninson IB. Expressions and activity of P-glycoproteln. a multidmg efHux pump, Jn human hematopoietic stem cells. Cell 1991:66:85-94.

77. Endicott JA, Juranka PF, Sarangi F, et al. Simultaneous expression of two P·glycoproteln genes in drug sensitive Chinese hamster ovary cells. Mol Cell Biol 1987;7:4075-4081.

78. Gros P. Ben Neriah Y, CfoopJM, et al. Isolation and expression of a complementary cDNA that confers multidrug resistance. Nature l 986;323:728-731.

79. Van Der Bliek AM, Baas F. Ten Houte de Lange T, et al. The human MDR3 gene encodes a novel P-glycoprotein homologue and gives rise to alternatively spliced mRNAs in liver. EMBO J 1987;6:3325-3331.

SO. Goldstein U. Glaskic H, Fojo A, et al. Expression of a multidrug resistant gene in human cancers. JNCI 1989;81:116-124.

81. Cordon-Cardo C, O'Brien JP, Boccln }, et al. Expression of the multidrug resistance gene product (P-glycoprotein) in human normal and tumor tissues. J Hlstochem Cytochem 1990;38:1277-1287.

82. Cordon-Cardo C, O'Brien JP, Casals D, et al. Multldrug-resistance gene (P-glycoprotein) Is expressed by endothelial cells at blood-brain barrier sites. Proc Natl Acad Sci USA 1989:86:695-698.

83. Weinstein RS, KusaacJR. Kluskons LF, et al. P-glycoprotein In pathology: The mu!tldrug resistance gene family in humans. Human Pathol 1990;21 :24-48.

84. Cordon-Cardo C, O'Brien JP. The multldrug resistance phenotype in human cancer. In: De Vita VT, Hellman S, Rosenberg SA. eds. Important advances in oncology 1991. Philadelphia: JB Lippincott, 1991 :19-39.

85. Marsh W, Center MS. Adriamycin resistance in HL60 cells and accompanying modifications of a i:nembrane protein contained in drug-sensitive cells. Cancer Res 1987;47: 5080-5086.

86. Danks MK, Yalowich JC, Bech WT. Atypical multiple drug resistance in a human leukemic cell line selected for resistance to teniposide (VM-26). Cancer Res 1987;47: 1297-1301. .

87. Ross WE. Sullivan DM. Chow KC. Altered function of DNA topolsomerases as a basis for antineoplastic drug action. In: DeVita VT, Hellman S, Rosenberg S, eds. Important advances in oncology 1988. Philadelphia: JB Lippincott, 1988:65-81.

88. Liu LF, D'Arpa P. Topoisomerase-targetlng anti tumor drugs: Mechanisms of.cytotoxicity and resistance. In: De Vita VT, Hellman S, Rosenberg SA. eds. Important advances in oncology 1992. Philadelphia: JB Lippincott, 1992:79-93.

89. Glisson B. Gupta R, Hodges P, et al. Cross resistance to intercalating agents in an eplpodophyllotoxin-resistant Chinese hamster ovary cell llne: Evldence for a common intracellular target. Cancer Res 1986;46: 1939-1942.

90. Pommier Y, Kerrigan D, Schwartz RE, et al. Altered DNA topoisomerase II activity in Chinese hamster cells resistant to topoisomerase II Inhibitors. Cancer Res 1986;46: 3075-3081.

91. Skipper H. Data and analyses having to do with the influence of dose Intensity and duration of treatment (single drugs and combinations) on lethal toxicity and the therapeutic response of experimental neoplasms. Birmingham: Southern Research Institute, Booklets'l3. 1986, and 2-13. 1987.

92. De Vita VT. Hubbard SM, Longo DL. The chemotherapy oflymphomas: Looking back, moving forward-The Richard and Linda Rosenthal Foundation Award Lecture. Cancer Res 1987;47:5810-5824.

93. Hryniuk WM. The importance of dose intensity in the outcome of chemotherapy. In: DeVlta VT, Hellman S, Rosenberg SA, eels. Important advances in oncology 1988. Philadelphia: JB Lippincott, 1988:121-142.

94. Hryniuk W, Levine MN. Analysis of dose intensity for adjuvant chemotherapy trials Jn stage JI breast cancer. J Clin Oncol 1986;4:1162-1170.

95. Hryniuk WM. Average relative dose intensity and the impact on design on clinical trials. Semin Oncol 1987:14:65-74.

96. Levin L. Hryniuk W. Dose intensity analysis of chemotherapy regimens in ovarian carcinoma. J Clin Oncol 1987;5:756-767.

97. Hryniuk W. Bush H. The importance of dose intensity in chemotherapy of metastatic breast cancer. J Clln Oncol 1984:2:1281-1288.

98. Hryniuk W. Is more better? J Clin Oncol [Editorial] 1986;4:621-622. 99. Canellos GP, Pocock SJ, Taylor SG Ill, eta!. Combination chemotherapy for metastatic

breast cancer: Prospective comparison of multiple drug therapy with L-phenylalanine mustard. Cancer 1976;38:1882-1886.

100. Canellas GP, DeVita VT. Gold GL, et al. Cyclical combination chemotherapy for advanced breast cancer. Br Med J 1974;1:218-220.

101. Bonadonna G. Brusamalino MP, Valagussa R, et al. Combination chemotherapy as an adjuvant treatment in operable breast cancer. N Engl J Med 1976;298:405-410.

102. Bonadonna G, Valagussa R. Dose-response effect of adjuvant chemotherapy in breast cancer. N Engl J Med 1981:304:10-15.

103. Horton J, Olson KB. Sullivan J, et al: 5-Fluorouracil In cancer: An improved regimen. Ann Intern Med 1970;78-.897-900.

l 04. Grage TB, Moss SE. Adjuvant chemotherapy in cancer of the colon and rectum: Demonstration of effectiveness of prolonged 5-FU chemotherapy in a prospectively controlled randomized trial. Surg Clin North Am 1981;61:1321-1329.

105. Higgins GA, Dwight RW. Smith JV. et al. Fluorouracil as an adjuvant to surgery in carcinoma of the colon. Arch Surg 1971;102:339-343.

106. Higgins GA Jr, Humphrey E. Juler GL, et al. Adjuvant chemotherapy in the surgical treatment of large bowel cancer. Cancer 1976;38:1461-1467.

107. Goldin A. Schepartz SA, Venditti JM. et al. Historical development and current strategy of the National Cancer Institute Drug Development Program. In: De Vita VT, Busch H, eds. Methods of cancer research. Vol XVI, Cancer drug development, part A. New York: Academic Press, 1979:165-247.

108. Zubrod CG, Schepartz S, Leiter J, et al. The chemotherapy program of the National Cancer Institute: History, analysts, and plans. Cancer Chemother Rep 1966;50:349-540.

109. Hirschberg E. Patterns of response of animal tumors to anticancer agents. Cancer Res 1963:23(suppl 5, part 2):521-980.

110. Driscol JS. The preclinical new drug research program of the National Cancer Institute. Cancer Treat Rep 1984;68:63-76.

111. Shoemaker RH, Wolpert·DeFilippes MK. Venditti JM. Potentials and drawbacks of the human tumor stem cell assay. Behring Inst Mitt 1984:74:262-272.

112. Rockwell S. Effects of clumps and clusters on survival measurements with clonogenic assays. Cancer Res 1985;45:1601-1607.

113. Shoemaker RH. Wolpert-DeFilippes MK, Kern DH, et al. AppliCation of a human tumor colony-forming assay to new drug screening. Cancer Res 1985;45:2145-2153.

114. Boyd M, Shoemaker R, Alley M. et al. New NCI disease-oriented drug screening program. In: Proceedings of the 5th NCl-EORTC Symposium on New Drugs in Cancer Therapy, Amsterdam, 1986.

115. Boyd MR, Shoemaker RH, McLemore TL, et al. New drug development. In: Roth JA, Ruckdescel JC. Weisenburger THE, eds. Thoracic oncology. Philadelphia: WB Saunders, 1987.

116. Mosmann T. Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxlcity assays. J Jmmunol Methods 1983;65:55-63.

117. Shoemaker RH. New approaches to antitumordrug screening: The human tumor colony fonningassay. Cancer Treat Rep 1986;70:9-12.

118. Rozencweig M. Von Hoff DD, Staquet MJ, et al. Predictive value of animal toxicology wlth anticancer agents prior to early clinical trials. Clln Res [Abstract] 1979;27C?):

391A. 119. DeV!ta VT, Oliverio VT, MuggiaFM, eta!. The drug development program and clinical

trials programs of the Division of Cancer Treatment, National Cancer Institute. Cancer Clin Trials 1979:2:195-216.

120. Muggia FM, Rozencweig M, Chiuten DF, el al. Phase II trials: Use of a clinical tumor panel and overvlew of current resources and studies. Cancer Treat Rep 1980;64: 1-9.

121. Wooley PV. Schein PS. Clinical pharmacology and phase I trial design. In: DeVita VT, Busch H, eds. Methods in cancer research. Vol XVII, Cancer drug development, part B. New York: Academic Press, 1979:177-199.

122. Muggia FM, McGuire WP, Rozencwelg M. Rationale.design and methodology of phase II clinical trials. In: DeVita VT, Busch H. eds. Methods in cancer research. Vol XVII, Cancer drug development, part B. New York: Academic Press. 1979: 199-215.

123. Collins JM. Zaharko DS, Dedrick RL, et al. Potential roles for preclinical pharmacology in phase I clinical trials. Cancer Treat Rep 1986:70:73.

124. Goldin A. Venditti JM. Progress report on the screening program at the Division of Cancer Treatment, National Cancer Institute. Cancer Treat Rev 1980;7: 167.

125. Marson! S, Hoth D. Simon R, et al. Clinical drug development: An analysis of phase II trials. 1970-1985. Cancer Treat Rep 1987;71:71.


r

83. Phillips GL, Fay JW, Herzig GP, et al. Intensive l,3·bis(2-chloroethyl)-1-nitrosourea (BCNU), NSC #4366650 and cyropreserved autologous marrow transplantation for refractory cancer: A phase I-JI study. Cancer 1983;52: 1792-1802.

84. Greene MH. Biole JD. Strike TA. Carmustine as a cause of acute nonlymphocytlc leukemia. N Engl J Med 1985:313:579.

85. Gerson SL, Miler K, Berger NA. 0 6alkylguanine-DNA alkyltransferase activity in human rnyeloid c~lls. J Clin Invest 1985;76:2106-2114.

86. O'Drlscol BR, Hasleton PS, Taylor PM, et al. Active Jung fibrosis up to 17 years after chemotherapy with cannustine (BCNU) in childhood. N Engl J Med 1990;232:378-382.

87. Hundley R. Lukens JN. Nitrosourea-associated pulmonary fibrosis. Cancer Treat Rep 1979;63:2128-2130.

88. Harmon WE. Cohen JH. Schneeberger EE, et al. Chronic renal failure in children treated with methyl CCNU. N Engl J Med 1979;300:1200-1203.

89. Cohen NA, Egorin MJ, Snyder SW, et al. Interaction of N,N',N'·triethylenethiophosphoramide and N,N',N'-triethylenephosphoramide with cellular DNA. Cancer Res 1991 ;51 :4360-4366.

90. HagenJ. Pharmacoklnetics ofthlo·TEPA and TEPA in the conventional dose· range and its correlation to myelosuppresslve effects. Ciincer Chemother Pharmacol 1991;27: 373-378.

SECTION 9 AnUcancer Drugs Derjved From Plants

ROSS C. DONEHOWER ERIC K. ROWINSKY

Many of the currently used antineoplastic agents are natural products initially isolated from plants. Although the search for new anticancer agents includes programs for the rational synthesis of compounds aimed at particular chemotherapeutic iargets, it seems likely that many important chemotherapeutic leads will continue to be identified by systematic screening of the natural products of plants, fungi, and marine animals. In this section of the chapter, the pharmacology of the vinca alkaloids, epipodophyllotoxins, and taxanes is reviewed.

VINCA ALKALOIDS

The vincas are natural alkaloids present in minute quantities in the common periwinkle plant (Catharanthus roseus). Although a number of these compounds have been investigated clinically, only vincristine and vinblastine have been approved currently for clinical use in the United States. The vincas are dimeric alkaloids formed from two different multiringed units (Fig. 18-32), an indole nucleus (catharanthine portion) and a dihydroindole nucleus (vindoline portion). Vincristine and vinblastine are structurally identical, with the exception of a single substitution on the vindoline nucleus, where vincristine possesses a formyl group and vinblastine a melhyl group. Despite this small difference, these two agents differ dramatically in their antitumor spectrum and clinical toxicitie.s. De~acetyl vinblastine (vindesine), initially identified as a metabolite of vinblastine, was introduced into clinical trials in the 1970s (see Fig. 18-32). Its developmental status remains uncertain, despite demonstrated antitumor activity in several malignancies, including non-small cell lung cancer. Two other semisynthetic derivatives of vinblastine have also entered clinical trial: vinzolidine and navelbine. These compounds are avail-

Vinca Alkaloids 409

91. Egorin MJ, Synder SW, Pan S-S. Daly C. Cellular transpon and accumulation of thiotepa in murine. human, and avian cells. Cancer Res 1989;49:5611-5617.

9~. Heideman RL, Cole DE. Balis F, et al. Phase I and pharmacokinetic evaluation ofthiotepa In the cerebrospinal fluid and plasma of pediatric patients: Evidence fordose·dependent plasma clearance of thiotepa. Cancer Res 1989;49:736-741.

93. Ng S-F. Waxman DJ. N,N',N'-triethylenethiophosphoramlde (thio-TEPA) oxygenation by constitutive hepatic P450 enzymes and modulation of drug metabolism and clearance in vivo by P450-induclng agents. Cancer Res 1991;51:2340-2345.

94. Egorin MJ. Snyder SW. Characterization of nonexchangeable radioactivity in Ll210 cells Incubated with [1 4CJthiotepa: Labeling of phosphatidylethanolamine. Cancer Res 1990;50:4044-4049.

95. Musser SM, Pan S·S, Callery PS. Liquid chromatography·thermospray mass spectrometry of DNA adducts formed With mitomycin C, porfiromycin and thlotepa. J Chromatogr 1989;474:197-207.

96. Lokich JJ, Egorin MJ, Cohen BE, et al. A phase I study of thiotepa administered by short-term and protracted continuous Intravenous Infusion. Cancer 1989;63:46-50.

97. Kinnan! S, McVey L, Loo D, Howell SB. A phase I clinical trial ofintraperltoneal thiotepa for refractory ovarlan cancer. Gynecol Oncol 1990;36:331-334.

98. Wolff SN. Herzig RH, Fay JW, et al. High-dose N,N',N'-triehylenethiophosphoramide (thiotepa) with autologous bone marrow transplantation: Phase I studies. Semin Oncol 1990;17:2-6.

able in oral and intravenous formulations in contrast wtth the other vincas, which can only be admi~~stered intravenously. There is investigative interest in navelbine (Fig. 18-33) because of the encouraging activity observed in early clinical trials. The developmental status of vinzolidine remains uncertain.

The vinca alkaloids have an unusually broad spectrum of antitumor activity. Vincristine is an essential part of combi·nation chemotherapy regimens for pediatric and adult acute lymphocytic leukemia. 1 Vincristine has also played an irnportant role in the combination chemotherapy of Hodgkin's and non-Hodgkin's lymphoma? Vincristine is commonly used in combination with other agents in the treatment of Wilms' tumor, Ewing's sarcoma, neuroblastoma, and rhabdomyosarcoma in children and multiple myeloma, breast cancer, and small cell lung cancer in adults. 3

Vinblastine has been an integral component of curative chemotherapy regimens for germ cell cancers of the testis and advanced Hodgkin's disease and is commonly used for several other diseases.3

Vindesine has been most thoroughly evaluated in non-small

OCOCH,

OH COOCH,

VINBLASTINE R =CH, VINCRIST!NE R =CHO

FIGURE 18-32. Structure ofvinblastine and vincristine.


410 Anticancer Drugs

FIGURE 18-33. Structure of navelbine.

cell lung cancer in which the cumulatiVe response rates from all studies approach those reported for the other vincas.3

Navelbine is currently undergoing extensive clinical testing in the United States in the hope that the encouraging results observed in early clinical trials in breast cancer and'lung cancer in Europe will be sustained.4

MECHANISM OF ACTION

Microtubules are ubiquitous in eukaryotic cells and are vital to the performance of many critical cellular functions, such as the maintenance of cell shape and intracellular transport, in addition to mitosis. Many of the unique functional aspects of microtubules are the r~sult of their existence in a dynamic equilibrium with tubulin dimers and their ability to polymerize and depolymeriz~ in response to critical physiologic messages in the cell, including those related to cell cycle traverse. Vinca alkaloids exert their cytotoxic effects by binding to a specific site on tubulin and preventing polymerization of tubulin dimers, disrupting the formation of microtubules. This binding occurs at sites that are distinct from the binding sides of other antimicrotubule agents, such as colchicine, podophyllotoxin, and taxol.5 The binding ofvincas to tubulin prevents the polymerization of the tubulin subunits into microtubules; microtub.ule structures eventually disappear with continued drug exposure. Although the primary cytotoxic effect of vinca alkaloids is putatively by mitotic arrest, there is evidence that the lethal effects may be attributable in part to other actions. They appear to induce major cytolytic effects in vivo on nonproliferating cells in the GI phase of the cell cycle, and some cells appear to be most sensitive to the cytotoxic effects of these agents when exposed in the GI or the S phase of the cell cycle.' Low concentrations of vincristine (0.1 µmo!) are capable of blocking formation of microtubules, a fact that could be explained by the hypothesis that the entire microtubule assembly process can be inhibited by the incorporation of a single drug-tubulin complex at the growing end of the microtubule. 6 Even lower concentrations may result in cytotoxicity for some cell lines in vitro, but duration of exposure appears to be an important factor, in addition ·to drug concentration.7

PHARMA CO KINETICS

Considering the broad range of clinical uses of vinca alkaloids, relatively limited pharmacologic data are available, principally

because of a lack of a suitably sensitive assay to quantitate the extremely low concentrations of drug that possess pharmacologic activity. Early studies were generally performed with radiolabeled compounds, with subsequent chromatographic separation of parent drug and metabolites. However, recently developed radioimmunoassays (RJAs) and enzymelinked immunosorbent assays (ELISAs) with sensitivities in the picomolar range have the requisite sensitivity, perhaps at the expense of specificity.

After conventional doses of vincristine (1.4 mg/m2) given as a brief infusion or bolus, peak plasma concentrations approach 0.4 µmo!. 8 The drug is extensively bound by serum proteins (48%) and the formed elements of blood, particularly platelets. After an intravenous bolus, the plasma disposition of vincristine is triphasic. Initial half-lives (t~-a) in adults are less than 5 minutes because of extensive tissue uptake and binding.'·" /I-Phase half-lives range from 50 to 155 minutes, and the terminal 'Y half-lives vary widely from 23 to 85 hours. Similar pharmacokinetic parameters are noted in children. 11

When pharmacokinetics have been studied using [3H]vincristine with purification by high-performance liquid chromatography, the mean terminal half-life has been 64 minutes; "and /I half-lives have been 0.85 and 7.4 minutes, respectively.8 A direct linear relation between the maximum clinically tolerated dose of the vinca alkaloids, vincristine, vinblastine, and vindesine and their serum clearance has also been noted. 8 ·

10 Vincristine has been demonstrated to enter the central nervous system of subhuman primates rapidly after intravenous injection, and concentrations above 1 nmol have been maintained in cerebrospinal fluid for longer than 72 hours after 8.n intravenous bolus. 12 However, other studies with rats, dogs, monkeys, and humans have indicated that vincristine penetrates the human blood-brain barrier poorly. 13

Spinal fluid concentrations have been 20- to 30-fold lower than concurrent plasma levels and never exceeded 1.1 nmol in humans. 13

Vincristine is metabolized primarily by the liver and excreted in the stool and, to a lesser extent, in urine. Seventytwo hours after the administration ofradiolabeled vincristine, 12% of the radiolabel is excreted in urine (50% of which consists of metabolites) and approximately 70% is excreted in feces ( 40% of which consists of metabolites). 9 Vincristine concentrates rapidly in the bile with an initial bile to plasma vincristine concentration ratio of 100 to 1 that declines to 20 to 1 at 72 hours after a bolus dose. 14 In addition, inetabolic products appear rapidly in the bile with only 46.5% present as the parent unmetabolized drug at 2 hours after infusion. Although not carefully studied, modifications in the doses of vincristine should be considered for patients with hepatic dysfunction, especially patients with obstructive liver disease.

The clinical pharmacology of vinblastine is similar to that of vincristine with extensive binding to serum proteins and the formed elements of blood. 15 After intravenous bolus injections of vinblastine, peak plasma concentrations reach 0.5 µmol.9

•10

·16 As with vincristine, the plasma disposition ofvin

blastine is triphasic. The initial distribution phase is rapid with a t~-a of less than 5 minutes, presumably secondary to extensive binding ta tissues and the formed elements of blood. Vinblastine appears to be more avidly sequestered in tissues than vincristine, with 73% of labeled drug retained in the body 3 days after drug administration. Values reported for /I

~f '


and 'Y half-lives have ranged from 53 to 99 minutes and 20 to 24 hours, respectively.9

•10

•16 Excretion of vinblastine is

principally through the biliary tract, which ag:lln suggests that toxicity may be exaggerated in patients with hepatic insufficiency. Less than 15% of an administered dose is excreted in the urine, and fecal excretion of the .parent compound is also low, suggesting extensive metabolism. The metabolic fate of vinblastine has not been fully characterized, but one metabolite, vindesine, appears to be as active as the parent compound. When vinblastine is administered by continuous intravenous infusion for 5 days at dosages from 1 to 2 mg/m2

/

day, steady-state drug concentrations of 1.1 to 6.6 nmol have been observed. 17

Peak plasma vindesine concentrations of 0.1 to 1.0 µmol are reached after bolus intravenous injections of this agent, and levels typically decrease to less than 0.1 µmo! by 1 to 2 hours. As with the other vincas, the plasma disappearance of vindesine is characterized by triexponential pharmacokinetics. Vindesine is rapidly distributed to body tissues, and its ty,,-a is typically less than 5 minutes. (3 and 'Y half-lives range from 55 to 100 minutes and 20 to 24 hours, respectively. This prolonged eliminatiofl suggests that drug accunlulatioil may occur with repeated doses at short intervals. 9 •

10•18 Vindesine has a

large volume of distribution (600 L for the terminal phase) and negligible renal clearance, with 1 % to 12% of the parent compound measured in the urine of patients on bolus or infusion schedules. 19 The liver is the principal organ responsible for the disposition of vindesine as determined in animal studies and in humans. 19•

20 In a patieilt with a biliary T-tube, the measured renal clearance ofvindesine was low (12 ml/min), but the biliary clearance was 29 ml/min, and biliary concentrations were significantly higher than simultaneously measured plasma levels.

The pharmacokinetic behavior of navelbine is similar to that of other vinca alkaloids, with drug disposition exhibiting biexponential Or triexponential behavior. After intravenous administration, there is an initial rapid distribution phase (5-9 minutes) and a long elimination half-life of 20 to 49 hours. 21

•22 Like the other vincas, the principal route of excre

tion is through the biliary tract (70-80%), and there is limited urinary excretion of the drug. Principal metabolites appear to be deacetylated forms, although other unidentified metabolites are also present in urine. Human studies of oral navelbine administered in a powder-filled capsule have shown the bioav:lllability of the parent compound to be approximately 40%, with peak plasma levels achieved I to 2 hours after an oral dose. 23 Similar studies of a liquid-filled gelatin capsule demonstrated 23% to 30% bioavailability.24

DRUG INTERACTIONS

Methotrexate accumulation in tumor cells is enhanced in vitro in the presence of vincristine or vinblastine,25 an effect mediated by a vinca-induced blockade of drug efflux from cells. The minimal concentrations of vincristfue required to achieve this effect occur only transiently in vivo, and experimental chemotherapy studies in mice have not demonstrated superiority of this sequence of administration. Pharmacokinetic interactions of vinca alkaloids and other drugs have not been studied in detail.

Vinca Alkaloids 411

TOXICITY

Despite similarities in structure, the vinca alkaloids differ significantly in their clinical toxicity profiles. Peripheral neurotoxiclty is the most frequent and dose-limiting toxic effect of vincristine. 26 It frequently occurs in patients who are older than 40 years and is related to cumulative dose. Initially, only symmetric sensory impairment and paresthesias may be encountered. However, neuritic pain and motor dysfunction may then develop, and loss of deep tendon reflexes, foot drop, wrist drop, ataxia, and paralysis can occur with continued administration. Patients may also complain of bone, back, and limb pain. Vincristine-induced peripheral neuropathic effects are usually first noted in adults who. have received a total dose of 5 to 20 mg. Children tolerate this toxicity better than adults, and the elderly are particularly prone. Patients with lymphomas appear to be more highly predisposed to developing neurotoxicity when compared with patients with other malignancies, and particular care is warranted in treating pati.ents with antecedent neurologic disorders, such as Charcot-Marie Tooth disease, or a history of poliomyelitis. The principal pathologic finding is primary axonal degeneration. Peripheral nerve conduction velocities are usually normal, although diminished amplitude of sensory and motor nerve action potentials and prolonged distal motor latencies are observed.27 Motor, sensory, cranial, and autonomic nerves may all be affected by vincristine, resulting in hoarseness, diplopia, j~w pain, facial palsies, paralytic ileus, urinary retention, and postural hy-

. potension. Fortunately, these complications are uncommon. Convulsions and altered mental status have also been reported.3 Currently, the only known treatment for vincristine neurotoxicity is discontinuation of the drug or reduction of the dose or frequency of administration. A number of antidotes have been proposed, including thiamine, vitamin B12, folinic acid, and pyridoxine, but none has been shown to be effective. However, coadministration of glutamic acid to patients receiving vincristine appears to decrease neurotoxicity in a randomized, prospective, double-blind trial. 28 The mechanism by which glutamic acid exerts this protective effect and whether this effect will be confirmed in subsequent trials are not known. Although neurotoxicity also occurs during therapy with vindesine, it is dose limiting in less than 1 Oo/o of patients. Significant neurotoxicity occurs infrequently at the usual clinical doses of vinblastine. Mild paresthesias have been reported in as many as 30% cif patients in P.hase II trials of navelbine.

Myelosuppression, in particular neutropenia, is the principal dose-limiting toxicity of vinblastine, vindesine, and navelbine. Thrombocytopenia and anemia are typically less.common and less severe. The timing of the onset of neutropenia is usually 7 to 11 days after an intravenous bolus of one of these agents with recovery by day 14 to 21. Hematologic toxicity has not been cumulative. Gastrointestinal side effects have been noted with all the vinca alkaloids. Mucositis, pharyngitis, and stomatitis are commoner with vinblastine than with vincristine, but may occur with any of these agents. Other gastrointestinal effects (which are probably the result of autonomic neuropathy), such as constipation, ileus, bloating, and abdominal pain, are seen most commonly with vincristine or high doses of vinblastine or vindesine. Nausea, vomiting, and diarrhea may also occur. All the vinca alkaloids are potent vesicants



and may cause significant tissue damage if extravasation occurs. Phlebitis may occur along the course of an injected vein with resultant sclerosis. Mild and reversible alopecia is a common finding, seen in as many as 20% of patients treated with vincristine. A variety of other side effects have been reported, including the syndrome of inappropriate antidiuretic hormone secretion, acute cardiac ischemia, fever without an obvious source, acute pulmonary edema, Raynaud's phenomenon, and liver damage.3

RESISTANCE

Resistance to the vinca alkaloids develops rapidly in vitro and has been ascribed to two mechanisms. The first involves mutations of the a or {J subunits of the tubulin heterodimer leading to decreased vinca alkaloid binding.29 The second, more wellcharacterized mechanism involves the multidrug resistance phenotype that confers broad resistance to many unrelated classes of bulky natural-product antineoplastic agents. There has been considerable interest in using pharmacologic agents in vivo to reverse multidrug resistance. Responses have been seen in multiple myeloma and lymphoma using verapamil in combination with a vinca alkaloid-containing regimen after the identical regimen was found to be inactive without verapamil. 29a

EPIPODOPHYLLOTOXINS

Extracts of the mandrake plant (Podophyllum peltatum) have been used for medicinal purposes for centuries as cathartics or as treatment for parasites or ven~real warts. Podophyllotoxin, an antimitotic agent that binds to a site on tubulin distinct from that occupied by the vinca alkaloids, was identified as the main constituent possessing cytostatic activity as early as the 1940s. These early podophyllotoxin derivatives possessed a prohibitive degree of clinical toxicity. However, two glycosidic derivatives of podophyllotoxin, etoposide (VP-16) and teniposide (VM-26). have demonstrated highly significant clinical activity against a wide"variety of neoplasms, including non-Hodgkin's lymphomas, germ cell malignancies, leukemias, and small cell lung carcinoma. 30 These compounds have a complex structure consisting of a multiringed moiety, known as an epipodophyllotoxin, linked to a sugar, glucopyranoside, by an ether linkage. The structures of etoposide and teniposide differ only by the substitution of a methyl group (etoposide) for the thenylidine (teniposide) on the glucopyranoside sugar (Fig. 18-34). These agents possess a number of basic similarities in their pharmacologic characteristics, toxicities, and spectrum of antineoplastic action.

Teniposide is not approved for general use in the United States, although it is widely available in other countries. Extensive clinical trials have been performed with this drug, and it is clear that teniposide is an active antitumor agent, particularly in childhood monocytic leukemia. A major problem has been an assessment of what degree of activity teniposide possesses relative to etoposide.

H

R~O

o~ 0 HO

OH 0

<o 0

0 -,(

I 0

,,,J:)_o'", OH

R=CH,- VP-16

R= 0- VM-26 . s

FIGURE 18-34. StructureofVP-16 and VM-26.

MECHANISM OF ACTION

The epipodophyllotoxins were initially thought to exert their cytotoxic effects by binding to tubulin and inhibiting microtubule assembly in a manner similar to the parent compound, podophyllotoxin. However, it was noted that these agents had no effect on microtubule assembly at concentrations th.at were highly cytotoxic.31 By cell cycle analysis, the epidophyllotoxins were found to arrest cells in late-S or early-G2 phase of the cell cycle, rather than the G2/M border that would have been expected of an antimicrotubule agent. 32 It was subsequently found that these drugs produced DNA strand breaks in intact cells but that these effects were not see11 when the epipodophyllotoxins were incubated in vitro with purified DNA, suggesting that direct chemical cleavage in DNA was not occurring. 33 The epipodophyllotoxins most likely exert their cytotoxic effects by interfering with the scission-reunion reaction of the enzyme topoisomerase II by stabilizing the putative cleavable enzyme-DNA complex in a cleavable state34

(see Fig. 18-18), The enzyme then covalently binds to DNA, forming single-strand, protein-associated breaks. On a molar basis, teniposide is approximately tenfold more effective than etoposide at inducing RNA strand breaks.35 Besides forming a cleavable complex, the epipodophyllotoxins inhibit the catalytic or "strand-passing" activity of topoisomerase II that permits the enzyme to catenate DNA circles and disentangle topologically constrained DNA. .

Etoposide is also able to inhibit transport of nucleos1des across the plasma membrane in a manner similar to podophyllotoxin, an action that may influence the efficacy of combinations of etoposide with antimetabolites.31

PHARMACOKINETICS

In adults with normal hepatic and renal function, .the disposition of etoposide in plasma is best des·cribed as a biphasic


process with .an initial half-life of approximately 1.5 hours and a terminal half-life that ranges from 3 to 11 hours.36 Peak plasma levels and areas under the plasma concentration versus time curve (AUC) increase linearly with doses over a 100 to 600 mg/m2 dose range.37 Mean peak plasma concentrations after a single intravenous dose of 290 mg/m2 are approximately 30 µg/ml. 37

•39 Etoposide does not accumulate in plasma

after daily administratio~ of 100 mg/m2 ·for 5 consecutive days. Jn addition, plasma and renal clearances are independent of dose. Estimates of plasma clearance range from 16 ± 7 to 28 ± 9. 7 ml/min/m2

, whereas the uriri.ary clearance of etoposide ranges from 7.4 to 13.6 ml/minim' and accounts for approximately 25% to 48% of total disposition. Using a twocompartment open model, the volume of distribution for etoposide at steady state ranges from 7 to 29 L/m2

•36 After ad

ministration of radiolabeled etoposide to humans, 41.9% to 87.5% of the total radioactivity is recovered in the urine within 48 hours, with 66.8% in the form ofunmetabolized drug. After an 80 mg/m2 intravenous bolus, the mean urinary excretion over 48 hours is 39.8%, and fecal recovery ranges from 0% to.16% of administered radioactivity. 39 Only 6% or less of an intravenous dose is recovered in the bile as the parent compound. Therefore, biliary excretion appears to be a minor route of drug elimination, and metab.olism probably accounts for most of the nonrenal clearance of etoposide in humans.

In children, approximately 55% of a dose of etoposide is excreted in the urine over 24 hours, and renal clearance accounts for approximately 35% of total body clearance over a dose range of 80 to 600 mg/m2

• 40 In one study of patients

with hepatic and renal dysfunction, stepwise multiple linear regression analysis of patient-specific variables identified

· cfeatinine clearance and serum albumin as the best predictors of systemic clearance of etoposide, whereas hepatic functiori appeared to be an insignificant factor. 41 The excess risk of hematologic toxicity in patients with abnormal renal function can be managed by dose modification based on serum creatinine42 or by adaptive control dosing based on measured plasma concentrations. 43 For patients with hepatic dysfunction, clearance rate, volume of distribution, and half-lives of etoposide are not significantly prolonged in patients with obstructive jaundice when compared with patients with normal hepatic function. Such patients appear to have a compensatory increase in renal excretion, and dose modification may therefore not be necessary. 44 However, total plasma etoposide concentrations may not accurately reflect the magnitude of systemic exposure to active drug because of the increased fraction of unbound drug in patients with liver dysfunction due to displacement by bilirubin. Caution is therefore justified in dosing such patients. A mathematical model to estimate the fraction of unbound etoposide has been developed that will require prospective pharmacodynamic validation but may prove useful in this situation.45 Etoposide is metabolized by opening of its lactone ring, yielding its principal urinary metabolite, the hydroxy acid. 41 A cis-lactone is also identifiable in plasma and urine.41 In addition, glucuronide or sulfate conjugates, or both, are excreted in urine in several species and represent 5o/o to 22% of the total dose in humans.46 These metabolites do not .have significant antineoplastic activity. After intravenous or oral administration, peak plasma etoposide concentrations and AUC values exhibit marked intrapatient and interpatient variability that confounds reliable estimates of absolute oral

Epipodophyllotoxins 413

bioavailability.36.47 Bioavailability appears to be approximately 50% using the currently available formulation, although the range of value noted in indiv:idual patients varies widely (25-75%).

Preclinical and several well-conducted clinical trials have shown that the cytotoxic effects of etoposide are highly schedule dependent. In a study that examined the pharmacokinetic basis for this observation, a dramatic difference in response rates (10% versus 89%) was observed in. untreated patients with extensive small cell lung cancer randomized to receive the same etoposide dose as a 24-hour infusion or as five consecutive daily 2:-hour infusions. Pharmacokinetic analysis of the study arms revealed that levels of 1 µg/ml were maintained for twice as long a period in the 5-day arm. 48

Instantaneous peak plasma concentrations of teniposide at a dose of 67 mg/m2 range from approximately 20 to 30 µg/ ml. 36•39 The plasma disposition of teniposide has been characterized as a biphasic process in some studies and as a triphasic process in others.36·

49 Values reported for the elimination half-life . of teniposide vary according to the pharmacokinetic model used to fit the data: 6 to I 0 hours in · studies demonstrating biexponential decal and between 20 and 48 hours in those demoµstrating triphasic characteristics. The pharmacokinetic behavior of teniposide does not differ significantly between children and adults.36•46·50 Voluffie of distribution at steady state ranges from 8 to 30 Lim' in adults and 3 to 10 Lim' in children. 36 The volume of distribution at steady state has been calculated to be approximately 28.5% of body weight with a mean central compartment of 3.5 L. 50

Total plasma clearance has been 7 to 17 ml/min/m2, and the

renal clearance of the drug has been reported to be less than that of etoposide, 1 to 3 ml/minim'. Urinary excretion of the parent compound comprises 4o/o to 14% of the administered dose. 49

Teniposide is more extensively metabolized than etoposide. Only 5% to 20% of the administered dose of teniposide is excreted as unchanged drug. Information pertaining to the metabolism of teniposide is scant, and only the hydroxy acid formed by opening of the lactone ring, the picro-lactone iso-. mer, and the aglycone glucuronide metabolites have been identified in plasma and urine.36

•46

•49 Only the aglycone has been demonstrated to possess anti-DNA activity. Etoposide has a threefold increased plasma clearance and sixfold increased renal clearance over teniposide in humans. Teniposide· has a more complete metabolism and a lower renal clearance than etoposide, which account for its sloWer elimination. In addition, another factor that influences the slower elimination rate of teniposide may be its higher degree of protein binding (>99% for teniposide versus 94% for etoposide).

DRUG INTERACTIONS

Agents with diverse mechanisms of action, such as cyclophosphamide, carmustine, cisplatin, vincristine, cytosine arabinoside, 5-fluorouracil, and hydroxyurea, have demonstrated synergistic interaction with epipodophyllotoxins. 51 The epipodophyllotoxins also appear to be synergistic with methotrexate and enhance net cellular accumulation of methotrexate and methotrexate polyglutamates in vitro at clinically relevant concentrations.52



TOXICITY

The principal dose-limiting toxicity of the epipodophyllotoxins has been a dose-related myelosuppression. Granulocytopenia predominates, with nadir counts occurring 7 to 14 days after initiation of a course of therapy. Nadir granulocyte counts on the daily oral schedule of etoposide occur somewhat later, days 17 to 21. Thrombocytopenia occurs less frequently, and nadir counts are observed 9 to 16 days after etoposide. Hematologic recovery is usually complete by day 20 after standard doses, and cumulative toxicity is not usually observed. Patients who have received extensive prior myelosuppressive therapy by chemotherapy or irradiation to marrow-bearing areas of the skeleton are at increased risk for developing severe myelosuppression and should receive reduced doses. Nausea and vomiting occur in approximately 30o/o to 40o/o of patients and are more frequent with oral than intravenous administration. Discontinuation of treatment because of adverse gastrointestinal effects is rarely required. Constipation, diarrhea, dysphagia, aftertaste, abdominal pain, stomatitis, and anorexia have· been reported, but are infrequent at conventional doses. Mucositis is a prominent toxicity at high doses and appears to be the dose-limiting nonhematOlogic toxicity of etoposide; severe hepatotoxicity also has been associated with high-dose etoposide. 53

•54

Etoposide induces transient hypotension after rapid intravenous administration that usually responds to slowing or cessation of the infusion and adi:ninistration of fluids or other appropriate supportive therapies. Anaphylactic-like symptoms, characterized by chills, fever, tachycardia, bronchospasm, and dyspnea, may be associated with hypotension and occur in approximately 0. 70/o to 2% of patients. These reactions are usually observed during or immediately after administration of etoposide and often respond promptly to the cessation of the infusion and administration of'pressor agents, corticosteroids, antihistamines, and volume expanders. Similar acute hypersensitivity reactions (e.g., hypotension, flushing, bronchospasm) may develop if teniposide is infused rapidly. Teniposide is formulated in the same cremophorvehicle as taxol, and these reactions are believed to result from to the effects of the excipient.55 Reversible alopecia, sometimes progressing to total baldness, occurs in at least 8% to 20% of patients. Increased pigmentation, pruritis, and, rarely, radiation recall dermatitis have been observed. 56 Therapy with epipodophyllotoxin has also been associated with subsequent development of acute nonlymphocytic leukemias.57 Minor toxicities have included chemical phlebitis and local irritation with extravasation.

RESISTANCE

Malignant cells develop resistance to epipodophyllotoxins through several mechanisms, and in most instances this resistance gene.rally includes etoposide and teniposide, regardless of which agent has been used for selection. In some cells, resistance is associated with decreased intracellular concentrations of etoposide and teniposide because of amplification of the MDR gene and increased cellular levels of the P-170 glycoprotein, the multidrug resistance efflux pump.58•58

a A second mechanism of resistance is an enhanced ability to repair DNA strand breaks. Resistance to etoposide can be in-

duced by prior treatment with radiation, and inhibition of DNA repair can block the development of etoposide resistance.59

•60

Altered topoisomerase activity because of low cellular enzyme concentrations or structural alterations in the prot€in that may affect binding has also been identified as a cause of resistance in cells that demonstrate cross-resistance to other topoisomerase-11 inhibitors, such as amsacrine and doxorubicin. aoa.6ob The predominant clinical mechanism of etoposide resistance has not been established.

TAXOL

The clinical development of taxol began in 1983 and for the first few years proceeded slowly. However, the unique chemical structure of this agent and its unique mechanism of action coupled with the significant antitumor activity observed in early clinical trials has made taxol perhaps one of the most important new agents to emerge in cancer therapy in the last decade. Interest in taxol began in the 1960s when a crude extract of bark from the Pacific yew, Taxus brevifolia, was shown to have broad antitumor activity in preclinical experimental tumor models in screening performed by the National Cancer Institute. In 1971, Wall and coworkers identified taxol as the active constituent of the bark extract. soc Taxol is composed of a complex 15-membered taxane ring system and an ester side chain attached at the C-13 position of the ring, which is essential for antitumor activity (see Fig. 18-3). The development of taxol was initially hampered by the difficulties inherent in large-scale isolation, extraction, and preparation of bulk compound for a natural product and the poor aqueous solubility of taxol. Interest in the drug was maintained during this time by the elucidation of the novel mechanism of action by which taxol exerts its cytotoxic effect and the ultimate availability of adequate drug supply for the necessary preclinical testing.61

·62 Should the clinical development oftaxol con

tinue to be successful, the issue of drug supply will become increasingly important. The current method of obtaining taxol depends on a process of extraction from tree bark. Because the tree is found principally in ancient forests of the Pacific northwest, this is an environmentally important issue. A renewable source of drug supply will be needed for the drug to be available for broad clinical use, and a number of options are under active exploration, including total synthesis or semisynthesis from a precursor. A taxol precursor, 10-deacetylbaccatin Ill, has been isolated from the needles of other yew species, and this compound can be converted by semisynthesis to taxol or analogs by addition of the appropriate side chains. One of these analogs, taxotere, has also been selected for clinical development on the basis of its antitumor activity and improved solubility.63

The most impressive clinical antitumor activity of taxol has been observed in advanced ovarian cancer and metastatic breast cancer. The initial activity reported in refractory ovar- · ian cancer64 has been confirmed in three subsequent studies, with response rates ranging from 21% to 40%.65- 67 A phase Ill trial in the Gynecologic Oncology Group should help clarify the role of taxol in advanced ovarian cancer. Significant activity (56-62%) has also been observed in metastatic breast cancer patients. 68 Preliminary results from· phase I studies of taxotere suggest that this agent also has activity in cancers of the ovary and breast. 69


MECHANISM OF ACTION

The laboratory of Horwitz and coworkers initially demonstrated that taxol promotes microtubule assembly in vitro and that taxol stabilizes microtubules in mouse fibroblast cells exposed to the drug.61·62 Taxol binds preferentially to microtubules rather than to tubulin dimers, with a binding constant of approximately 1 µmo!. 10 Although the binding site for taxol on microtubules is distinct from the binding sites for exchangeable guanosine triphosphate (GTP) and for colchicine, podophyllotoxin, and vinblastine, the specific binding site for taxol on microtubules has not been identified. _Unlike other antimicrotubule agents, such as colchicine and the vinca alkaloids, which induce microtubule disassembly, taxol shifts the equilibrium toward microtubule assembly and stabilizes microtubules at concentrations as low as 0.05 µmol/L, which can be easily achieved in patients. Overall, taxol decreases the critical concentration of tubulin required for microtubule assembly in the presence or absence of factors that are usually essential for assembly, such as exogenous GTP or microtubuleassociated proteins.71 Taxol-treated microtubules are stable even after treatment with calcium or low temperatures, conditions that usually promote disassembly. 72 This unusual stability inhibits the normal dynamic reorganization of the microtubule network. Concentrations of taxol (0.1-10 µmo!) produce two distinct morphologic effects on cellular microtubules. During all phases of the cell cycle, cells form abundant arrays of disorganized microtubules aligned in parallel bundles.62·73-75 Additionally, although taxol-treated cells show evidence of entry into mitosis as manifested by chromosomal condensation and breakdown of the nuclear membrane, they lack normal mitotic spindle apparatus. Instead of forming two mitotic spindle asters enucleated by bipolar centrioles, large numbers of abnormal asters that do not require centrioles for enucleation are formed.13- 75 These distinct morphologic effects suggest that taxol may adversely affect critical microtubule functions during interphase and mitosis, but the precise rea.sons for cell death are unclear.

PHARMA CO KINETICS

The pharmacology of taxol has been studied in a number of phase I trials, and the results reported have been consistent. 76

Peak plasma concentrations achieved with brief infusions (6 hours) range from 1.3 to 13.0 µmo! at the recommended dose of 175 to 250 mg/m2.77

•78 On the more conventional 24-hour

infusion schedule, steady-state plasma· concentrations of 0.6 to 3.5 µmo! have been maintained. 79 The disposition of taxol in plasma has been best characterized by a biexponential model. a and {3 half-lives have ranged from 0.27 to 0.31 hours (mean 0.29) and 1.3 to 8.6 hours (mean 5.0), respectively. Mean values for steady-state volumes of distribution have been large (55-183 L/m2; mean, 110 L/m2>". Systemic clearances have ranged from 100 to 993 ml/min/m2, but the principal mechanisms of systemic clearances have not been fully elucidated. When cisplatin administration precedes taxol in combination chemotherapy, taxol clearance has been demonstrated to decrease by approximately 30% and plasma concentrations to increase accordingly. 80 Total urinary excretion has been insignificant (range, 1.4-6.6%; mean,' 4.8%). The importance of hepatic metabolism and biliary excretion in

Mechanism of Action 415

taxol disposition has been emphasized by a recent study in rats demonstrating that 12% and 29% of an injected dose are recovered in bile as unchanged taxol and metabolites, respe.ctively. 81 The principal metabolites are derivatives hydroxylated at the m position of the benzoate at C-2 or the phenyl group at C-3' of the ester side chain, both of which have significantly diminished cytotoxic activity. Preliminary data suggest that human metabolism is qualitatively similar. 82

DRUG INTERACTIONS

Interactions of other drugs with taxol have not yet been studied in detail at the level of plasma pharmacokinetics or cellular effects. Administration of cisplatin before a taxol infusion de·creases the plasma clearance of taxol, significantly increasing Systemic exposure. The resultant myelosuppression has been shown to be greater than for the reverse sequence. In vitro cytotoxicity of the combination of these two agents appears to be greater for the sequence of taxol followed by cisplatin rather than the reverse sequence or simultaneous exposure. 83

TOXICITY

The principal dose-l~miting toxicity of taxol is neutropenia.76

The neutropenia is dose dependent and is not influenced by the schedule of administration, nor is it cumulative. Onset of neutropenia has usually occurred by day 8 after treatment, nadir neutrophil counts are observed by day 8 to 11, and recovery occurs by day 15 to 18. Episodes of fever during neutropenia have not been frequent, perhaps because of the short duration of taxol-induced myelosuppression. The major risk factor for neutropenia appears to be the extent of prior myelotoxic chemotherapy or irradiation. Anemia and thrombocytopenia have been uncommon.

A major concern during early clinical studies of taxol was the frequent occurrence of hypersensitivity reactions84 consisting of hypotension, bronchospasm, dyspnea, urticaria, abdominal and extremity pain, and diaphoresis. It has been assumed that these reactions represent a nonimmunologically mediated release of histamine and other vasoactive substances induced by the drug or, more likely, its cremophor vehicle. With prolonged infusions (24 hours) and premedicati9n with corticosteroid, diphenhydramine, and H2 antagonists, the incidence has decreased to less than 1 Oo/o of infusions. An important study ongoing in Canada and Europe will address the question of whether the cumbersome 24-hour infusion is necessary to prevent this toxicity when premedication is given.

Clinically significant neurologic toxicity has been uncommon when taxol is given as a single agent at doses less than 170 mg/ml, although mild neurosensory changes have been frequent with numbness and paresthesias occurring in the characteristic ''glove and stocking'' distribution. Neurologic examinations reveal distal sensory loss and loss of deep tendon reflexes. Electrophysiologic studies support axonal degeneration and demyelination as mechanisms for taxol neurotoxicity .76·80·85 Transient myalgias and arthralgias, which are occasionally painful, have been observed in patients treated at higher doses. When taxoI and cisplatin are given in combination chemotherapy, neurotoxicity becomes more prominent and is dose limiting when G-CSF is used to limit the neutropenia. 80

•86



Other toxicities that have been observed with trucol include nausea and vomiting, mucositis, total body alopecia, and rare local venous toxicity. Cardiac effects include bradycardia and rare instances of atrioventricular block, ventricular tachycardia, and myocardial infarction. 87

RESISTANCE

Investigations with cell lines made resistant to taxol have characterized two potential mechanisms of acquired drug resistance. Some resistant cells have mutations in tubulin, resulting in impaired microtubule assembly. Continuous exposure to taxol is required for polymerizatioh to proceed normally,. thereby promoting the formation of functional microtubules.

A second well-documented mechanism of acquired trucol resistance involves the MDR phenotype that confers crossresistance to vincristine, doxorubicin, and etopbside.88 The multidrug resistance is more fuily discussed in the section on drug resistance in this chapter.

REFERENCES 1. Aur IiJA, Simone JV, Verzosa MS, et al. Childhood acute lymphocytlc leukemia: Study

VJll. Cancer.1978;42:2123-2134. 2. DeV!Ca VT, Serpick AA, Carbone PP. Combination chemotherapy in the treatment of

advanced Hodgkin's disease. Ann Intern Med 1970;73:891-895. 3. Rowlnsky EK, Donehower RC. The clinical pharmacology and use of antimlcrotubule

agents in cancer chemotherapeutics. Pharmacol Ther 1991;52:35-84. 4. Depierre A, Lemarie E, Dabouis G, et al. A phase II study ofnavelbine In the treatment

of non-small cell lung cancer. Am J Clin Oncol 1991;14:115-119. 5. Bryan J. Definition of three classes of binding sites in isolated mlcrotubule crystais.

Biochemistry 1972; 11 :2611-2616. 6. Madoc-Jones H, Mauro F. lnterphase action ofvinblastine and vincristine: Differences

In their lethal action through the mitotic cycle of cultured mamalian cells. J Cell Physlol 1968;72:185-196.

7. Jackson DV, Bender RA. Cytotoxic thresholds ofvincristine in a murine and human leukemia cell line In vitro. Cancer Res 1979;39:4346-4349.

8. Bender RA, Castle MC, Margileth DA, et al. The pharmacokinetics of j3HJ-vlncristlne in man. Clin Pharmacol Ther 1977;22:430-438.

9. Nelson RL. The comp~tlve clinical pharmacology and ph3nnacokinetics ofvindesine, vtncristlne, and vinblastine in human patients with cancer. Med Pediatr Oncol 1982;10: 115-127.

IO. Nelson RL, Dyke RW. Root MA. Comparative pharmacokineticsofvlndesine, vincristine, and vinblastine in patients with cahcer. Cancer Treat Rev l980;7(Suppl): 17-24.

11. Sethi VS, Kimball JC. Pharmacoklnetlcs of vlncrlstlne sulfate In children. Cancer Chemother Pharmacol 1981;6:111-115.

12. El Dareer SM, White VM, Chen FP, et al. Distribution and metabolism of vincristin~ in mice, rats, dogs and monkeys. Cancer Treat Rep 1977;61:1269-1277.

13. Jackson DV, Sethi VS, Spurr CL. et al. Pharinacoklnetics of vincristine ln the cere· brospinal fluid of humans. Cancer Res 1981;41:1466-1468.

14. Jackson DV, Castle MC, Bender RA. Biliary excretion of vincristine. Clin Pharmacol Ther 1978;24:101-107.

15. Owellen RJ. Hartke CA, Hains FO. Pharmacokinetlcs and metabolism ofvinblastine , in humans. Cancer Res 1977;37:2597-2602.

16. Creasey WA, Scott Al, Wei CC, et al. Pharmacological studies with vlnblastine in the dog. Cancer Res 1975;35:1116-1120.

17. Lu K, Yap HY, Wetts S, et al. Comparative clinical pharmacology of vinblastine in patients with advanced breast cancer. Proc Am Soc Clin Oncol 1979;20:371.

18. Hande K. Gay J, Gover J, et al. Toxicity and pharmacology of bolus vlndesine injection and prolonged vlndes!ne infusion. Cancer Treat Rev 1980;7:25-30.

19. Jackson DV, Sethi VS, Long TR. et a1. Phannacoklnetics of vindesine by bolus and infusion. Cancer Chemother Pharmacol 1984;13:114-119.

20. Rahmani R, Zhou XJ, Placidi M, et al. In vivo and in vitro pharmacokinetlcs and metabollsm of vinca alkaloids in rat: I. Vindesine. Eur J Drug Metab Pharmacokinet 1990;15:49-55.

21. Jehl F, Quoix EM, Leveque D, et al. Pharmacokinetic and preliminary metabolic fate of navelbine in humans as determined by high performance liquid chromatography. Cancer Res 1991:51:2073-2076.

22. Rahmani R, Bruno R, Iliad ls A, et al. Clinical pharmacokinetics of the antltumordrug, navelbine. Cancer Res 1987;47:5796-5799.

23. Annand JP, Marty M. Navelbine: A new step in cancer therapy. Semln Oncol 1989;16(Suppl.4):41-45.

24. Lucas VS, Donehower RC. Rowlnsky EK, et al. Clinical results of the absolute bio·

availability and pharmacokinetics of weekly navelbine liquid-filled soft gelatin capsules at full therapeutic doses in patients with solid tumors. Proc Am Soc Clin Oncol 1992; 11: 111.

25. Bender RA, Bleyer WA, Frisby SA, et al. Alteration of methotrexate uptake in human leukemia cells by other agents. Cancer Res 1975;35:1305-1308.

26. Legha SS. Vincristlne neurotoxlclty: Pathophyslology and management. Med ToXicol 1986;L421-427.

27. Bradley WG,· Lassman LP, Pearce GW. et al. The neuropathy of vincristine in man: Clinical. electrophysloiogical and pathological studies. J Neurol Sci 1970;10:107-131.

28. Jackson DV, Wells HB, Atkins JN, et al. Amelioration of vlncristine neurotoxiclty by glutamic acid. Am J Med 1988;84:1016-1022.

29. Cabral FR, Brady RC, Schiber MJ. A mechanism of cellular resi~tance to drugs that Interfere with mlcrotubule assembly. Ann NY Acad Sci Hl86;46:748-756.

29a. Meltzer PS, Dalton WS. Grogan TM, et al. Drug-resistance in multiple myeloma and non-Hodgkin's lymphoma: Detection of P-glycoprotein and potential circumvention by addition of verapamil to chemotherapy. J Clin Oncol l 989;7:415-424.

30. O'Dwyer PJ. Leyland-Jones B, Alonso MT, et al. Etoposlde (VP-16-213): Current status of an active ~ticancer drug. N EnglJ Med 1985;312:692-700.

31. Loike D, Horwitz SB. Effects of podophyllotoxin and VP-16 on micro tubule assembly in, vitro and nucleoslde transport In HeLa cells. Biochemistry 1976;15:5435-5442.

32. Krishan A. Palka K, Frei E. Cytofluoroipetrlc studies on the action of podophyllotoxins and eplpodophyllotoxins (VM-26, VP-16-213) on the cell' cycle traverse of hyman lymphoblasts. J Cell Biol 1975;66:521-530.

33. Wozniak AJ, Ross WE. DNA damage as a basis for 4'demethylepippdophyll 9-(4, 6-0· ethyiidlne-P·D·glucopyranoside) (etoposide) cytotoxicity. Cancer Res 1983;43:120-124.

34. Yang L, Rowe TC, Liu LF. Identification of DNA topolsomerase II as an intracellular target of antltumor eplopodophyllotoxlns in Simian Virus 40-infected monkey cells .. Cancer Res 1985;45:5872-5876.

35. Long BH, Brattain MG. The activity ofetoposlde (VP16·213) and tenlposide (VM-26) against human lung tumor cells In vitro: Cytotoxicity and DNA breakage. In: lssell BF, Muggia FM, Carter SK, eds. Etoposide (VPl 6): Current status and new developments. Academic Press, 1984:63-86.

36. Clarke Pl, Slevin ML. The clln!Cal pharmacology of etoposide and teniposlde. Clin Pharm 1987;12:223-252.

37. Hande KR, Wedlund PJ, Noone RM. et al. Pharmacokinetlcs of high-dose etoposlde (VP-16-213) administered to cancer patients. Cancer Res 1984;44:379-382.

38. D'lncalci M, FarlnaP, SessaC, eta!. Pharmacokinetlcs ofVP-16-213 given by different administration schedules. Cancer Chemother Pharmacol 1982:7:141-145.

39. Creavcn PJ. The clinical pharmacology of VM-26 and VP-16-213: A brief overview. Cancer Chemother Pharmacol 1982:7:133-140.

40. Sinkule JA. Hutson P, J1ayes FA, et al. Pharmacoklnetlcs of etoposide in children and adolescents with refractory solid tumors. Cancer Res 1984;44:3109-3113.

41. Arbuck SG, Douglas HO, Crom WR, et al. Etoposlde pharmacokinetlcs In patients with normal and abnormal organ functions. J Clln Oncol 1986;4:1690-1695.

42. Simon J, Clark D, Slevin M. Renal function and etoposide pharmacokinetics: Is dose modification necessary? Proc Am Soc Clin Oncol 1991 ;10:281.

43. Mick R, Rataln M. Modeling interpatlent pharmacodynamic variability of etoposide. Proc Am Soc Clin Oncol 1991;10:274.

44. Hande KR, Wolff SN, Greco FA. et al. Etoposlde kinetics In patients with obstructive jaundice. J Clin Oncol 1990;8:1101-1107.

45. Stewart CF, Arbuck SG, Fleming RA, et al. Changes In the clearance of total and unbound etoposlde in patlents with liver dysfunction. J Clin Oncol 1990;8: 1874-187.9.

46. Evans WE, Slnkule JA, Crom WR, et al. Pharmacokinetlcs ofteniposlde (VM-26) and etoposlde (VP-16-213) in children with cancer. Cancer Chemother Pharmacol 1982;7: 147-152.

47. Harvey VJ, Stein ML, Smithe, MM. et al Variable bioavallability following repeated oral doses of etoposide. Eur J Cancer Clin Oncol 1985;21:1315-1319.

48. Slevin ML, Clark PL. Joel SP, et al. A randomized trial to evaluate the effects of schedule on the activity of etoposide in small-cell lung cancer. J Clin Oncol 1989;7: 1333-1340.

49. O'Dwyer PJ, Alonso MT, Leyland-Jones B, et al. Teniposlde: A ri:!vlew of 12 years' experience. Cancer Treat Rep 1984;68:1455-1464.

50. Crom WR, Glynn-Barnhart AM, Rodman JH, et al. Pharmacoklnedcs of anticancer drugs In children. Clin Pharm 1987:12:168-213.

51. Ratain MJ. Schilsky RL, Wojack BR, et al. Hydroxyurea and etopos!dei Ii:i vitro synergy and phase I clinical trial. JNCI 1988;6:1412-1416.

52. YalowichJC, Fry OW, Goldman ID. Tenlposlde (VM-26) and etoposide (VP-16-213)· induced augmentation of methotrexate transport and polyglutamatlon In Ehrlich ascites tumor cells in vitro. Cancer Res 1982;42:3648-3653.

53. Wolfe SN, Fer M, McKay CM. High-dose VP-16-213 and autologous bone marrow transplantation for refractory malignancies: A phase I study. J Clin Oncol 1983;1 :701-705. .

54. Johnson DH, Greco FA, Wolfe SN. Etoposide-lnduced hepatic injury: A potendal com· plication of high-dose therapy. Cancer Treat Rep 1983;67:1!;)23-1024.

55. O'Dwyer PJ, King SA, Fortner CL, Leyland-Jones B. Hypersensitivity reactions to teniposlde (VM-26): An analysis. J Clin Oncol 1986;4:1262-1269.

56. Fontana JA. Radiation recall associated with VP-16-213 therapy. Cancer Treat Rep 1979:63:224-225.

57. Winick N, McKenna R, Bowman WP. et al. Secondary acute myeloid leukemia in children with B-lineage acute lymphoblastlc leukemia treated with an eplpodophyl· lotoxln. Proc Am Soc Clin Oncol 1992;11:279.

58. Lee T, Roberts D. Flux oftenlposide (VM-26) across the plasma membrane ofteniposlde resistant sub lines of L1210 cells. Cancer Res 1984;44:2986-2990.


58a. Gupta RS. Genetic, biochemical and cross-resistance studies with mu tan cs of Chinese hamster ovary cells resistant to the anticancer drugs VM-26 and VP-16-213. Cancer Res 1983:43:1568-1574.

59. Hill BT, Bellamy AS. Establishment of an etoposide-resistant human epithelial tumor cell line in vitro: Characterlzation of patterns of cross-resistance anddrug sensitivities. Int J Cancer 1984;33:599-608.

60. Arnold A. Whitehouse J. Interaction of VP-16·213 with the DNA repair antagonist chloroquine. Cancer Chemother Phannacol 1984;7:123-126.

60a. Glisson B. Characterization of acquired epipodophyllotoxin-reslstant Chinese hamster ovary cell line: Loss of drug-stimulated DNA cleavage activity. Cancer Res 1986;46: 1934-1938.

60b. Pommier Y, Kerrigan D, Schwartz RE, et al. Altered DNA topoisomerase II activity in Chinese hams1er cellS resistant to topoisomerase II inhibitors. Cancer Res 1986;46: 3075-3081.

60c: Wani MC, Taylor HL, Wall ME. Coggon P, McPhail AT. Plant antltumor agents: VI. The isolation and structure of taxol, a novel antileukemic and antltumor agent from Taxus brevifolia. J Am Chem Soc 1971;93:2325-2327.

61. Schiff PB. Fant J. Horwitz SB. Promotion of mlcrotubule assembly in vitro by taxol. Nature 1979;22:665-667.

62. Schiff PH, Horwitz SB. Taxol stabilizes mlcrotubules in mouse fibroblast cells. Proc Natl Acad Sci USA H!80;77:1561-1563.

63. Bissery MC, Guenard D. Gueritte-Voegelein F, et al. Experimental antltumor activity of taxotere (RP 56976, NSC 628503), a taxol analogue. Cancer Res 1991 ;51 :4845-4852.

64. McGuire WP, Rowinsky EK, Rosenshein NB, et al. Taxol: A unique antineoplastic agent wlth significant activity in advanced ovarian epithelial neoplasma. Ann Intern Med 1989;111:273-279.

65. Thigpen T, Blessing J. Ball H, et al. Phase II trial of taxol as a second-line therapy for ovarian carcinoma: A gynecologic oncology group study. Proc Am Soc Clin Oncol 1990:9:604.

66. EinzigAI, Wiernik P. Sasloff J, et al. Phase II study oftaxol in patients with advanced ovarian cancer. Proc Am Assoc Cancer Res 1990;31:1l14.

67. Sarosy G, Kohn E, Link C, et al. Taxol dose intensification in patients with recurrent ovarian cancer. Proc Am Soc Oncol 1992;1 l :226.

68. Holmes FA, Walters RS. Thleriault RL, et al. Phase II trial oftaxol: An active drug in metastatic breast cancer. JNCI 1991;83:1797-1805.

69. Pazdur R, Newman RA, Newman BM, et al. Phase I trial.oftaxotere (RP56976) Proc Am Soc Clin Oncol l992;ll:lll.

70. Parness J, Horwitz SB. Taxol binds to polymerized microtubules in vitro. J Cell Biol 1981;91:479-487.

71. Hamel E, de] Campo M, Lowe MC, et al. Interactions oftaxol, mlcrotubule-associated

References 417

proteins and guanine nucleotides in tubulin polymerization. J Biol Chem 1981;256: 11187-11894.

72. Thompson WC, Wilson L, Purlch DL. Taxol induces microtubule assembly at low ·.temperature. Cell Motll 1981:1:445-454.

73. Rowlnsky EK, Donehower RC, Jones RJ, et al. Microtubule changes and cytotoxicity in leukemic cell lines treated with taxol. Cancer Res 1988;48:4093-4100.

74. RobensJR, Rowinsky EK, Donehowerfl.C, et al. Demonstration of the cell cycle positions for taxol-induced "asters" and "bundles" by measurement of f\ourescence, FeulgenDNA content, and auto radiographic labeling of the same cells. J Hlstochem Cytochem 1989:37:1659-1665.

75. Roberts JR, Allison DC, Dooley WC, et al. Effects oftaxol on cell cycle traverse: Taxol· induced polyploidization as a marker for drug resistance. Cancer Res 1990;50:710-716.

76. Rowlnsky EK. Cazenave LA, Donehower RC. Taxol: A novel lnvestigational antimicrotubule agent. JNCI 1990;82:1247-1259.

77. µ,ngnecker SM. Do~ehower RC, Cates AE, et al. High-performance liquid chromatographic assay for taxol in human plasma and in the pharmacokinetics in a phase I trial. Cancer Treat Rep 1986;71:53-59.

78'. Wiernik PH, Schwartz EL. Strauman JJ, et al. Phase I clinical and pharmacoklnetic study of taxol. Cancer Res 1987;47:2486-2493.

79. Rowlnsky EK, Burke PJ, Karp JE, et al. Phase I and phannacodynamic study of taxol in refractory acute leukemias. Cancer Res 1989;49:4640-4647.

80. Rowlnsky EK, Gilbert M. McGuire WP, et al. Sequences oftaxol and clsplatln: A phase I and pharmacologic study. J Clin Oncol 1991;9:1692-1703.

BI. Monsarrat B, Mariel E, Cros S, et al. Taxol metabolism, isolation and identification of three major metabolites of taxol In rat blle. Qrog Metab Disp 1990;18:895-901.

82. Monsarrat B, Alvinerie P. Gares M, et al. Hepatic metabolism and bllary excretion of taxol. J ~ell Pharmacol 1992 (in. press).

83. Citardl M. Rowinsky EK, Schaefer KL, et al. Sequence-dependentcytotox1city between cisplat!n and the antimicrotubule agents taxol and vincristine. Proc Am Assoc Cancer Res 1990;31:2431.

84. Weiss R. Donehower RC, Wiemlk PH, et al. Hypersensitivity reac1Jons from taxol. J Clln Oncol 1990;8:1263-1268.

85. Lipton RB, Apfel 'sc, Dutcher JP, et al. Taxol produces a predominantly sensory neuropathy. Neurology 1989:39:368-373.

86. Forastiere AA. Rowinsky EK. Chaudry V, et al. Phase I trial of taxol and clsplatin plus G-CSF In solid tumors. Proc Am Soc Clln Oncol 1992;11:117.

87. Rowinsky EK, McGuire WP, Guarnieri T, et al. Cardiac disturbances during the ad-ministration of taxol. J Clin Oncol 1991;9:1704-1712. ·

88. Greenberger LM, WJlliams SS, Horwitz SB. Biosynthesis of heterogeneous forms of multid.rug resistance-associated glycoproteins. J Biol Chem 1987;262:1-5 •.


CHAPTER 40

Cancer; Principles & Practice of Oncology, Fourth Edition, edited by Vincent T. DeVita, Jr., Samuel Hellman, Steven A. Rosenberg. J.B. Llppjncott Co., Philadelphia© 1993.

Jay R. Harris Monica Morrow

Gianni Bonadonna

Cancer of the Breast

In North America, breast cancer is the inost common malignancy among women and accounts for 27o/o of their cancers. Eighteen percent of the female cancer deaths are due to breast cancer, but since 1985, Jung cancer has equaled or exceeded breast cancer as a cause of cancer death in women. Among women in the United States between the ages of 40 and 55 years, breast cancer is the leading cause of death. Jn 1992, approximately 181,000 new cases of breast cancer will be diagnosed in the United States, and 46,000 women will die from breast cancer.' Between 1940 and 1982, the agestandardized incidence rose by an average of 1.2% per year, and between 1982 and 1986, the incidence rose more sharply at 4% per year. Age-adjusted breast cancer mortality rates have been remarkably stable in the United States. The relatively constant mortality despite increases in incidence may be the result of increases in a more benign form of the disease, earlier detection, or advances in treatment.

The natural history of breast cancer is characterized by a Jong duration and marked heterogeneity within and among patients. Breast cancer is among the more slow-growing tu- · mors, and the preclinical period before diagnosis and the clinical phases after initial treatment and even after the appearance of metastasis are measured in years and decades. Nevertheless, some patients have aggressive forms of the disease and do poorly. Other patients have such indolent forms of the disease that it is difficult to demonstrate that therapy has any effect on survival. During the long clinical phase, there is ample opportunity for clonal mutation and evolution, and it seems probable that individual patients may have multiple tumor clones, each with its own growth rate, propensity to metastasize, and sensitivity to drugs. Advances in molecular biology during the next decade should enable a more precise estimate of a p&tient's clinical course than is now possible based on clinical criteria.

1264

In this chapter, we describe the salient features of the disease, stressing practical information of importance to clinicians and findings that are new since the J;µ;t edition of this book. Readers interested in a more comprehensive source should consult another book in this series.2

ONCOGENES AND GROWTH FACTORS IN BREAST CANCER

Advances in molecular biology have enabled researchers to examine the mechanisms of neoplastic growth at the level of gene expression. Delineation of the genetic, molecular, and biochemical events that ~e responsible for the initiation and transformation of mammary epithelial cells are crucial for understanding the cause, pathogenesis, and prognosis of this disease and for developing new methods of treatment and prevention.a-a Two impOrtant examples are presented here.

It is hoped that identification of oncogenes in breast tumors can help in assessing prognosis. Some of the oncogenes linked to human mammary carcinoma are ERBB2 (i.e., human homolog of the gene originally called NEU in rat neuroblastomas), HRAS (Ha-ras), MYC, and WNT2 (i.e., updated name of INT2), of which ERBB2 has been the rnost studied. The ERBB2 (or NEU) protein is a member of the tyrosine kinase family, and because of the similarity of its transmembrane topology to the epidermal growth factor receptor, it is thought to be a growth factor receptor, although the putative ligand has not been identified. In more than 1500 human breast carcinomas, amplificatioil of the ERBB2 oncogene or its overexpression at the level of mRNA transcription were found in 21 % of the tumor specimens, showing increased gene copy numberranging from twofold to more than a 100-fold greater than normal.9 Jn 1987, Slamon and colleagues published find-


ings of a strong association between the oncogene's amplification and nodal status but not with.other previously defined prognostic factors, including progesterone receptor status, tumor size, or patient age. For the patients with histologically positive lymph nodes, a significant correlation was found between the degree of gene amplification and both the time to relapse and survival. These findings have been confirmed by some but not all investigators. 10

•11 Some of the conflicting

findings may be a result of small sample sizes, nonconsecutive case series, differences in study design, or the use of different techniques to demonstrate ERBB2 gene amplification. Only prospective studies involving large numbers of patients can clarify its usefulness and that of other oncogenes in assessing prognosis. ·

An improved understanding of the biology of breast cancer may provide new avenues of therapy. During the past 2 decades, a new model of the growth regulation of human breast cancer was developed (Fig. 40-1). 12

•15 Breast tumors have

the capacity to make growth factors that stimulate their own growth in an autocrine fashion, such as the transforming growth factor-a (TGF-a) and insulin growth factors (IGFs), ot that affect stromal tissues or tumor invasiveness by paracrine mechanisms, such as TGF-il and platelet-derived growth factor.3•13•

14 The synthesis and secretion of several of these growth factors are regulated by estrogen in some estrogen receptor (ER)-positive breast cancer cell lines, and they are produced constitutively by certain ER-negative cell lines. 12·13

This new model of breast cancer growth regulation proposes that stromal tissues such as fibroblasts, mononuclear cells, or endothelial cells play an important role in the development and growth of mammary carcinoma by producing diffusible products that enhance the growth of the breast cancer cells. There appear to be important communications in the form of secreted polypeptide messengers among the various cellular elements of a breast tumor. 14 This new model of breast cancer regtilatory mechanisms provides exciting opportunities for new therapies to inhibit tumor growth by interfering specifically with the actions of these growth factors. Such therapies are currently being investigated in phase 1-11 studies.

SCREENING FOR BREAST CANCER

Screening for cancer represents an important advance in the management of the _disease and is described in detail elsewhere in this book. The main methods for earlier detection of breast cancer have been mammography and physical examination performed by a trained health professional. The ability of mammography to detect cancers well before they are apparent on physical examination bas been established. The usefulness of mammography has been enhanced by technical improvements that provide better visualization of the breast parenchyma with less exposure to radiation, improvements in film quality and processing, improved techniques for imaging (e.g., compression, focal spot size reduction, magnification, ancillary use of ultrasonography), better guidelines for the diagnosis of cancer, and greater availability of well-trained mammographers. Using these newer techniques, a large percentage of cancers detected :ire 2 cm or smaller with uninvolved axillary nodes or at the noninvasive

Screening for Breast Cancer 1265

Prolact;n"' ··::

?;

A

B

..... E ---1'"'~:\~,c:::,_: .. ~· -+'~•

~::::::" . :·.. '::' ::-:·i.':

Breast Cancer Cell

IGF-I I :I __ ,,...-1GF-ll

\.i......_ """- --EGF

,-- ,, 1-.,,...-=---t-Q Growth Factors

Other Organs

l Estromedins.,/

Secreted Growth Factors A toe ine (TGFa, TGFp, IGF-I, IGF-IJ, u r 53K, POGF) ----

Breast cancer Cell

Paracrine

Strama! cell FIGURE 40-1. (A) Simplified mechanisms of breast cancer growth regulation based on data from the early 1970s. Prolactin was considered important, mainly because it has a pivotal role in rodent mammary carcinoma. The importance of this hormone in human breast cancer development and growth remains questionable. Estrogen (E) stimulation of cells involves the binding of E to estrogen receptor (ER) protein, followed by the tight coupling of the receptor-hormone complex to DNA, resulting in alterations of specific gene transcription and protein synthesis. (B) Schema depicting possible growth regulatory pathways in human breast cancer based on recent experimental data. The model preserves the estrogen response pathway mediated through cellular ER, and there is considerable evidence to suggest that E can directly affect cell proliferation. E may also indirectly influence tumor growth through the secretion of growth factors (estromedins) produced at distant sites. In addition to steroid hormones, a host of polypeptide hormones and growth factors may play a role in the regulation of breast cancer growth. (With permission from Osborne CK, Artega CL. Autocrine and paracrine growth regulation in breast cancer: Clinical implications. Breast Cancer Res Treat 1990; 15:3-11)

stage (i.e., ductal carcinoma in situ). The use of randomized trials to assess the effect of screening on breast cancer mortality eliminates potential sources of bias, and the trials have demonstrated that screening reduces breast cancer mortality by approximately 25%. "·22

·

The finding that screening reduces breast cancer mortality has important implications for the natural history of the disease. Some evidence suggests that metastases occur very early in the course of the disease and that breast cancer should be considered a systemic disease from its onset. 23 However, the reduction in breast cancer mortality by screening provides


1266 Cancer of the Breast

compelling evidence that early diagnosis and treatment of breast cancer can avert metastasis. Breast cancer therefore should not be considered a systemic disease in all patients.

Although the randomized trials provided an overall estimate of the effect of screening, they leave several issues unresolved. One important issue is the effect of screening within different age groups. Jn the available studies, the beneficial effects of screening are fairly consistently restricted to women between the ages of 50 and 69. There is no firm evidence that screening reduces breast cancer mortality in women 40 to 49 years of age, although it can be argued that only in recent years has the sensitivity of mammography been adequate to detect small lesions reliably in this age group. The available results from trials provide little information on the value of mammography in women 70 years or older and have not defined the optimal periodicity of screening or the relative effects of mammography and physical examination in reducing breast cancer mortality.

RISK FACTORS

The risk of an American woman developing breast cancer during her lifetime is currently about 11 %, and 3% to 4% will die from the disease. Several factors increase the risk of developing breast cancer, including family history, reproductive history, diet, hormone usage, and radiation exposure. Despite the recognition of these risk factors, approximately 70% of the women who develop breast carcinoma do not have any identifiable risk factors.

Numerous studies have linked breast cancer incidence to the age of menarche, menopause, and first pregnancy. Age at menarche and the establishment of regular ovulatory cycles seem to be strongly associated with breast cancer risk. 24

•25 A

review of case-control studies suggests that a 200/o decrease in breast cancer risk exists for each year that i.nenarche is delayed.24 The late onset of menarche is associated with a delay in the establishment of regular ovulatory cycles, which is thought to have an additional protective effect by some investigators, although there is dispute on ·this point. 25-27 A woman's level of physical activity, even if moderate, can have an impact on the likelihood of ovulatory cycles and may alter breast cancer risk. 28 The age-specific incidence of breast cancer rises at a steep rate with age up to the time of menopause and then slows to a rate one sixth of that seen in the premenopausal period. It has been suggested that this age-specific incidence curve is shaped largely by the effects of ovarian activity. 29

Age at menopause is another factor in breast cancer risk. The relative risk of developing breast cancer for a woman with natural menopause before age 45 is 0. 73 compared with a woman with natui-al menopause between the ages of 45 and 54.30 Oophorectomy before age 50 decreases breast cancer risk, with an increasing magnitude of risk reduction as the age at oophorectomy decreases.31 From these data, it seems likely that the total duration of menstrual life is an important factor in breast cancer risk, although the mechanisms through which risk is altered remain uncertain.

Parity and age at first birth are other endogenous hormonal factors that influence breast cancer risk. Nulliparous women are at greater risk for the development of breast cancer than ·

parous women, with a relative risk of about 1.4.32 It has become increasingly apparent that the effect of term pregnancy on breast cancer risk varies with the age at first birth, with women whose first term pregnancy occurs after age 30 having a twofold to fivefold increase in breast cancer risk compared with women having a first term pregnancy before age 18 or 19.33•34

Abortion, whether spontaneous or induced, before full-term pregnancy has no protective effect, and in several studies, it has been shown to increase breast cancer risk. The apparently contradictory effects of pregnancy on risk are explained in a variety of ways. 32·35·36 Breast tissue may undergo differentiation as a result of the hormonal changes of pregnancy, and these differentiated cells are less likely to undergo malignant transformation, or the persisting changes in hormone levels after pregnancy may alter the proliferative rate of the breast epithelium.29·37·38 In incomplete pregnancy, the breast is exposed only to the high estrogen levels of early pregnancy, and this may be responsible for the increased risk seen in these women.

The associations between the use of oral contraceptives or postmenopausal estrogen replacement and breast cancer were reviewed by Henderson.39 Of 27 reported studies of oral contraceptive use, only two showed an overall increase in risk for all patients in the study population. One of these studies found a twofold increase in breast cancer risk with any use of oral contraceptives. 40 These data suggest that if oral contraceptives increase overall breast cancer risk, the magnitude of the increase is small. Despite this, there is continuing concern about the potential risks in a variety of subgroups. Several studies have demonstrated an increased risk of breast cancer in long-terin users of oral contraceptives.40

-43 Risk estimates

range from 1. 7 for use longer than 8 years to 4.1 for use for 10 or more years.40

•41 However, the Centers for Disease Con

trol's Cancer and Steroid Hormone study failed to find any increase in risk related to duration of oral contraceptive use or in users with a family history of breast cancer or a personal history of benign breast disease. 44 Currently, there is no compelling evidence to avoid the use of oral contraceptives for any subgroup of patients. Most studies of oral contraceptives

-involve women treated with higher estrogen doses than those currently used. Since 1960, the average estrogen content of contraceptive pills has dropped from 150 to 35 mg, and currently available contraceptives may have even less of an effect than that previously observed. Due to uncertainties regarding the latent interval between oral contraceptive use and the onset of cancer and their recent widespread use by very young women, firm conclusions about the safety of oral contraceptives must await further follow-up.

The use of postmenopausal estrogen replacement therapy may be associated with a small increase in breast cancer risk in the· range of 1.5 to 2.0 for moderate-dose conjugated estrogen therapy for periods of 10 to 20 years. 39 The use of estrogen replacement for sh.art periods of time appears safe, and little information is available about the effect of longterm, low-dose therapy: In evaluating the potential risk of estrogen replacement therapy, its proven benefits in reducing osteoporosis and subsequent fracture and lowering the risk of coronary artery disease should be kept in mind."

A possible relation between breast cancer and diet has been suggested by the large variation in international breast cancer incidence rates: National per capita fat consumption correlates with incidence and mortality from breast cancer (Fig. 40-

T


a. &. 25 •1'€»EAl..IV'O'i •u• o FEMALE CAN!IOA• ~~ & •REI........., 11Wl'T2Eu'\""° :S 20 BEl.GIU,,f

!::. ·~'::' W Ol.1.il'Rl~RWAY·~~ !;;[ 15 e1TAi;'cz£~ C:C ePOl'fTUGlii.. ~FIM..AN'J

Q~ 10 ~ •""-"" c;11,,f A V'Ef\EMl.1 -e ~~IA~N

fil ~ eP.JEll!IJRICO

I-5 ePHJUPPf~,:~ICO

~ ~-~ ~ ~ asl= w 00~~.~20~~7.0~~760~~60~~-,700,,---,-1~20~-,-,.7.o~.~~,60;,,-~.,.:,eo

~ TOTAL DIETARY FAT INTAKE (g/day)

FIGURE 40-2. The relation between total dietary fat intake by country and the death rate from breast cancer. (With permission from Carroll K. Experimental evidence of dietary factors and hormone dependent cancers. Cancer Res 1975;35:3374-3383)

2).46 That these differences are not solely due to genetics is suggested by studies of migrants. Japanese women immigrating to the United States and first-generation Americanbom Japanese were found to have an incidence of breast cancer almost equal to that of whites in the same area and considerably higher than that of women in Japan. 47 Although this suggests that environmental factors are important in breast cancer incidence, it does not implicate diet as the sole cause of the observed differences.

Epidemiologic studies of diet and studies of populations with special diets have produced inconclusive results. Kinlen compared breast cancer rates between nuns who were vegetarians or ate only small amounts of meat and single British women who ate regular diets and observed no differences.48 Breast cancer mortality among Seventh Day Adventists, a group that eats a diet low in animal fats, is not significantly lower than expected." Jn the largest prospective study of dietary fat, 89,538 nurses between the ages of 34 and 59 were studied.50

No relation between breast cancer risk and total fat, saturated fat, linoleic acid, or cholesterol intake was found. The difference in fat intake between women at the lowest and highest extremes of fat intake was only 25%. This suggests that dietary fat reduction in the context of the usual American diet is unlikely to reduce breast cancer risk. Jn another prospective

·study with more than 10 years of follow-up, breast cancer risk was actually lower in women with the highest total fat intake, although this difference was not statistically significant after adjustments for total caloric intake were made.51 Currently, there is no conclusive evidence that dietary modification reduces breast cancer incidence or alters prognosis after a diagnosis of breast cancer. 52

A positive but modest association between alcohol use and breast cancer risk is seen in most studies. A metaanalysis of 16 case-control and cohort studies found a relative risk of breast cancer of 1.4 for women with an alcohol intake of two drinks per day.53 A clear-cut dose-response association was observed. The largest cohort study of alcohol and breast cancer, based on data from the American Cancer Society, found no increase in risk in occasional users of alcohol, but risk was elevated in all other categories of alcohol use. 84 Several studies have found that alcohol has its greatest effect on breast cancer risk in women under the age of 30.55

·56

Risk Factors 1267

Women with a family history of breast cancer in a first- or second-degree relative are at increased risk for developing the disease. The risk of developing breast cancer is increased 1.5 to 3.0 times if a mother or sister has the disease, and risk may be greater if a sibling is affected than if a mother is affected. 57

·58 For women whose relatives have unilateral breast

carcinoma, its occurrence premenopausally or postffienopausally does not seem to significantly affect risk. 57 Jn some studies, risk increases with the number of cancer-affected firstdegree relatives, but this is an inconsistent finding. 59•60 Bilateral premenopausal breast cancer in a relative has been associated with the highest risk of breast cancer development. 57 ·61 For most persons with a family history of breast cancer, the lifetime probability of developing breast cancer is rarely greater than 30%, and the magnitude of risk conferred by a positive family history is similar to that seen with maI)y other risk factors. Only 5% of carefully studied breast cane.er patients are thought to have a pedigree consistent with hereditary breast cancer. 62 Breast cancer is observed as part of cancer family syndromes in association with other tumors. These are listed in Table 40-1.

Epidemiologic studies have shown that women exposed to ionizing radiation due to nuclear war or medical diagnostic and therapeutic procedures are at increased risk for the development of breast carcinoma.63

- 66 Multiple chest fluoroscopies, breast irradiation for mastitis, and thymic irradiation increase breast cancer risk. 54

-55 There is a long latent period

for radiation-induced breast cancer, and the risk of developing the disease is related to the age at radiation exposure. Radiation exposure after age 40 results in a minimal increase in risk, and radiation in adolescence is associated with the greatest risk of breast cancer development. 64 Girls irradiated during infancy for thymic enlargement had a linear dose-response risk for subsequent breast cancer development. 66

An assessment of the breast cancer risk associated with benign breast disease cannot be made without specific knowledge of the histologic features of the biopsy. Fibrocystic disease includes a heterogeneous group of pathologic changes associated with various degrees of breast cancer risk. A useful system for classifying benign breast diseases was proposed by Dupont and Page,67 and subsequently adopted at a consensus meeting of the American College of Pathologists. 68 Benign breast conditions were classified as nonproliferative or proliferative, and on the basis of a review of more than 10,000 breast biopsies, relative risks of breast cancer were determined. 67 Women with proliferative disease were found to have

TABLE 40-1. Breast Cancer in Hereditary Syndromes

Syndrome Sites of Other Tumors

Li-Fraumeni syndrome

Cowden's disease

Muir syndrome

Sarcomas (e.g., soft tissue, bone), brain tumors, leukemia, adrenocortical carcinoma

Facial trichilemmomas, papillomatosis of lips and .oral mucosa, acral keratoses, gastrointestinal polyps, uterine leiomyomata

Basal cell carcinoma, benign and malignant gastrointestin~ tumors



a relative risk of 1.9, and the subcategory of women with atypical hyperplasia were found to have a relative risk of 4.4. Nonproliferative breast disease was associated wi~h no excess risk of breast cancer. Sixty-nine percent of the reviewed bi op- · ·sies were found to have nonproliferative breast disease, and of the biopsies demonstrating proliferative changes, only 3.6% were atypical.67 The histologic diagnoses grouped under the headings of proliferative and nonproliferative disease are shown in Table 40-2.

The risk of breast cancer development after a diagnosis of proliferative breast disease with or without atypia is greatest in the first 10 years after the biopsy, with the relative risk associated with atypical hyperplasia halved after this interval and the relative risk for women with proliferative di~ease without atypia returning to the level of the index population. 69

A marked interaction between atypia and a family history of a first-degree relative with breast cancer was also found by Dupont and Page.67 This subgroup of patients had a risk 11 times that of women with nonproliferative disease. The effect of a family history of breast carcinoma in conjunction with atypical hyperplasia on the incidence of cancer for a 25-year period is shown in Figure 40-3. The absolute risk of breast cancer development in women with a positive family history and atypical hyperplasia was 20% at 15 years, compared with 8% in women with atypical hyperplasia and a negative family history of breast carcinoma. No increased risk of breast cancer development was observed in women with a diagnosis of proliferative disease who used estrogens after their breast biopsies. 70

Risk factor assessment for the individual patientis poorly understood by many clinicians. Risk may be expressed as absolute risk or relative risk. Relative risk cannot be multiplied by lifetime risk to determine the risk status of an individual patient. Many believe that risk is most meaningfully discussed with a patient as the risk over a given time interval. Although the lifetime risk of breast cancer development for an American woman is about 11 %, more than half of this risk is expressed after age 65. Moreover, the risk of dying from breast cancer is one third the risk of developing the disease. Risk figures by age interval are shown in Table 40-3. For most women with identified risk factors, the lifetime absolute risk of developing breast cancer is in the range of 20% io 40%. Gail and colleagues developed a model that incorporates age at menarche, age at first live birth, number of previous biopsies, and number of first-degree relatives with breast cancer to provide an in-

·TABLE 40-2. Classification of Benign Breast Diseases

Nonproliferative

Adenosis Cysts, macro or micro Duct ectasia Fibroadenoma Fibrosis Mas ti tis Metaplasia, apocrine or

squamous Mild hyperplasia

Proliferative

Moderate or florid hyperplasia Papilloma with fibrovascular core Atypical hyperplasia, ductal or lobular

a: w "' 80

~ 5 60 w :li ... 0 40 w

~ ffi 20

~ ~

0

NON-PL: Non-Proliferative Lesi~ns POWA: Prolilerative Disease Without Atypla

AH: Atypical Hyperplasia FH: Family History of Breast Cancer

5 10 15 2.0

TIME FROM ENTRY BIOPSY IN YEARS

25

FIGURE 40-3. Freedom from breast cancer in relation to types of benign breast lesions and family history. (With permission from Du-· pant W, Page D. Risk factors for breast cancer in women with proliferative breast disease. N EnglJ Med 1985;3J2,J46-151)

dividualized risk estimate for breast cancer development at different ages over various time intervals. 71 For example, a~ 30-year-old nulliparous female with menarche at age 12, a history of one benign breast biopsy and one relative with breast cancer is estimated to have a 2.3% risk of developing breast cancer in the next 10 years. A 50-year-old with the same risk factors has a 7 .6% risk of developing breast cancer in the next 10 years. The 30-year risk for the 30-year-old woman is 14.9%, and the corresponding risk for the 50-year-old woman is 19.9%. A precise estimate of an individual patient's risk is often helpful, because the extensive publicity about breast cancer causes many women at increased risk to feel that they are doomed to die from the disease.

There are no interventions proven to decrease risk. The data on the use of exogenous hormones are controversial enough that a firm recommendation to avoid their use cannot be made. Similarly, the link between dietary fat and breast cancer is weak, and there is little evidence to suggest that altering diet or alcohol consumption in adult life changes breast cancer risk. Patient education with instruction in breast self-examination, regular physical examinations every 4 to 6 months, and mammography in accordance with standard guidelines are appropriate s\irVeillance techniques.

PREVENTION OF BREAST CANCER

Prevention of breast cancer is possible only by bilateral prophylactic mastectomy. This is a radical approach, particularly because breast-conserving therapy is increasingly employed for the treatment of established cancer. Research is now directed toward developing agents for the primary chemoprevention of breast cancer, an approach that, if succ.essful, will

· be more applicable to the general population.


Breast Biopsy 1269

TABLE 40-3. Probability of Developing and Dying From Breast Cancer

Risk of Developing Risk of Developing Invasive Breast Risk of Dying From

Age (y) Breast Cancer (%) Cancer(%) Breast Cancer (%)

Birth to IJO 10.2 9.8 3.6 20 to 30 0.04 0.04 0.000 20 to 40 0.49 0.42 0.09 35 to 45 0.88 0.83 0.14 35 to 55 2.53 2.37 0.56 35 to IJO 10.27 9.82 3.56 50 to 60 1.95 1.86 0.33 50 to 70 4.67 4.48 1.04 50 to IJO 8.96 8.66 2.75 65 to 75 3.17 3.08 0.43 65 to 85 5.48 5.29 1.01 65 to l!O 6.53 6.29 1.53

(Seidman H, Mushinski M, Gelb SEA. Probabilities of eventually developing or dying of cancer: United States, 1985. CA 1985;35:36-56). The data are derived from white females.

PROPHYLACTIC MASTECTOMY

Little agreement exists on the indications for prophylactic · breast removal. Some have suggested the use of prophylactic . mastectomy for women with a family history of breast cancer, carcinoma of the contralateral breast, high-risk benign histologies, lumpy breasts, or a high degree of anxiety.12- 75 This list includes women with true hereditary breast cancer who may have a risk of developing breast cancer that is as high as 50% and women with lumpy breasts due to nonproliferative fibrocystic disease who have no increased risk of developing breast cancer. 76 There are currently no absolute indications for prophylactic mastectomy. Decisions about surgery should be made only after a detailed discussion with the woman of her relative and absolute risk of developing breast cancer.

Another problem with prophylactic mastectomy is that the efficacy of the procedure is difficult to assess due to inadequate follow-up and poor definition of the risk classification of women who have undergone the surgery. In the Mayo Clinic series of 1500 women treated with prophylactic subcutaneous mastectomy, only 5 (0.0033%) were identified who subsequently developed breast carcinoma at a maximal follow-up of 22 years.77 However, many of these women were at relatively low risk for the development of breast carcinoma, and not all patients were systematically followed. Prophylactic mastectomy does not offer 100% protection against breast cancer. Pennisdollowed 1244 women for 7 years after prophylactic subcutaneous mastectomy and found a cancer incidence of 0.5%.78 Subcutaneous mastectomy is an unsatisfactory operation for breast cancer prophy !axis because glandular breast tissue is left behind beneath the nipple and often on the skin flaps and in the tail of the breast. 75·79 In animals, a reduction in the volume of breast tissue is not associated with a proportionate reduction in breast cancer risk.80

If prophylactic surgery is undertaken, the operative procedure should be a total mastectomy with sacrifice of the nippleareolar complex. Even with this procedure, breast tissue may be left behind on the flaps or in the axilla unless great care is taken with operative technique. 81 In considering prophylactic mastectomy, it should be kept in mind that for nearly

all women, their chances of not having breast cancer. exceed their risk of deveioping the disease, and even bilateral mastectomy is not an absolute guarantee that breast cancer will not develop .

CHEMOPREVENTION

One of the most important new strategies to address the breast cancer problem is the use ·of agents for the chemoprevention or the chemosuppression of the disease. The two m'ain agents being evaluated are tamoxifen and retinoids. Much of the impetus for the use of tamoxifen as a chemopreventive agent comes from the trials of its use as an adjuvant systemic agent, which demonstrated a reduction of approximately 40% in the number of contralateral breast cancers in tamoxifen-treated women compared with controls. A trial to test the value of tamoxifen as a prevention for breast cancer is currently underway in the United Kingdom and will be underway soon in the United States by the National Surgical Adjuvant Breast and Bowel Project (NSABP). Retinoids have been shown to be effective in preventing or reducing the incidence of breast cancer in several animal studies. s2-s4 A clinical ·trial of the synthetic retinoid N-( 4-hydroxyphenyl) retanamide (HPR) has been undertaken by the National Cancer Institute in Milan. Three thousand women previously treated for a unilateral Tl or T2, node-negative breast carcinoma have been randomized to HPR or no treatment, and the incidence of contralateral breast cancer will be assessed. No data are currently available from this trial. There is no indication for the use of tamoxifen or the retinoids as chemopreventive agents outside of a clinical trial.

BREAST BIOPSY

INDICATIONS FOR BIOPSY: CLINICAL ABNORMALITIES

The most common clinical indications for breast biopsy are · a dominant breast mass or a pathologic nipple discharge. Although this seems straightforward, the determination of what



Constitutes a dominant breast mass is frequently difficult, particularly in the premenopausal woman. Normal breasts are a mixture of fat and glandular tissue. The glandular tissue, most of which is located in the upper outer quadrant, changes throughout the menstrual cycle, and this type of physiologic nodularity should not be considered an indication for a·breast biopsy. Domiriant masses are characterized by their persistence throughout the menstrual cycle. Breast pain, in the absence of a dominant mass, should not be an indication for a breast biopsy. Breast pain is an uncommon presenting symptom of breast carcinoma, seen in only 7o/o of cases reported by Preece, and if it is not accompanied by clinical or mammographic evidence of a breast mass, medical management, rather than surgical intervention, is appropriate. 85

If a dominant breast mass is detected in a premenopausal woman, it should be aspirated to determine if it is a cyst. Cysts require biopsy only if the aspirated fluid is bloody, the palpable abnormality does not completely resolve after the aspiration of fluid, or the same cyst recurs multiple times. This policy can be safely followed because intracystic carcinoma accounts for fewer than 1 % of all breast cancers." Routine cytologic examination of breast cyst fluid is not indicated because of the low likelihood of carcinoma in the absence of bloody fluid or recurrent cysts. 87

•88 Noncystic masses in premenopausal

women that are clearly different from the surrounding breast tissue should be biopsied using one of the techniques described later. Observation for orie to two menstrual cycles is only appropriate for vague asymmetry or nodularity if it is uncertain that a dominant mass exists. In women older than 35, a mammogram is an essential part of the prebiopsy workup of a breast mass. Communication with the radiologist allows a complete radiographic evaluation, which may include compression and magnification views to define the extent of associated microcalcifications, or ultrasound if a cyst is suspected.

In postmenopausal women, clinical examination of the breasts is frequently easier due to atrophy of the nodular glandular elements. Without e~ogenous estrogen, cysts are uncommon after the perimenopausal years, and the masSes should be regarded with a higher degree of suspicion than those in premenopausal womei:i. Benign breast conditions causing palpable masses are less frequent, and small areas of nodularity that might be observed in premenopausal women should be considered for prompt biopsy in the postmenopausal patient.

Although nipple discharge is a common complaint, it is an uncommon sign of breast carcinoma. Three percent to 11 % of women with malignant disease have an associated nipple discharge. 89- 91 The likelihood of a nipple discharge being secondary to malignancy increases as patient age increases. A 32o/o incidence of carcinoma was found for women older than 60 presenting with a discharge and no mass, compared with a 7o/o incidence in women yoUnger than 60. 91 Discharges are considered pathologic if they are unilateral, spontaneous, and localized to one duct. Discharges secondary to carcinoma may be bloody or serous. Bloody discharges are reported in 70% to 85% of the cancers that present with nipple discharge.""°'' If a discharge occurs, it should be tested for occult blood and localized to one quadrant of the breast. A mammogram should be obtained to look for dilated ducts or occult masses. Galactography cannot reliably differentiate benign from malig-

nant causes of discharge and is not usually helpful in cases where clinical indications for surgery are present.93 A persistent pathologic discharge is an indication for duct exploration. In a review of 1956 cases of pathologic discharge, 44% were due to papilloma or papillomatosis, 23% to duct ectasia, 16% to fibrocystic disease, and only 11 % were associated with malignancy.94 Terminal duct excision is the procedure of choice for the diagnosis andtreatment of mpple discharge in the absence of a breast mass. If a mass is associated with the discharge, the mass should be biopsied.

BIOPSY TECHNIQUES

Palpable Masses

Four techniques are available for the diagnosis of breast masses. They are fine-needle aspiration, core-cutting needle biopsy, incisional biopsy, and excisional biopsy. Fine-needle aspiration and core-cutting needle biopsy are office procedures. Excisional biopsy, with rare exceptions, is an outpatient procedure that can be done using local anesthesia.

Fine-needle aspiration is becoming increasingly popular for the diagnosis of dominant breast masses. The procedure has the advantages of being quick, relatively painless, and inexpensive to perform. The main drawbacks of the procedure are the need for an experienced cytopathologist to interpret smears and the risk of false-negative results. Strawbridge, reviewing 890 lesions with histologic confirmation of cytologic results, found a 9.6% false-negative rate." Bell and coworkers reported a 4% false-negative rate in their series of 1680 aspirates.97 Kline and colleagues, in a review of 3545 breast aspirates, reported a 9.6% incidence of false negatives." In half of Kline's cases, the needle tract did not extend into the tumor, most·of which were smaller than 1 cm in diameter. Extremely flbrotic tumors were found to be a source of falsenegative results, and invasive lobular carcinoma was more likely to be a source of false-negative results than infiltrating ductal carcinoma.95 In addition to false-negative aspirates, specimens with material insufficient for diagnosis are reported in 1% to 10% of aspirates.88

·95·96 Barrows found that the ex

perience of the physician performing the aspirate was the factor that most strongly predicted a successful aspirate. False-· positive aspirates are extremely uncommon, reported in fewer than 1 % of cases in most large series. 8~·95•96 However, :fineneedle aspiration cannot identify cases of gross ductal carcinoma in situ as noninvasive tumors, potentially leading to overtreatment.

Fine-needle aspiration is an excellent technique for establishing a diagnosis ofbr_east cancer before placing an incision in the breast. A treatment option can then be selected by the patient, and only a single operative procedure is needed. Because of the false-negative rate of up to 10% seen with fineneedle aspiration, dominant masses for which a malignant diagnosis is not obtained should be excised. Several physicians have suggested that patients with benign cytologic results and a mammogram and physical examination suggesting benign breast disease can be followed expectantly, with a rate of missed breast cancer of less than 1 %. In view of the low morbidity of breast biopsy, this seems unnecessarily risky if a dominant breast mass is present. 97- 99

Core-cutting needle biopsy has many of the advantages of


fine-needle aspiration. It is a rapid, relatively painless office procedure. Because a core of tissue is obtained for histologic examination, more details of tumor structure are available, and ductal carcinoma in situ can be identified. Fentiman reported a 79% accuracy rate for core-cutting needle biopsy, and no false-positive results were seen. 100 Minkowitz found an 89% sensitivity for the technique, which increased to 94% for lesions larger than 2.5 cm.101 Aspiration cytology and corecutting needle biopsy were prospectively compared for 81 women by Shabot and coworkers. The accuracy of fine-needle aspiration was 96.2%, compared with 78.9% for core-cutting needle biopsy.102 The choice of core-cutting needle biopsy or fine-needle aspiration depends on the availability of an experienced cytopathologist.

Excisional biopsy has been the standard technique for the diagnosis of breast masses. It has the advantage of allowing a complete evaluation of the tumor size and its histologic characteristics before instituting definitive therapy. When an excisional biopsy is done for the diagnosis of a breast mass, it should be done as a definitive lumpectomy, with the mass excised with a surrounding margin of normal breast tissue. The biopsy incision should be placed directly over the mass in a curvilinear fashion. The exception to this rule is the young woman with a clinical diagnosis of fibroadenoma; the mass can be shelled out without removing additional breast tissue. Breast tissue should not usually be reapproximated after biopsy to avoid distortion of the breast contour, and drains should be avoided for the same reason. Proper handling of the specimen is crucial to the success of the biopsy. The specimen should be excised as a single piece of tissue and orienting sutures placed for the pathologist. The tissue should not be placed in formalin and should be promptly sent to the pathology department to allow inking for evaluation of the margins and sampling for hormone receptors. Most excisional biopsies are done as outpatient procedures. There is no evidence that a one-step procedure (i.e., biopsy under general anesthesia followed by definitive surgery if positive) is associated with any survival benefit compared with biopsy followed later by definitive surgery, and outpatient breast biopsy is more cost effective. 23,103-105

Incisional biopsy is used to establish a diagnosis of breast cancer in masses too large to excise completely. It is frequently employed in women with metastatic disease or those with locally advanced breast cancer who will be r.eceiving systemic therapy as an initial treatment approach. The problems of poor wound healing, which can occasionally occur in this setting, can be avoided by the use of fine-needle aspiration with immuncihistochemical hormone receptor determination rather than incisional biopsy.

Mammographic Abnormalities

The classic mammographic signs of clinically occult malignancy are clustered microcalcifications and a stellate mass. A review of 5500 biopsies from 17 reports of 100 or more cases identified microcalcifications as the indication for biopsy in 45o/o of cases, masses in 430/o, masses containing microcalcifications in 6%, and asymmetric density in 5%. 106 The incidence of positive biopsies done for mammographic findings ranges from 9% to 65%, with most investigators reporting a 15% to 30% positive biopsy rate. 101

-110 Moskowitz found

Breast Biopsy 1271

that the predictive value of microcalcifications for the presence ofcancer was 11.5±1.7%, based on his review of 40,431 screening mammograms. 111 The predictive value of masses that were thought to be definitely malignant was 7 4%, but masses thought to be possibly malignant yielded carcinoma in only 5.4% of cases. Because the predictive value of any mammographic sign depends on the skill of the observer and the quality of the mammograms, positive biopsy yields in some institutions may be lower than those reported in the literature.

Special mammographic views are frequently helpful in the evaluation of equivocal findings on two-view mammography. ,-/ Tabar found that 62% of abnormalities identified on a singleview screening examination were shown to be of no concern with additional views-"0 Sickles evaluated 302 women with indeterminate microcalcifications with magnification views ·and demonstrated 61 % of the lesions to be benign and 11 % to have previously unsuspected malignant characteristics. 112

Even if biopsy is clearly indicated on the basis of the initial mammogram, magnification views are useful in delineating the extent of microcalcifications in women considering breastconserving therapy.

Magnification views of mass lesions are helpful in equivocal cases. Magnification mammography increased diagnostic accuracy from 29% to 69% in 216 cases with indeterminate masses or microcalcifications on initial studies. 113 Other radiologic techniques are not particularly helpful in the evaluation of occult breast masses, although sonography is useful for determining if mass lesions are cysts. Transillumination light scanning, computerized tomography (CT), and magnetic resonance imaging (MRI) have not been shown to be superior to mammography.

Nonpalpable Masses

Because only 15% to 30% of mammographic abnormalities are shown to be ·cancer, a biopsy technique that reliably excises the lesion with a good cosmetic result is essential. Blind excisions of large amounts of breast tissue or the use of skin marking as a localization technique are unsatisfactory, with failure to excise the mammographic abnormality reported in 1 7% of cases, even with the removal of large amounts of breast tissue. 114 Localization can be done using straight needles, hook wires, or dye injection along a needle tract. The most important factor in the success of localization is how close the wire is placed to the mammographic abnormality. Gallagher reported wire placement to within 2 mm of the target in 96% of cases, with 96% of the lesions being successfully removed on the first attempt. us The average specimen size was 6 cm3.

Communication between the mammographer and the surgeon is important to ensure that the incision for the needle localization biopsy is placed over the termination of the wire rather than at the entry of the wire into the skin to avoid tunneling to the lesion. Specimen radiography should be carried out in all biopsies done for microcalcifications to confirm the presence of the calcifications in the biopsy specimen. Compression and magnification views are useful for specimen evaluation. Although nonpalpable masses can frequently be identified grossly at the time of biopsy, it is still useful to obtain x-ray films of the specimen to ensure that the gross lesion corresponds to the mammographic abnormality. The specimen



should be marked with orienting sutures and inked to allow evaluation of margins.

Frozen section is reliable in the diagnosis of palpable breast masses, but indications for its use in the evaluation of nonpalpable breast abnormalities are limited. If a definite mass is palpable in.traoperatively, a frozen section may be done and tissue saved for estrogen receptors if the frozen-section specimen is positive. Tinnemans evaluated the reliability offrozensection diagnoses in 297 clinically occult lesions. 116 Diagnosis was deferred to permanent section in 12% of cases, and if a definitive diagnosis was made, thei:e was a 3o/o incidence of false-negative results and two instances of false-positive results. Because the abnormalities being sought by needle localization are usually small and often represent histologically borderline lesions, they may be difficult to diagnose on frozen section. Because needle localization is rarely undertaken with a plan to proceed to definitive therapy at the same operation, a careful examination of all the available pathologic material in paraffin sections would seem to be the most prudent course. Failure to remove the mammographic abnormality occurs in 2o/o to 8% of localizations. 106 In these cases, persistence of the lesion on mammogram should be confirmed and repeat biopsy undertaken.

Aspiration cytology, using a stereotactic device for needle positioning, has been employed for the diagnosis of nonpalpable breast abnormalities. Azavedo and coworkers aspirated 2594 mammographic abnormalities.117 Seventy-seven percent of the patients had benign aspirates and received no further therapy if the mammographic finding was thought to be of low suspicion. One of these 2005 women developed a carcinoma 14 months later. Overall, only 21.9% of the mammographic abnormalities were biopsied, and 76% of those biopsied were carcinoma. Diagnostic sensitivities and specificities over 90o/o are reported from several centers using this procedure.111-119 The major use of aspiration cytology of nonpalpable breast masses has been to avoid biopsy of minimally suspicious lesions' unless cytologic evidence of atypia is found. 120 Stereotactic fine-needle aspiration is unlikely to be of major use in the management of highly suspicious mammographic abnormalities, because it does not reliably differentiate invasive and in situ disease. The availability of stereotactically directed core biopsies may partially resolve this problem, but only complete excision of the mammographic

abnormality allows a reliable assessment of invasion to be made.

CLASSIFICATION OF TUMOR TYPES

Histopathologic examination of breast cancer makes available information that establishes the diagnosis of the lesion, aids in determining patient prognosis, and leads to a better understanding of the nature of breast cancer.

Several pathologic classifications of mammary carcinomas are in use. The most commonly used are those presented by the Armed Forces Institute of Pathology (AFIP) 121 and the World Health Organization. 122 Breast carcinomas are classified as ductal or lobular, corresponding to t.he ducts and lobules of the normal breast. However, there is evidence that most tumors arise in the terminal duct section of the breast, regardless of pathologic type. 123 A comparison of the histologic types of infiltrating breast carcinomas from the AFIP is given in Table 40-4. 121

CARCINOMA IN SITU

Tumors arising from duct epithelium that are confined within the lumen of the ducts or lobules of the breast are referred to as carcinoma in situ. carcinoma in situ has been classified as ductal or lobular, depending on the cytologic features and pattern of growth. Ductal carcinoma in situ (DCIS), also known as intradt.J.ctal carcinoma or noninvasive ductal carcinoma, and lobular carcinoma in situ (LCIS) are characterized by a proliferation of malignant epithelial cells confined to the mammary ducts or lobules, without light-microscopic evidence of invasion through the basement membrane into the surrounding stroma. The distinction between DCIS and LCIS is usually not difficult, but overlaps exist. The natural history and management of these lesions are discussed later.

INFILTRATING DUCTAL CARCINOMAS

A variety of histologic types of invasive (infiltrating) carcinomas of the breast have been described. Infiltrating ductal carcinomas in which no special histologic features are recognized are designated NOS (not otherwise specified) and

TABLE 40-4, Comparison of Histologic Types of Infiltrating Breast Carcinoma

Characteristic Ductal Lobular Medullary Colloid Come do Papillary

Percent"' 78.1% 8.7% 4.3% 2.6% 4.6% 1.2% Average age (y) 50.7 53.8 49.0 49.7 48.6 51.9 Average size (cm) 3.1 3.5 3.4 3.8 3.9 3.4 Involved nodes 60% 60% 44% 32% 32% 17% Survival

At 5 years 59% 57% 69% 76% 84% 89% At 10 years 47% 42% 68% 72% 77% 65% At 15 years 38o/o 34% 62% 62% 74% 65%

"' Percent of all infiltrating carcinomas. (McDivitt R, Stewart F, Berg J. Tumors of the breast. In: Atlas of tumor pathology. Bethesda, MD: Armed Forces Institute of Pathology, 1967)

L1 62 of 122 Celltrion, Inc., Exhibit 1006

account for almost 70% of breast cancers. They are characterized by their stony hardness to palpation. When transected, a gritty resistance is typically encountered, and the tumor retracts below the cut surface. Histologically, various degrees of fibrotic response and associated DCIS are present. These tumors. commonly metastasize to the axillary lymph nodes, and their prognosis is the poorest of the various ductal types.

MEDULLARY CARCINOMA.

Medullary carcinomas are circumscribed lesions that can attain large dimensions, but they demonstrate only low-grade infiltrative properties. They constitute 5% to 7% of all mammary carcinomas and are characterized by a wellcircumscribed border, poorly differentiated nuclei, syncytial growth pattern, and an intense infiltration with small lymphocytes and plasma cells. The 5-year survival rate after treatment for medullary carcinoma is better than for NOS ductal carcinomas. The more favorable prognosis requires the presence of all of these characteristics; tumors with some of these features (i.e., atypical medullary tumors) do not have a more favorable prognosis. 124

TUBULAR CARCINOMA

A tumor in which tubule formation is conspicuou~ is known as tubular or well-differentiated carcinoma. This diagnosis. is made only if 75% or more of the tumor is composed of these elements. Axillary metastases are uncommon, and the prognosis is considerably better than for NOS ductal carcinoma.

MUCINOUS CARCINOMA

Another ductal type, the mucinous or colloid carcinoma, comprises about 3% of all mammary carcinoma. It is slow growing and can reach bulky proportions. When the tuinor is predominantly mucinous, the prognosis tends to be good.

Rarer types of ductal carcinomas include papillary, adenocystic, apocrine, secretory, squamous, and carcinosarcoma or metaplast.ic duct carcinoma. Jn many cases, NOS ductal carcinomas contain small areas of these special types.

INFILTRATING LOBULAR CARCINOMA

Another histologic type of breast cancer is infiltrating lobular carcinoma. It is relatively uncommon, accounting for only 5% to 10% of breast tumors in most series. The clinical presentation is often an area of ill-defined thickening in the breast, unlike the dominant lump characteristic of ductal carcinoma. Microscopically, lobular carcinomas typically are composed of small cells. in a linear arrangement ("Indian filing") with .a tendency to grow around ducts and lobules (i.e., targetoid growth). Lobular carcinomas are characterized by a greater proportion of multicentric tumors in the same or the opposite breast than are found in NOS ductal carcinoma. Overall, infiltrating lobular carcinoma has a similar likelihood of axillary nodal involvement and prognosis as infiltrating duct carcinoma. However, the sites of metastases for these two types tend to differ. Ductal carcinomas more characteristically metastasize to bone or to intraparenchymal sites within lung,

Local and Regional Spread of Breast Cancer 1273

liver, or brain, and ·lobular carcinomas more often show a predilection for meningeal and serosal surfaces. 125

INFILTRATING COMEDOCARCINOMA

Infiltrating comedocarcinoma is a type of infiltrating ductal carcinoma. It is composed primarily of the comedo type of ductal carcinoma in situ, which is characterized by large poorly differentiated nuclei with frequent mitoseS and central ne'crosis with areas of invasion. This. term has not been used recently because of possible confusion with pure comedo ,-.,-' DCIS. Tumors with an extensive intraductal component ap-pear to have important implications in breast-conserving treatment and are discussed in the section on local treatment.

PAGET'S DISEASE

Paget's disease of the breast occurs in 1 % of all patients with breast cancer. Clinkally, the patient presents with a relatively long history of eczematous changes in the nipple with itching, burning, ooiing, bleeding, or some combination of these. The nipple changes are associated with an underlying carcinoma in the breast that can be palpated in one half to two thirds of the patients. The subadjacent tumor may be the intraductal or invasive duct type. The prognosis is related to the histologic type of the associated tumor. Histologically, the nipple epidermis contains tunior cells singly and in nests. Treatment of Paget's disease is discussed iater in .this chapter.

INFLAMMATORY BREAST CARCINOMA

Inflammatory breast cancer is characterized clinically by prominent skin edema, redness and warmth, a visible erysipeloid margin, and induration of the undei:lying tissue. Jn approximately half of these patients, a mass is not palpated. These criteria in the past were sufficient for the diagnosis, but currently, pathologic corroboration must be obtained by a skin biopsy demonstrating involvement of the dermal lymphatics with cancer cells. Jnflainmatocy cells rarely are present. The prognosis of patients with inflammatory breast cancer is poor, even if the disease is apparently localized. The management of inflammatory cancer is discussed later in this chapter. ·

LOCAL AND REGIONAL SPREAD OF BREAST CANCER

The features of breast cancer related to its local and regional involvement provide the basis for local. treatment, which is discussed later in this chapter.

LOCAL INVOLVEMENT

The primary site of breast cancer is described by the quadrant of the breast in which it is found. In one series of 696 patients, 48% of the tumors were located in the upper outer quadrant, 15% in the upper inner quadrant, 11 % in the lower outer quadrant, 6% in the lower inner quadrant, and ·17% in the central region (i.e., within 1 cm of the areola).126 An additional 3% were called diffuse because of multifocal origin or in-



volvement of the entire breast. The higher frequency of breast cancer in the upper outer quadrant is thought to be attributable simply to the greater amount of breast tissue in that quadrant. In this series of patients, no differences in survival based on quadrant location were found. The relation between the location of the primary tumor and prognosis was examined in another large NSABP series. Relapse and ultimate survival were reiated to the pathologic status of the axillary nodes, arid there were no significant differences in prognosis by primary tumor location (Table 40-5). 127

The spread of cancer through the breast was summarized by Haagensen. 128 This spread occurs by direct infiltration into the breast parenchyma, along mammary ducts, and through breast lymphatics. Direct infiltration tends to occur by ramifying projections that have a characteristic stellate appearance on gross examination. If untreated, direct involvement of overlying skin or deep pectoral fascia is common. Involvement along ducts is observed frequently and may include wide segments of the breast. It 'is unclear whether this intraductal involvement represents true spread of a primary cancer along previously uninvolved ducts or a "field cancerization" that results in simultaneous transformation along entire lengths of ducts. Spread can occur by the extensive network of breast lymphatics. Investigators have emphasized lymphatic spread vertically down to the lymphatic plexus in the deep pectoral fascia underlying the breast, and spread to the central subareolar region has been .described. These multiple mecha: nisms of spread emphasize the likelihood of cancer being present in the breast well beyond the palpable primary mass.

A detailed study of the sites of cancer in a breast containing a primary tumor has been performed by Holland and associates.129 They examined 264 mastectomy specimens from patients with clinically unifocal breast cancer measuring 4 cm or less. In only 40o/o'of cases was the cancer in the breast restricted to the primary tumor (Fig. 40-4). The probability of finding additional foci of cancer decreased as a function of the distance from the primary tumor: 41 % of the specimens had additional foci of cancer 2 cm or more from the primary tumor, and 11 % had additional foci 4 cm or more from the primary tumor. Of the cases with additional foci beyond 2 cm, the additional foci were intraductal in approximately two thirds of the patients.

TABLE 40-5. Five-Year Relapse Rate According to the Location of the Primary Tumor and Nodal Status

Location* Negative Nodest Positive Nodest

UOQ 17 (208) 63 (239) UIQ 25 (75) 59 (37) LIQ 22 (23) 55 (22) LOQ 26 (46) 70 (44)

• U, upper; 0, outer; Q, quadrant; I, inner; L, lower. t Five-year relapse rate, with the number of patients in the subgroup in parentheses.) (Fisher B, Slack N, Ausman R. Location of breast carcinoma and prognosis. Surg Gynecol Obstet 1969;129,705-716)

Cumulative •1. of cases--------------~ 100

BO

60

40

20

0 0 2

Invasive {non-diffuse} cancers, pathologic sizes ~ 4 cm, n "264

4 6 8 10 cm Distance from the (patho:logica11y estimated) reference tumor

FIGURE 40-4. Distribution of tumor foci at different distances from the reference tumor. (A) Cases without tumor foci outside the reference tumor. (B) Cases with tumor foci within 2 cm. (C) Cases with noninvasive tumor foci at a distance greater than 2 cm. (D) Cases with invasive tumor foci at a distance greater than 2 cm. (With permission from Holland R, VelingS, MravunacM, HendriksJ. Histologi.c multifocality of TIS, Tl-2 breast carcinomas. Implications for clinical trials of breast-conserving surgery. Cancer 1985;56:979-990)

REGIONAL NODAL INVOLVEMENT

The most common sites of regional lymph node involvement in breast cancer are the axillary, internal mammary, and supraclavicular lymph node regions. A knowledge of the likelihood of involvetnent of these areas and their significance is critical for planning treatment. The axillary lymph node region is the principal site of regional metastases from carcinoma of the breast, and approximately 40% of patients have evidence of spread to the axillary nodes. The likelihood of axillary nodal involvement appears to be related directly to the size of the primary tumor. 130

To some extent, the incidence of histologic involvement of axillary nodes depends on the extent of the pathologic analysis of the specimen. Pickren was the first to show that a more thorough clearing and sectioning of the axillary specimen resulted in a greater yield of positive nodes.131 Of 51 specimens analyzed in routine fashion and found to be negative, 11 (22%) showed evidence of involvement on more careful analysis in that study.

Detection of axillary involvement by physical examination has high false-positive and high false-negative rates (Table 40-6).132

-135 If axillary lymph nodes are palpable, histologic

evidence of metastatic disease is not found in approximately 25% of patients. Conversely, if axillary nodes are not palpable, histologic involvement is detected in approximately 30% of patients. These shortcomings of clinical evaluation are of particular importance because histologic involvement of ax-

TABLE 40-6. Accuracy of Physical Examination in Predicting Histologic Involvement of Axillary Nodes

Accuracy of Series Series Series Series . Prediction ]132 2133 3134 4ws

False-positive rate (%) 25 31 26 29 False-negative rate (o/o) 32 28 27 29


Local and Regional Spread of Breast Cancer 1275

TABLE 40-7. Ten-Year Survival by Axillary Node Status for Patients Treated With Radical Mastectomy

Negative Nodes

Investigations (%)

Valagussa136 80 Haagensen 137 76 Schottenfeld 135 72 Fisher138 65 Spratt and Donegan139 68 Payne140 76 Ferguson141 72

illary nodes has a high correlation with prognosis. Table 40-7 shows IO-year survival figures according to axillary involvement from seven separate series of patients treated with radical mastectomy. 13

5-141 Patients with histologically negative

axillary nodes have a markedly greater likelihood of survival than patients with histologic involvement. The prognosis is inversely related to the number of involved nodes.142 The combination of the presence and extent of metastases to the axilla represents the single most important prognostic factor for patients with breast cancer.

The axilla is a triangle bounded by the axillary vein superiorly, the latissimus dorsi laterally, and the serratus anterior medially (Fig .. 40-5). For the purposes of analysis, the axilla is commonly divided into three levels: proximal, which is tis-

FIGURE 40-5. Anatomy of the axilla. The levels are defined in relation to the pectoralis minor muscle. (Kinne DW. Primary treatment of breast cancer. In: Harris JR, Hellman S, Henderson IC, Kinne DW, eds. Breast diseases. 2nd ed. Philadelphia, JB Lippincott, 1991)

4 or More Positive Positive Nodes 1-3 Positive Nodes (%) Nodes(%) (%)

38 50 24 48 63 27 43 25 38 13 27 35 39 52 27

sue inferior to the lower border of the pectoralis minor muscle (level I); middle, which is tissue directly beneath the pectoralis minor (level II); and distal, which is tissue superior to the pectoralis minor (level Ill). Prognosis is related to the level of axillary involvement (Table 40-8)."5 Involvement of the upper level nodes carries a worse prognosis than involvement of proximal level nodes alone. In a series of 182 mastectomy specimens examined by clearing, involvement of nodes at the apex of the axilla was found in 15, and all 15 patients relapsed, indicating the grave prognosis associated with involvement high in the axilla. 128 Involvement ofupper-level nodes usually is associated with a high total number oflymph nodes involved; in this group of 15 patients, the mean number of involved nodes was 16.2 (range, 4-37). In another study, axillary node involvement and survival were examined in 385 patients to determine whether the total number of involved nodes or the level of axillary involvement was the better indicator of prognosis.143 For any given number of involved nodes, survival was independent of the level of involvement, and the investigators concluded that prognosis was related more directly to the total number of nodes involved than to the level of involvement.

The distribution of axillary node involvement by level has been studied in two large series, with nearly identical re-

TABLE 40-8. Ten-Year Survival Related to Primary Tumor Size and Level of Axillary Involvement

Axillary Status <2

Negative 82 Positive

Proximal only 73 Middle or distal - • All 68

* Insufficient data.

Size of Primary Tumor(cm)

2-5 >5

65 44

74 39 28 37 51 37

Total

72

65 31

(Schottenfeld D, Nash A, Robbins G, Beattie E. Ten-year results of the treatment of primary operable breast cancer. Cancer 1976;38: 1001-1007)

~-



TABLE 40-9. Five-Year Relapse Rate(%) According to Size of Primary and Axillary Node Involvement

Axillary Status

Negative Axillary Nodes Fisher et al146

Nemoto et al 142

Valagussa et aP36

Positive Axillary Nodes Fisher et al146

Nemoto et al142

Valagussa et al1 36

Size of Primary Tumor (cm)

<2 2-5 >5

12 24 27 13 19 25 8 24 19

50 60 79 39 50 65 37 64 74

sults. 144·145 Involvement of level I alone was seen in 54o/o to 58% of patients and level I and II in 20% of patients. Involvement of levels II or III without involvement of level I (i.e., skip metastases) was seen in only 2% to 4% of patients with nodal involvement. These results indicate that involvement of the a.Xilla is usually sequential. A level I dissection is highly effective for determining nodal involvement but frequently underestimates the extent of involvement.

Prognosis is related to the size of the primary tumor and to axillary node involvement. Whether these two factors independently predict the outcome is addressed in Table 40-9. If axillary nodes are involved, the size of the tumor still has prognostic value. For example, in the data from Valagussa and colleagues, the 5-year relapse rate was 37% for patients with positive nodes and small (,;2 cm) tumors and 79% for patients with positive nodes and large (>5 cm) tumors. 136 In the data from Fisher and coworkers, this correlation was analyzed further according to the number of positive axillary nodes (i.e., 1 to 3 or 4 or more). 146 Within each subgroup with positive axillary nodes, the size of the primary tumor was still an independent prognostic factor. However, if axillary nodes are negative, the relation is less clear. The prognosis for patients with small ·tumors and negative nodes is exceptionally good, with a 5-year relapse rate of approximately 10%. For tumors larger than 2 cm, the prognosis is not as good.

However, the prognosis for patients with large tumors and negative nodes is not significantly worse than that of patients with 2- to 5-cm tumors and negative nodes. These data imply that the results of an axillary sampling are of value for prognostic purposes in patients with large primary tumors, because patients with histologically negative axillary nodes do relatively well even without adjuvant therapy. The 30-year survival rate was 61 % if the primary tumor was no larger than 2 cm, 46% if it was 2 to 5 cm, and 50% if it was larger than 5 cm. 147

For patients with involvement of level I axillary nodes, the 30-year survival rate was 40% if the tumor was no larger than 2 cm, 31% if it was 2 cm to 5 cm, and only 14% if it was __ larger than 5 cm. ·

The axillary nodal region is the principal drainage site for carcinoma of the breast, and a histologic analysis of the axilla provides a useful guide to prognosis. The more practical issue of what treatment is required for the axillary region is discussed in later sections.

The second major site of regional metastases for carcinoma of the breast is the internal mammary lymph node chain, which lies at the anterior ends of the intercostal spaces by the side of the internal thoracic artery. Because of their intrathoracic location and their uncommon clinical presentation, the frequency of internal mammary node involvement was not appreciated as early as was axillary node involvement. One of the first to document this second route of spread was Sampson Handley, who reported his results of internal mammary node biopsy in 1000 patients in 1975 (Table 40-10).148

These results illustrate- that internal mammary node involvement is more common for inner quadrant or central tumors than for outer quadrant tumors and that axillary lymph node involvement is more likely than internal mammary node involvement. In the Handley study, even in patients with inner or central tumors, axillary involvement was more common than internal mammary node involvement ( 42% versus 28%). If the axillary nodes were involved, internal mammary node involvement was uncommon (8%). Another larger series of patients reported from Italy confirmed the Handley results. 132

These investigators stress the importance of primary tumor size in relation to internal mammary node involvement. Internal mammary node involvement was seen in 19% of patients with tumors smaller than 5 cm and 37% of patients with tumors larger than_ 5 cm.

TABLE 40-10. Internal Mammary Node Involvement(%) in Relation to Location of the Primary and Axillary Node Involvement

Node Involvement UIQ' LIQ

Totalt 27% 33% (67/248)> (20/61)

Axilla not involved 14% 6% (20/143) (2/36)

Ndlla involved 45% 72% (47/105) (18/25)

* U, upper; I, inner; Q, quadrant; L, lower, 0, outer. t All patients regardless of axillary status. :j: Numbers in parentheses are crude results.

Primary Site

Central UOQ

32% 14% (70/216) (54/382)

7% 4% (5/76) (71170) 46% 22%

(65/140) (47/212)

(Handley R. Carcinoma of the breast. Ann R Coll Surg Engl 1975;57,59-66)

LOQ

13% (12/93)

5% (2/40) 19%

(10/53)

,,r-


The significance of internal mammary node involvement is similar to that of axillary node involvement. In a large series reported by Veronesi and coworkers, the 10-year rate of disease-free survival was 73% if the axillary nodes and the internal mammary nodes were negative, 47% if axillary nodes alone were positive, 52% ifthe internal mammary nodes alone were positive, and only 25% if both areas were positive.149 In practice, biopsy of the internal mammary nodes is associated with a greater likelihood of morbidity than biopsy of axillary nodes and is rarely performed.

The principal route of spread to the supraclavicular lymph node areas is through the axillary node chain. In one series of patients undergoing routine supraclavicular dissection, involvement of the region was found in 23 (18%) of the 125 patients who had involvement of axillary nodes but in none of the 149 patients who did not have involvement of axillary nodes.145 The significance of supraclavicular node involvement was first shown by Halsted, who performed a supraclavicular dissection in 119 patients. Forty-four (37%) women were found to have involvement of these nodes, and only two were free of cancer at 5 years. 150 Supraclavicular node involvement represents a late stage of axillary nodal involvement and carries a grave prognosis.

DISTANT METASTASES

Metastatic spread from carcinoma of the breast can involve a variety of organs. The likelihood of organ involvement has been studied in several autopsy series, and the results are shown in Table 40-11. 128

•151

•152

TABLE 40-11. Percentage of Patients With Metastatic Breast Cancer at Various Sites in Three Collected Series

Series 1151 Series 2152 Series 3128

Site (n = 160) (n = 43) (n = 100)

Lung 59 65 69 Liver 58 56 65 Bone 44 71 Pleura 37 23 51 Adrenals 31 41 49 Kidneys NR' 14 17 Spleen 14 23 17. Pancreas II 17 Ovaries 9 16 20 Brain 9 22 Thyroid 24 Heart II Diaphragm II Pericardium 5 21 19 Intestine 18 Peritoneum 12 9 13 Uterus 15 Lymph nodes 72 76 Skin 34 7 30

* NR, not recorded.

Pretreatment Evaluation 1277

PRETREATMENT EVALUATION

The pretreatment evaluation of the breast cancer patient is directed at identifying the clinical stage of the patient's disease and disease sites that would alter the treatment plan. A complete history and physical examination is the first step in this evaluation. The components of the breast-directed history and physical examination are listed in Table 40-12. A chest roentgenogram (posteroanterior and lateral views), complete blood count, and liver chemistries should be obtained for all patients. The likelihood of identifying metastases by the routine use of scans in the asymptomatic patient is greatly infhb_., enced by the clinical stage of disease. Bone scans are fre-

TABLE 40-12. Breast-Directed History and Physical Examination

History Breast and axillary symptoms: first observed and evolution

Breast mass Nipple discharge: spontaneous or induced, color Nipple and skin retraction Axillary mass Arm swelling or pain

History of prior breast biopsies, cyst aspirations Reproductive history

Age at onset of menses Date Of last menstrual period, regularity of cycles Number of pregnancies, children, abortions Age at first birth Age at menopause History of hormone use

Family history: relationship, age at diagnosis of breast cancer Date of last mammogram Review of systems directed toward evidence of metastases

Physical Examination Breast mass

Size (measured) Location (clock position and distance from areola) Consistency Fixation to skin, pectoral muscle, chest wall

Skin changes Erythema Edema Dimpling Satellite nodules Ulceration·

Nipple changes Retraction Discoloration Erosion Discharge: color, location

Nodes Axillary size, number, fixation Supraclavicular Infraclavicular Arm edema



quently used as a preoperative screening test, but Khansur and coworkers found positive scans in only 2.1% of 187 women with stage I and II disease. 153 Baker reported a similar low incidence of occult bony metastases detected by scanning, with disease demonstrated for only 1.8% of stage I and II patients.154 An incidence of bony metastases of less than 5o/o for patients with stage I and II disease has been reported in several other studies, including a review of ll 18 patients by Lee. 155

-158 False-positive scans are obtained frequently. par

ticularly for older patients. In contrast, positive bone scans are obtained for 20% to 25% of asymptomatic women with stage III breast cancer, making this a worthwhile screening procedure in locally advanced breast cancer. 153

•154

·156

The yield of screening liver scans is even lower than that seen with bone scanning, and the test is of little benefit in the preoperative evaluation of stage I and II breast cancer. Nomura and colleagues reviewed 129 patients with stage I, II, and III breast carcinoma and had a 0% yield from liver scanning. 159

Abnormal liver scans secondary to causes other than metastatic disease are common. Weiner found 12 abnormal scans in 234 patients, but only 4 (1 %) were due to metastatic disease.160 Sears and associates reviewed 100 patients with five abnormal liver scans, only one of which was a true positive finding. 161 Liver scans should be reserved for patients with abnormal liver chemistries, hepatomeg~ly, or significant weight loss suggesting hepatic metastases. Because falsepositive scans are common, histologic confirmation of metastases should be considered before abandoning definitive primary therapy on the basis of an abnormal liver scan.

Bilateral mammography is an essential part of the preoperative workup. Ideally, a mammogram should be obtained before a breast biopsy is done to allow adequate evaluation of the index breast for breast-conserving therapy. In these patieTI.ts, the use of magnification views is recommended to assess the presence and extent of the microcalcifications suggestive of associated DCIS. In patients who will undergo mastectomy, complete visualization of the contralateral breast to exclude an occult synchronous tum.or is necessary. In a report from Guy's Hospital, the use of routine preoperative mammography resulted in ·a synchronous breast canCer incidence of 2.4%, a fivefold increase over prior clinical detection rates.162

The evaluation of serum tumor markers has not been shown to be of benefit preoperatively. Although carcinoembryonic antigen (CEA) may be useful in monitoring response to therapy, it is infrequently elevated in primary breast cancer. Lee found that only 3% of patients with ·stage I breast carcinoma and 6% with stage II disease had CEA levels greater than 5 mg/ml. 163 Other markers, such as assays to identify sialomucin (e.g., CA 15-3, CA 549), are more commonly elevated in primary breast cancer, with abnormalities seen in 20% to 50% of patients. 164

-166 However, 20o/o of patients with benign breast

disease have elevated CA 15-3 levels, and elevated levels are seen in benign gastrointestinal disease, diminishing the usefulness of this marker as a screening test. 166 The value of tumor markers in primary breast cancer has not been clearly established. The available markers have no real utility in the preoperative evaluation of the patient. The identification of a highly specific and sensitive marker, such as a shed tumor product, is currently an active area of research.

STAGING

Staging refers to the grouping of patients according to the extent of their disease. It is useful in choosing treatment for . individual patients, estimating prognosis, and comparing the results of different treatment programs. Staging of breast cancer is performed initially on a clinical basis, according to the physical examination and laboratory radiologic evaluation. The most widely used clinical staging system is the one adopted by the International Union against Cancer (UICC) and the American Joint Committee on Cancer (AJ CC) Staging and End Results Reporting. It is based on the tumor-nodes, _ metastases (TNM) system as detailed in the 1988 Manual for · Staging of Cancer:

T PRIMARY TUMORS TX Primary tumor cannot be assessed TO No evidence of primary tumor Tis Carcinoma in situ: intraductal carcinoma, lobular

carcinoma, or Paget's disease with no tumor Tl Tumor 2 cm or less in its greatest d~mension

a. 0.5 cm or less in gi-eatest dimension b. Larger than 0.5 cm, but not larger than 1 cm in

greatest dimension c. Larger than 1 cm, but not larger than 2 cm in

greatest dimension T2 Tumor more than 2 cm but not more than 5 cm in·

its greatest dimension T3 Tumor· more than 5 cm in its greatest dimension T4 Tumor of any size with direct extension to chest wall

or to skin. Chest wall includes ribs, intercostal muscles, and serratus anterior muscle, but not pectoral muscle. a. Extension to chest wall b. Edema (including peau d'orange), ulceration of

the skin of the breast, or satellite skin nodules confined to the same breast

c. Both of the above d. Inflammatory carcinoma

Dimpling of the skin, nipple retraction, or any other skin changes except those in T 4b may occur in Tl, T2, or T3 without affecting the classification.

N REGIONAL LYMPH NODES NX Regional lymph nodes cannot be assessed (e.g., pre-

viously removed) NO No regional lymph node metastases Nl Metastasis to movable ipsilateral axillary node(s) N2 Metastases to ipsilateral axillary nodes fixed to one

another or to other structures N3 Metastases to ipsilateral internal mammary lymph

node(s)

M DISTANT METASTASIS MO No evidence of distant metastasis Ml Distant metastases (including metastases to ipsilat

eral supraclavicular lymph nodes)


Another clinical staging system, the Columbia Clinical Classification, is Jess widely used but is of historic importance. Like the UICC-AJCC system, patients are grouped according to the extent of disease in the primary tumor site, nodal areas, and distant metastases:

Stage A

Stage B

Stage .C

Stage D

No skin edema, ulceration, or solid fixation of the tumor to the chest wall. Axillary nodes are not involved clinically. No skin edema, ulceration, or solid fixation of the tumor to the chest wall. Clinically involved nodes, but less than 2.5 cm in transverse diameter and not fixed to overlying skin or deeper structures of the axilla. Any one of the five grave signs of advanced breast carcinoma: 1. Edema of the skin of limited extent (involving

less than one third of the skin over the breast) 2. Skin ulceration 3. Solid fixation of the tumor to the chest wall 4. Extensive involvement of axillary lymph

nodes (measuring 2.5 cm or more in transverse diameter)

5. Fixation of the axillary nodes to overlying skin or deeper structures of the axilla

All other patients with more advanced breast carcinoma: 1. A combination of any two or more of the five

grave signs listed under stage C 2. Extensive edema of the skin (involving more

than one third of the skin over the breast 3. Satellite skin nodules 4. Inflammatory type of carcinoma 5. Clinically involved supraclavicular lymph

nodes 6. Internal mammary metastases as evidenced

by a parasternal tumor 7. Edema of the arm 8. Distant metastases

Clinical evaluation of spread to the axilla has high falsepositive and false-negative rates. For this reason, pathologic staging based on histologic study of the axillary specimen is preferable. For the individual patient, prognosis is better determined by pathologic staging than by clinical staging (Table 40-13). For patients who have clinical indications of spread of tumor but negative histologic evaluations, the survival rate (72%) is similar to that of the entire group of patients with histologically negative nodes (76%), not to that of the group with histologically positive nodes (48%).137 Similarly, if a patient does not have clinical evidence of axillary involvement but microscopic involvement is detected pathologically, the survival rate (57%) is similar to that of the entire group of patients with microscopic involvement (48%).

Pathologic stage is commonly given as stage 1 (i.e., axillary nodes not involved) or stage II (i.e., axillary nodes involved). Refinements of this simple staging format have been made, such as subdividing stage II according to the number of positive axillary nodes. Because prognosis is clearly related to the ex-

Staging 1279

TABLE 40-13. Ten-Year Survival(%) According to Clinical and Pathologic Assessment of Axillary Nodes

Patholoflic Assessment

Node Node Clinical Assessment Negative Positive All Patients

NO 77 57 71 NI 72 34 44 All patients 76 48

(Haagensen C. Treatment of curable carcinoma of the breast. Int J Radiat Oncol Biol Phys !977;2:975-980) · ~~

tent of axillary-involvement (see Table 40-7), it has become convention to subdivide axillary involvement into one to three nodes positive or more than four nodes positive. Another refinement is based on the recognition that micrometastatic involvement of axillary lymph nodes is not associated with the poor prognosis seen with macrometastatic involvement. A. comparison of the significance of these two types of axillary metastases has been the object of recent pathologic study. In one study, occult metastases were demonstrated in the regional lymph nodes by an extended histopathologic technique in 24% of 78 cases of invasive breast cancer that would have been regarded as pathologic stage I (i.e., no nodal metastases) after ro_utihe pathologic examination. Patients in whom the largest nodal metastases measured 2 mln or less in the greatest diameter (i.e., micrometastases) were compared with those in whom the lesions were larger than 2 mm (i.e., macrometastases). Life table analysis revealed no significant difference in survival rates between patients with micrometastases and those without nodal metastases, and both of these groups exhibited a significantly greater likelihood of survival than patients with macrometastases. In another study, by Huvos and c_oworkers from Memorial Hospital in New York City, prognosis was related to pathologic extent of axillary nodal involvement.167 For the 62 patients with no involvement of the axillary nodes, the 8-year survival rate was 82% (51 of 62 patients). When micrometastatic involvement (<2 mm) of level I axillary nodes was found, the 8-year survival rate was 94% (17 of 18). The survival rate was 62% (28 of 45) for patients with macrometastatic involvement of level I axillary nodes. Other refinements of the pathologic staging scheme are based on the recognition that extension of metastatic disease beyond the lymph node capsule or involvement of an axillary node larger than 2 cm has been associated with a worse prognosis, independent of the number of nodes involved.

The Postsurgical Treatment Pathologic Classification was developed by the UICC-AJCC in _1988:

PRIMARYTUMOR (pl) pTx Criteria to assess the primary tumor can not be

met pTO pTl-4

No evidence of primary tumor Same as UICC-AJCC classification



NODAL INVOLVEMENT (pN) pNX Regional lymph node metastasis cannot be

pNO pNl pNla pNlb

pN2-3

assessed No regional lymph node metastasis Metastasis to movable ipsilateral axillary node(s) Only micrometastasis (none larger than 0.2 cm) Metastasis to lymph node(s), any larger than 0.2 cm

i. Metastasis in one to three lymph nodes, any larger than 0.2 cm and all smaller than 2 cm in greatest dimension

ii. Metastasis to four or more lymph nodes, any larger than 0.2 cm and all smaller than 2 cm in greatest dimension

iii. Extension of tumor beyond the capsule of a lymph node metastasis smaller than 2 cm in greatest dimension

iv. Metastasis to a lymph node 2 cm or more in greatest dimension

Same as clinical UICC-AJCC class.ification

LOCAL TREATMENT OF BREAST CARCINOMA

The modern era of breast cancer surgery began with Halsted's description of the radical mastectomy in 1894. 168 At that time, breast cancer was thought to begin with a tumor in the breast and to spread in an orderly fashion through lymphatics from the low axillary nodes to the high axillary nodes and then to distant sites. The radical mastectomy incorporated this concept of tumor biology by extirpating the tumor and its draining lymphatics en bloc with a wide margin of normal tissue. Halsted contended that the histologic tumor type, degree of

local spread, patient age, and thoroughness of the operative procedure all influenced surviv'1.l after surgery. 150 Although our understanding of tumor biology has changed considerably since Halsted's time, the surgical treatment of breast cancer that he popularized remained unchallenged dogma for the next 70 years.

Radical mastectomy is the en bloc removal of the breast, the skin overlying the tumor (usually witlJ a 5-cm margin), the pectoralis major and minor muscles afld all of the axillary contents (Fig. 40-6). Skin graft closure of the operative defect is not uniformly required but wa.S advocated by Haagensen to allow more radical skin excision. 134 The radical mastectomy is technically feasible in virtually all women with breast cancer. Of 1640 women seen at Memorial Hospital from 1940 to 1943, 88.9% were thought to have tumors resectable for cure by this technique. 147 In 1974, Adair and Berg reported the 30-year follow-up of the 1458 women in this series who underwent surgical resection. 147 Fifty-seven percent of the patients died from breast cancer, 24% died from other causes, and only 13% survived 30 years free of cancer. Six percent of the total group were lost to follow-up.

Although radical mastectomy does not cure most women with breast carcinoma, it is an effective means of maintaining local control of the primary tumor. Local recurrence on the chest wall or in the axilla is relatively rare after the procedure, occurring in only 6% of 935 patients reported by Haagensen'" and 17% of 704 women in Donegan's series. 170 It is pos~ible that complete local control of the primary tumor in the breast and axilla may play a role in long-term survival for a small number of women as indicated by the observation of Adair and coworkers that 33% of their 30-year survivors treated with surgery alone had positive axillary nodes.147 Similar observations were made by Brinkley and Haybittle, 171 Fentiman, 172 .and Rosen and colleagues, 173 who found that about 25% of 20-year survivors had positive axillary nodes.

A B FIGURB 40-6. ·Radical mastectomy. (A) Extent of diSsection. Notice the vertical orientation of the incision. (B) Postresection anatomy. Notice the absence of the pectoralis major and minor muscles. (Kinne DW. Primary treatment of breast cancer. In: Harris JR, Hellman S, Henderson IC, Kinne DW, eds. Breast diseases. 2nd ed. Philadelphia: JB Llppincott, 1991)

T


Radical mastectomy, in addition to failing to cure a large number of women, is associated with significant long-term morbidity. Limitation of arm elevation and chronic lymphedema are seen in 2q% to 53% of women after this procedure, and re IQ.oval of the pectoralis major muscle leaves a noticeable cosmetic defect on the patient's chest wall. 174·175 Breast reconstruction after radical mastectomy usually requires tissue flaps to provide adequate skin and muscle coverage.

The failure of radical mastectomy to cure many women with breast cancer was thought by some surgeons to be due to its failure to extirpate all of the draining lymphatics of the breast. It was recognized in the 1950s that approximately one quarter of the lymphatic drainage of the breast is through the ipsilateral internal mammary nodes, and this drainage occurs from all quadrants of the breast. 176 In an attempt to address this, the extended radical mastectomy was developed. This operation is a radical mastectomy with en bloc removal of the internal mammary nodes. Nonrandomized reports suggested that s11rvival might be improved in selected patients by this more radical procedure. 177

·178 However, a prospective ran

domized trial reported by Veronesi and V alagussa failed to show a survival difference between women treated with radical and those treated with extended radical mastectomy. 179 Isolated internal mammary node metastases are uncommon, occurring in 4.9% of 7070 patients in whom axillary and internal mammary nodes were sampled. Metastases to this node group seem to have the same prognostic significance as axillary node metastases. 136

·180 The extended radical mastectomy is an op

eration of historic interest, demonstrating that extension of the Halstedian principles of breast cancer surgery does not result in improved survival.

Modified radical mastectomy is now the standard operative treatment for patients with invasive breast cancer in the United States. Surveys by the American College of Surgeons indicate that only 28% of women were treated with modified radical mastectomy in 1972, but by 1981, this figure rose to 72o/o. 181 A survey of breast cancer treatment in New Mexico through 1985 found that 72% of node-negative women and 79% of node-positive women continued to be treated with modified radical mastectomy .182 A concomitant decrease in the number of radical mastectomies performed has been observed, with this procedure accounting for only 3.2% of breast cancer operations in 1981. 181

.

A modified radical mastectomy (i.e., total mastectomy .and axillary dissection) includes removal of the entire breast and some or all of the axillary lymph nodes. The pectoralis minor muscle may be removed or transected, but it is usually preserved (Fig. 40-7). Although it still involves removal of the entire breast, the modified radical mastectomy is less morbid than the radical mastectomy. The incidence of arm edema and shoulder dysfunction is decreased, and the cosmetic defect is less noticeable in a variety of types of clothing (Fig. 40-8) .183 Breast reconstruction is more easily performed after modified radical mastectomy.

Although the modified radical mastectomy may not seem to differ significantly from the radical mastectomy, it represented a major departure from Halstedian principles of cancer surgery, because it is not an en bloc procedure. The switch to modified radical mastectomy occurred as it became increasingly apparent that treatment failure after breast cancer surgery is usually due to systemic tumor dissemination before

Local Treatment of Breast Carcin~ma 1281

surgery rather than an inadequate operative procedure. Several retrospective studies showed similar survival rates for women treated with the radical or modified radical mastectomy .184·185 These findings were confirmed in two prospective randomized trials. Between 1969 and 1981, 606 women with stage I or II breast carcinoma were randomized to treatment with radical mastectomy or modified radi9al mastectomy in Manchester, England. 186 No differences ih disease-free survival, overall survival, or local recurrence rates were seen between the groups. In a smaller trial involving 311 women, Maddox and associates failed to demonstrate any significant benefit in survival or local control for women undergoing radical mastectomy. 187 The NSABP provided further evidence that radical en bloc surgery did not prolong survival with a trial (B-04) in which patients with clinically node-negative breast cancer were randomized to three treatment groups: radical mastectomy; total mastectomy with observation ofthe untreated axillary nodes, and a delayed dissection if positive nodes appeared; or total mastectomy with radiation therapy to the regional lymph nodes.188 No survival differences were found among the groups, despite the fact that approximately 40% of patients in the axillary observation arm were presumed to have histologically positive nodes (as seen in the group undergoing axillary dissection) that were untreated. Although this trial did not directly compare radical mastectomy and modified radical mastectomy, it provided further support for the use of modified radical mastectomy, by demonstrating the limitations in the Halsted notion of cancer spread.

There are few, if any, indications for radical mastectomy. If the pectoral fascia has been violated at the time of biopsy or the tumor seems to abut the fascia or invade a portion of the pectoral muscle, a small portion of the muscle directly beneath the tumor can be excised to obtain a negative deep margin of reseCtion. 189 Large tumors that involve greater amounts of the pectoral muscle are best treated with chemotherapy or radiation therapy as the initial therapeutic modality. In these situations, this multimodal approach frequently allows a standard modified radical mastectomy to be done after a response to the initial therapy.

Total mastectomy, referred to as simple mastectomy, involves removal of the entire breast, including the nipple areola complex, with preservation of the pectoral muscles and the axillary nodes. Because of the importance of axillary dissection as a staging procedure, total mastectomy is not considered a standard surgical approach for the management of infiltrating

· carcinomas. The indications for total mastectomy include patients with ductal carcinoma in situ who elec::t mastectomy (among whom the incidence of axillary node metastases is less than 1 %); women undergoing prophylactic surgery to prevent the development of breast cancer; patients who develop a recurrence in the breast after breast-conserving surgery that had included an axillary dissection, and patients who require mastectomy for local control of tumor in whom information about axillary node status does not influence therapy. This includes women with metastatic disease undergoing "toilet mastectomy" or elderly women with significant comorbidities.

OPTIONS IN LOCAL TREATMENT

The major options for local treatment are modified radical mastectomy and breast-conserving treatment consisting of



FIGURE 40-7. Modified radical mastectomy. (A) Extent of dissection. Notice the horizontal orientation of the incision. (B) The clavipectoral fascia is incised and reflected caudad. The pectoralis minor muscle is divided near the coracoid process. (C) Postresection anatomy. The pectoralis muscles are retracted, not resected. (Kinne DW, DeCosse JJ. Modified radical mastectomy for carcinoma of the breast. Am Surg 1982;48:543-556)

A

B

conservative surgery and irradiation. Breast-conserving surgery involves removal of the primary tumor and a variable margin of surrounding normal breast tissue, usually accompanied by an axillary dissection. A variety of terms, imprecisely defined, are used to describe this approach to the surgical therapy of breast cancer. They include lumpectomy, tumorectomy, segmental mastectomy, and local excision, which imply the removal of a relatively small amount of normal breast tissue, and partial nrastectomy and quadrantectomy, which usually imply the excision of a larger amount of breast tissue.

In the past, excision of less than the entire breast was thought to be contraindicated due to the multicentricity of breast carcinnma. The actual frequency of multicentricity is debatable, with reported incidences in mastectomy specimens ranging from 9% to 75%. 190 The large discrepancy in the reported incidence of multicentricity is related to a variety of factors: differences in the definition of multicentricity, quantitation of invasive and noninvasive carcinoma, variations in

c

the extent of tissue sampling, and different techniques of pathologic examination. It is important to differentiate foci of cancer in the vicinity of the tumor (i.e., multifocality) and totally independent foci of cancer (i.e., true multicentricity). The work of Holland and coworkers described earlier was instrumental in demonstrating that breast cancer is commonly multifocal, but uncommonly multicentric. 129 The clinical importance of multifocal tumor is readily apparent in reviewing the local failure rates of women treated only with excision of the primary tumor. The NSABP found a 39% incidence at 8 years of recurrent tumor in the breast in patients treated with an excision of the primary tumor in which the margins of resection were histologically negative.191 Lagios observed a 19% local failure rate at a mean follow-up of only 24 months in women treated with gross tumor excision without consideration of margin status. 192 Several additional experiences with local excision alone confirm local failure rates in the range of 25% to 35%. 193- 195 In all these series, the site of recurrence in the breast is at or near the site of the primary

T


FIGURE 40-8. Cosmetic results of a radical mastectomy on the patient's left and a modified radical mastectomy on the right.

tumor in most cases. These data emphasize that multifocal breast cancer commonly remains after an excision, even if the margins of resection are histologically negative.

Although it is clear from the preceding data that breast conservation using excision of the primary tumor alone provides inadequate local control, there is now ample data from retrospective and prospective studies indicating that the addition of radiation therapy to tumor excision reduces breast failure rates to the range of 4.4% to 16% at IO years, and survival does not differ from that seen after mastectomy. 191 ·197- 205

Since 1970, there have been seven prospective r"andomized trials using modern radiation therapy techniques in which breast-conserving surgery and radiation therapy has been compared with mastectomy. At the National Cancer Institute of Italy in Milan, 701 women with clinical stage I breast carcinoma ( <2 cm, negative axillary nodes) were randomized to treatment with conservative surgery or radical mastectomy .199

·20° Conservative surgery consisted of a quadrantec

tomy and complete axillary dissection. Radiation therapy was administered to the breast alone through two opposing tangential fields, giving a dose of 5000 cGy in 5 weeks. Another 1000 cGy was given to the tumor site by orthovoltage radiation. After 1975, all patients with histologically positive axillary lymph nodes were treated with 12 cycles of cyclophosphamide, methotrexate, and 5-fluorouracil (CMF regimen). Microscopically involved axillary nodes were found in 25% of the radical mastectomy group and 27% of the conservatively treated group. The latest report of the Milan trial was published in 1990.200 No differences in relapse-free and overall survival rates were observed. This was true for node-positive and nodenegative patients. Recurrence in the ipsilateral breast occurred in 3% of the quadrantectomy patients. There were 19 cases of contralateral breast carcinoma in the quadrantectomy group and 20 in the mastectomy group.

Local Treatment of Breast Carcinoma 1283

The NSABP began a three-arm trial (protocol B-06) in 1976 comparing mastectomy with lumpectomy with or without radiation therapy.'" A total of 1843 evaluable patients with clinical stage I or II carcinoma whose primary tumors clinically . measured no more than 4 cm were entered, but 174 patients

· refused their assigned treatment and were excluded from analysis. All patients underwent axillary dissection. Those with involved lymph nodes received adjuvant chemotherapy. In contrast to the Milan trial, in which an entire quadrant of the breast was removed, a lumpectomy was performed in the NSABP trial that involved resection of the tumor with only enough normal tissue 3.round it to ensure that the microscopic margins of the specimen were tumor free. It was considered impossible to obtain tumor-free margins in 10% of the patients randomly assigned to lumpectomy, and total mastectomy was carried out in these patients. Radiation therapy was delivered to the breast alone with supervoltage equipment using opposed tangential fields, often without wedge filters to compensate for the slope of the breast, to a dose of 5000 to 5300 cGy in 5 to 6 weeks. The regional lymph nodes were not treated, and no boost was given to the tumor site. After 8 years of followup, no differences in distant disease-free survival or overall survival rates were observed between patients undergoing mastectomy and those undergoing lumpectomy with or without radiation therapy. Although the use of radiation therapy did not effect survival, it significantly reduced the incidence of local recurrence in the breast. The probability of a local recurrence 8 years after surgery was only 1 Oo/o for women who received radiation and 39% for those who did not. The

. benefit of radiation was seen in patients with positive or negative riodes.

Additional smaller and more recent trials have been performed by the Institut Gustave-Roussy in Paris,201 the National Cancer Institute of the United States,'02 Guy's Hospital in London,203 the European Organization for Research on Treatment of Cancer,204 and the Danish Breast Cancer Group.205

All of these trials show comparable results for conservative surgery and radiation therapy and mastectomy. The randomized trials comparing these treatments are summarized in Table 40-14.

Based on the information obtained in these randomized trials and from nonrandomized series with long-term followup, a panel of experts at a Consensus Development Conference on the Treatment of Early Stage Breast Cancer convened by the National Cancer Institute and held in June 1990 concluded: "Breast conservation treatment (excision of the primary tumor and adjacent breast tissue followed by radiation therapy) is an appropriate method of primary therapy for the majority of women with stage I and II breast cancer and is preferable because it provides survival equivalent to total mastectomy and axillary dissection while preserving the breast.' '206

BREAST-CONSERVING SURGERY AND AXILLARY DISSECTION

The goal of breast-conserving surgery is to maintain local tumor control in the breast while preserving a good cosmetic appearance. The surgical techniques used for 'breastconserving surgery differ from those employed for mastec-



TABLE 40-14. Modern Randomized Trials of Conservative Surgery and Irradiation Compared With Mastectomy

Dates of No. of ~nvestigations Trial Patients Survival Equivalent

Gustave-Roussy201 1972-1979 179 Yes NCI Milan200 1973-1980 701 Yes NSABP B-06"1 1976-1984 1843 Yes NCI Bethesda202 1980-1986 112 Yes EORTC204 1980-1986 903 Yes Guy's Hospital203 1981-1986 399 Yes Danish Breast Cancer Group205 1983-1987 619 Yes

tomy. Breast-conserving surgery should be done through an incision directly over the tumor. Incisions that are transverse or curvilinear usually give the best cosmetic results (Fig. 40-9). The use of circumareolar incisions, with tunneling through the breast tissue to the carcinoma, should be avoided, because this approach makes control of margins more difficult and increases the amount of tissue that must be treated with a boost ofradiation therapy. A crucial step in breast-conserving surgery is avoidance of thin skin :flaps.207 Preservation of the subcutaneous fat beneath the incision and the breast tissue superficial to the tumor helps maintain normal breast contour. Attempts to obliterate the lumpectomy cavity with sutures and the use of breast drains worsen the cosmetic result by altering breast contour.

The extent of breast tissue that should be excised is the subject of some debate. The NSABP reported a 10% incidence of recurrence in the breast 8 years after excision with histologically negative margins.191 Margins were considered negative if tumor cells were not directly at inked resection margins, regardless of how close to the margins the cells were. The lowest rates of local failure are reported by Veronesi and colleagues, who observed a 2.8% incidence of recurrence at the primary tumor site and a 1.6% incidence of new tumors elsewhere in the breast after treatment with quadrantec-

FIGURE 40-9. Breast-conserving surgery and axillary dissection. Notice the use of separate incisions. (Kinne DW. Primary treatment of breast cancer. In: Hanis JR, Hellman S, Henderson IC, Kinne DW, eds. Breast diseases. 2nd ed. Philadelphia, JB Lippincott, 1991)

to my. 197 A randomized trial comparing quadrantectomy to tumor excision with a margin of 1 cm of grossly normal breast tissue was conducted at the National Cancer Institute of Italy in Milan between 1985 and 1987.208 After quadrantectomy, patients received external-beam irradiation alone, but women randomized to tumorectomy received an iridillm 192 implant and external-beam therapy. Local recurrence was less common in patients treated with quadrantectomy (2.5%) than patients treated with more limited excision (7.2%). A significant cosmetic advantage was observed for patients treated with Jumpectomy, even though quadrantectomy was performed by surgeons with extensive experience in the technique. At the time of latest follow-up, survival was the same in the patients treated with quadrantectomy and with excision. These results are consistent with the findings of Holland and associates.129 Because the probability of finding additional microscopic foci of cancer decreases as a function of distance from the primary tumor, the removal of larger amounts of breast tissue would be expected to decrease local failure rates.

The extent of breast resection appears to be only one of several variables that influence the likelihood of local recurrence after breast-conserving surgery and radiation therapy. The most important of these appear to be the presence of an extensive intraductal component (EiC), patient age at diagnosis, and the use of adjuvant systemic therapy. An EiC is defined for an infiltrating ductal carcinoma in which DCJS is prominently present within the tumor (typically greater than 25%) and present in grossly normal adjacent breast tissue or in which the t\lmor is composed primarily of DCIS with areas of focal invasion. Data from the Joint Center for Radiation Therapy (JCRT) series indicate that 28% of patients with infiltrating ductal carcinoma have an EiC. Local recurrence occurred in 27% of EiC-positive tumors but only 8% of EICnegative tumors.209 The effect of an EiC on local recurrence decreases as the extent of breast resection increases. One study reported a decrease of local recurrence rates in EiCpositive tumors from 36% to 9% as the volume of breast tissue resected increased from 35 to 74 cm3

.210 This result is con

sistent with a pathologic study of mastectomy specimens by Holland and coworkers, in which an EiC predicted the likelihood of finding extensive multifocal involvement beyond the edge of the tumor.211 Age less than 35 to 40 years has been associated with a substantially increased risk of local failure in several series.197·209·

212·213 In part, this is explained

by an association between young patient age and the presence of an EIC.198·

214

T


The use of adjuvant chemotherapy or hormonal therapy in conjunction with radiation therapy appears to reduce the incidence of local failure in the breast. In the NSABP trial (B-13) testing the value ofmethotrexate and fluorouracil in nodenegative, ER~negative women, patients randomized to -the treatment arm had a 0.9% rate ofrecurrence in the ipsilateral breast after 4 years of follow-up, compared with 3. 7% for women who did not receive cytotoxic therapy. 215 Similar results were seen in a trial (B-14) testing the value of tamoxifen in n~de-negative, ER-positive patients treated with tamoxifen had a 0% recurrence rate in the ipsilateral breast, compared with 2.2% in the control group.216 This finding is supported by the data from NSABP trial B-06, in which node-positive ~omen, all of whom received chemotherapy, had a 6% rate of failure in the breast after lumpectomy and irradiation, compared with 12% for node-negative women undergoing the same local treatment with no systemic chemotherapy.191

The value of obtaining histologically negative margins of resection in preventing local recurrence is unclear. The lack ofa universally accepted definition of what constitutes an involved margin, problems with sampling breast specimens, ancj the presence of additional foci of cancer in the breast, even if margins are negative, limit the reliability of margin status as a predictor of local failure. 217

•218 Veronesi was unable to

correlate margin status and local failure in his tumorectomy patients. 208 The 10% incidence of local recurrence in NSABP trial B-06, in which tumor-free margins were required, indicates that negative margins alone do not ensure local control after conservative surgery and radiation therapy. 191

The available data suggest that the amount of breast tissue that must be excised in breast-conserving procedures needs to be individualized. Although the lowest local failure rate is obtained with quadrantectomy, it is at the expense of a lesser cosmetic result. For most women, eXcision of the primary tumor with a 1- to 2-cm margin reduces the tumor burden to a point where it can be controlled with radiation therapy. Important factors in determining the adequacy .of surgical resection are the presence and extent of tumor at the inked resection margins, the presence or absence of an EiC, and the patient's relative concerns about cosmetic outcome and local recurrence.

Until recently, axillary dissection was considered an essential aspect of local treatment, whether by mastectomy or by a breast-conserving approach. The principal purpose of axillary dissection is to provide prognostic information that can direct further therapy. Axillary dissection is effective in obtaining local control in the axilla and simplifies the technique of irradiation in patients treated with breast-conserving treatment. However, because of the widespread use of adjuvant systemic therapy, even if axillary nodes are histologically negative, the role and the conduct of axillary dissection are being questioned.

If an axillary dissection is to be performed, it can be the classic axillary dissection in which nodal tissue at all three levels is removed or a more limited dissection of levels I and II. As nodal dissection has come to be regarded as a staging procedure; rather than a therapeutic one, a more limited dissection has become increasingly popular in the United States. Many surgeons think that a complete three-level axillary dissection should be carried out if the axillary nodes are grossly positive or if knowledge of the exact number of positive nodes


influences treatment, because there is a 20o/o to 30% chance that additional positive nodes will be identified at level III if metastases are present in the low axillary nodes.144·145•219·220

Major complications of axillary dissection are infrequent and include injury or thrombosis of the axillary vein and injury to the motor nerves of the axilla. Minor complications are reported in 8% to 13% of patients and include wound infection, seroma formation, mild shoulder dysfunction, loss of sensation in the distribution of the intercostobrachial nerve, and edema of the breas.t and arm. 221 ·222 The extent of axillary dissection correlates with the degree of breast edema seen after breastpreserving surgery and the risk of arm edema. Clark and coworkers found a 6% incidence of breast edema if no axillary surgery was done, compared with a 25% incidence after axillary sampling and a 79% incidence after complete dissection. 233 The incidence of lymphedema of the arm after axillary dissection varies with the definition of lymphedema, but an incidence of 6o/o to 8% is commonly reported. 183·224- 226 F8.ctors that contribute to the development of lymphedema are radiation therapy to the dissected axilla, postoperative wound com: plications, subsequent cellulitis of the arm, obesity, the extent of axillary dissection, and advanced age. 174·224·

227·228 The mor

bidity of axillary dissection can be minimized by the use of appropriate surgical techniques. Avoiding unnecessary dissection superior to the axillary vein and preservation of th~ fat and lymphatic tissue surrounding the vein decreases edema · of the arm and breast. The identification of the long thoracic nerve minimizes the ris~ of injury to this structure. In the clinically negative axilla, it is reasonable to preserve the thoracodor~al neurovascular bundle and the intercostobrachial nerve.

The appropriateness of eliminating axillary dissection depends on the institutional approach to the use of adjuvant systemic therapy. Women who would not be candidates for adjuvant therapy unless positive nodes were identified, such as those with invasive tumor lesS than 1 cm in diameter and those participating in clinical trials, continue to require axillary dissection. If different chemotherapy is given depending on the presence and extent of nodal involvement, axillary dissection continues to be necessary. If decisions regarding adjuvant therapy are being made independent of nodal status, the purposes of axillary dissection are for local control and to provide prognoStic information. In the woman who is undergoing breast-conserving treatment, radiation therapy to the axilla is an equally effective method of maintaining local control. In women undergoing mastectomy, axillary dissection remains the preferred method of achieving local control.

IRRADIATION IN BREAST-CONSERVING TREATMENT

The technical details of irradiation are important in obtaining satisfactory results using breast-conserving treatment. It is important that the techniques of surgery and irradiation be coordinated. In general, the larger the surgical resection, the less intensive the irradiation needs to be. The optimal combination of surgery and irradiation to achieve the dual objectives of local tumor control and preserv'ation of the cosmetic appearance is controversial and varies from patient to patient, depending on the extent, nature, and location of the tunior, the patient's breast size, and the patient's relative concerns



about local recurrence and preservation of the cosmetic appearance,

There is a consensus about some elements in the technique of irradiation. It is considered standard to require patient immobilization, treatment ~imulation, treatment planning with the use of tissue compensators to improve dose homogeneity, and supervoltage equipment (i.e., 4, 6, or 8 MV or "Co with 6 MV preferred for dose homogeneity and skin sparing). More sophisticated treatment planning is now available to assess three-dimensional dose distributions and to account for the lower density of lung tissue in the treatment field, but the impact of these developments on patient outcome has not been demonstrated, and they are not considered standard at this time.229 (In standard treatment planning, the lung is considered to have unit density; the use of actual lung density results in a small alteration in the dose distribution.) Each field should be treated daily, and bolus should not be used. To minimize the risk of radiation pneumonitis, not more than 3 cm oflung (as projected on the beam radiograph at isocenter) should be treated.'" The dose to the breast should be in the range of 4500 to 5000 cGy given at 180 to 200 cGy per day for 5 days each week.

There is controversy about the need for delivering additional dose to the primary. site (i.e., boost) and for treating nodal areas in addition to the breast. The rationales for using a boost are that.histologic studies show that residual cancer after resection of the primary tumor is typically in the vicinity of the primary site, recurrences after treatment are most commonly seen at or near the primary site, and boost treatrnent can be delivered without significant morbidity. Although boost irradiation is usually employed in this country, it is likely that patients treated with more extensive breast resections in whom the margins of resection are clearly negative do not benefit greatly from the use of a boost. It is generally agreed that a boost should be employed in patients with focally positive margins of resection. Boost irradiation is typically delivered using an electron beam or interstitial implantation, and the dose to the primary site is increased to approximately 6000 cGy. There is not a consensus about the advisability of treating nodal areas with irradiation, and this needs to be considered in relation to the use and extent of axillary dissection. In patients who do not undergo axillary dissection, the axillary and supraclavicular regions .are usually treated with irradia- · tion. There is genera] agreement that axillary irradiation should not be given after a complete axillary dissection, that overlap between adjacent fields (if used) should be avoided, and that a photon-only "hockey stick" (used to treat the internal mammary nodes) should be avoided. The use of supraclavicular irradiation in patients with positive axillary nodes is reasonable to decrease the risk of recurrence in this area, but it must be balanced against the small increase in complications. 231 The major trials providing the justification for breast-conserving treatment (i.e., NSABP B-06 and Milan trials) were conducted using breast irradiation alone, establishing this technique as a reasonable option in all patients (after axillary dissection).

PATIENT SELECTION FOR LOCAL TREATMENT

Patient selection is based on the results of patient evaluation. Critical elements in patient evaluation are history and physical

examination, mammography, histologic assessment of the resected breast specimen, and assessment of the patient's wishes. There appears to be a consensus about the guidelines for an adequate mammographic evaluation. 232 This includes dedicated mammography in an approved unit, ideally with magnification views of the primary site to determine the presence_and extent of associated microcalcifications. In patients with mammographic microcalcifications, if there is any doubt about the adequacy of the resection after review of the original and specimen mammograms, a follow-up mammogram with magnification views is recommended. The use of high-quality mammograms can identify about two thirds of patients with EiC-positive tumors.233

The surgeon should orient the resected breast specimen by the use of sutures or other means, and the pathologist should ev.alUate the margins of resection for tuffior involvement by the use of inking or some other method. A careful gross description is essential to detail the size of the specimen, the size of the tumor, and the proximity of the tumor to the mar

. gin. The pathology report should mention the presence of microscopic and gross involvement of the margins, although no established definition of "involved margins" has been demonstrated to predict local recurrence after breastconserving treatment.

Perhaps the most difficult aspect of patient evaluation is the assessment of the patient's wishes about breast preser- · vation in the face of having just been diagnosed with a potentially lethal disease. The patient and her physician need to discuss the implications of the options of mastectomy and breast-conserving treatment with regard to long-term $urvival, the possibility and consequences of local recurrence, the treatment options if local recurrence occurs, psychological adjustment, cosmetic outcome, sexual adaptation, and functional competence, and the guidelines, effectiveness, and cost of follow-up procedures. For most patients, this decision does not have an impact on the likelihood of survival, but it may influence the quality of life.

Randomized clinical trials have demonstrated equivalent survival for patients treated with mastectomy and breastconserving treatment, but there are some absolute and some felative contraindications to breast-conserving treatment. An ongoing pregnancy is an absolute contraindication to the use of breast irradiation. It is feasible in many cases to perform breast-conserving surgery in the third trimester and deliver 'breast irradiation after delivery. Women with two or more gross cancers situated in separate quadrants of the breast or those with diffuse malignant-appearing microcalcifications are not considered candidates for breast-conserving treatment. Another absolute contraindication is a history of prior therapeutic irradiation to the breast region that would necessitate retreatment to an excessively high total radiation dose to a significant volume. For example, breast-conserving treatment is not feasible in a woman with a history of Hodgkin's disease treated with mantle irradiation. A history of collagen vascular disease is considered a contraindication to breast-conserving treatment because of the reports indicating that these patients tolerate irradiation poorly.

Tumor size is not an absolute contraindication to breastconserving treatment, although there is little published experience in treating patients with tumors larger than 4 to 4.5 cm. A relative contraindication is a large tumor in a small

r

,,___] 76 of 122 Celltrion, Inc., Exhibit 1006

breast, for which adequate resection would result in significant cosmetic alteration. Breast size can be a relative contraindication. Treatment of women with large or pendulous breasts is feasible if there is sufficient expertise to ensure reproducibility of patient set-up and the availability of 6 to IO MeV radiation to obtain adequate dose homogeneity. The location of the tumor in the breast may be a consideration in the choice of local treatment. In a patient with a tumor beneath the nipple, an adequate resection usually involves removal of all or part of the nipple-areola complex. Whether this is preferable to mastectomy needs to be assessed by the patient and -her physicians. The use of breast-conserving treatment in an elderly patient, especially with impaired performance, is generally contraindicated because of poor tolerance of the treatment.

There ls still some controversy about whether certain pathologic and clinical features influence the likelihood of recurrence after breast-conserving treatment and therefore should affect the choice of local therapy. These include young patient age ( <35 years), focal involvement of the margins of resection, and an extensive intraductal component. It is controversial whether a breast that is difficult to evaluate by mammography and physical examination should influence the choice of local therapy. Further research is required to evaluate the significance of these factors in the selection of patients.

There are certain clinical and pathologic features that should not mitigate against the use of breast-conserving treatment. These include the presence of clinically suspicious (and mobile) axillary lymph nodes or microscopic tumor involvement in axillary nodes. The changes in the breast after conservative surgery and irradiation do allow for the detection of local recurrence at an early stage using physical exainination and mammography, and the use of irradiation in this setting has not been shown to result in a significant risk of second tumors in the treated area. '

BREAST RECONSTRUCTION AFTER MASTECTOMY

Breast reconstruction is an important option for women undergoing mastectomy for the treatment of breast cancer. Reconstruction may be done at the time of mastectomy or as a secondary procedure, .but the possibility of reconstruction should be discussed with the patient before definitive surgery. Reconstruction should be offered to all women who are not candidates for or do not desire breast-conserving surgery with radiation therapy. The indication for breast reconstruction is the patient's desire for the _operative procedure. The only true contraindications are significant comorbid conditions that would interfere with the patient's ability to tolerate a longer operative procedure in the case of immediate reconstruction or additional procedures in the case of delayed reconstruction. Patient age, the need for adjuvant chemotherapy, or a poor long-term prognosis are not contraindications to reconstruction. In the past, there were major concerns that the performance of breast reconstruction would obscure local recurrence after mastectomy, and the .Procedure was restricted to women with early stage carcinoma, usually after a delay of 2 or more

· years from the time of extirpative surgery. As our knowledge of the pathophysiology of breast cancer

has inCreased and as more women are diagnosed with early


stage breast cancer, there has been an increasing trend toward performing breast reconstruction at the time of mastectomy. Several nonrandomized studies of immediate reconstruction have demonstrated that the incidence of local recurrence after the procedure is similar to that after mastectomy alone. 234- 237 Local recurrences after reconstruction were detected at an early stage and, in many cases, treated without loss of the reconstructed breast. The advantages of an immediate reconstruction include avoidance of an additional op-_ erative procedure, decreased psychological trauma, and improved coordination of the efforts of the oncologic and reconstructive surgeons to produce an optimal cosmetic result without compromising the oncologic aspects of the surgery.238

The potential disadvantages of immediate reconstruction are a possible ii1crease in the number of surgical complications and a delay in the institution of adjuvant therapy. Vinton and colleagues compared the incidences of infection, lfematoma, seroma, and epidermolysis among 305 women who underwent modified radical mastectomy and 90. who underwent modified radical mastectomy with immediate reconstruction during the same time period.239 No increase in the frequency of complications was found in the reconstruction group. Eberlein and colleagues reported 199 consecutive patients treated with immediate reconstruction, including 109 with flaps and 90 with implants.240 Thirty percent of the reconstruction patients. underwent chemotherapy, and no delays in initiating treatment were found, a finding similar io that of Frazier.234

The two techniques of reconstruction involve the use of implants and the use of ti~sue from elsewhere in the body (Table 40-15). In January of 1992, the Food and Drug Administration (FDA) declared that silicone gel-filled implants could not be used until more information is available about their long-term safety. This moratorium does not apply to saline-filled implants. In February, a panel of experts appointed by the FDA recommended that silicone gel-filled implants be allowed in breast cancer patients pending the results of further study, and this was approved by the FDA in April of 1992. The complications associated with implants are discussed later. The availability of adequate skin coverage, the size and shape of the patient's contralateral breast, and the patient's cosmetic expectations influence the choice of reconstructive technique.

Until recently, the most commonly used technique for reconstruction was the placement of a silicone gel implant beneath the pectoralis major muscle. This approach is best suited for women with small or moderate-size breasts with minimal ptosis, and requires adequate skin to cover an implant of a size similar to the contralateral breast. The use of limited skin excision, with operative exposure gained by the incision, usually leaves enough skin to cover an implant. Oncologic surgeons agree that the only skin that it is necessary to excise for reasons of cancer control is the nipple-areola complex and the biopsy scar. If insufficient skin is available to achieve symmetry with the contralateral breast or for larger or ptotic breasts, a tissue expander may be employed. This technique involves placement of a prosthesis that is only partially inflated beneath the pectoral muscle. Using a subcutaneous injection port, the prosthesis is gradually filled with saline over a period of weeks to months until the des.ired breast size and shape are achieved.

Many questions have been raised about the safety of silicone



TABLE 40-15. Types of Reconstruction After Mastectomy

.Type Advantages Disadvantages

Local Tissue (plus implant) Submuscular implant Simple, single stage Lack of prqjection, little

natural ptosis · Tissue expansion Gradual recruitment and

stretching of skin, improved projection

Many postoperative visits, complications, time, second-stage surgery to place permanent implant

Permanent expander Expander is also permanent prosthesis, second stage minor

Many postoperative visits, complications, time

Distant Tissue (with or without implants) Myocutaneous latissimus Dependabl!3, nice Donor site scar, turn

patient during surgery Big operation (doubles

dorsi contour axillary fold Transverse rectus abdominis Excellent bulk

myocutaneous (TRAM) flap Rarely need !mplant Abdominoplasty

mastectomy time), wound-healing problems, vascularity, hernias

Microvascular free flaps Bulk depends on donor site

5% failure rate, long and complicated surgery (2 teams), donor site deformity co~parable to myocutaneous flaps

breast implants. Silicone implants have been available for the past 30 years, and approximately 2 million women have had implants for reconstruction or augmentation.241 The major recognized complication of this procedure is fibrous encapsulation or capsular contracture, w:Pich is the formation of excessive scar tissue around the implant, possibly leading to deformation of the appearance and pain. There is no convincing data to substantiate the idea that silicone breast implants increase br~ast cancer risk in humans, although it is true that women with breast implants have not been systematically followed. 241

-244 Two epidemiologic studies did not

demonstrate an increased incidence of breast carcinoma in women with silicone implants.242·243 Review of the Surveillance, Epidemiology, and End Results program (SEER) data from 1973 to 1986 did not reveal an increase ·in the number of breast sarcomas, a neoplasm that can Pe induced in rats with silicone implants.244

There is concern that leakage of silicone from implants may lead to collagen-vascular conditions. Over the past 26 years, there have been ·more than 100 cases reported of women who developed connective tissue disease in conjunction with the use of free silicon~ injections or silicone breast prostheses.245-248 However, only one third of the reported cases have occurred in women with silicone breast implants, and it is uncertain whether the observed numb~r of cases of connective tissue diseases in women with breast implants exceeds the number expected by chance. Based on the lack of good long-term information on the safety of silicone implants, the FDA concluded that the implants should not be used outside of a clinical trial until there is better do~umentation of their safety. Women with these implants have been advised that there is no information to indicate that implants should be removed, but that they should be monitored for any problems. Further research is needed to clarify this issue.

Another technique of reconstruction is the use of myocutaneous flaps to transfer skin, fat, and muscle from distant parts of the body. The most commonly used flaps are the laHssimus dorsi and transverse rectus abdominis (TRAM) myocutaneous flaps. The use of a flap for reccinstruction requires a more lengthy operative procedure than the implant method, and postoperative recovery is somewhat longer because of tw6 separate incision sites. The latissimus flap is often used in conjunction with a prosthesis because the flap alone provides insufficient bulk to achieve symmetry in most cases. The survival rate for this flap is high, with only a 1 % incidence of complete flap loss reported in most series.249 The TllAM flap usually allows an adequate breast mound to be fashioned without the use of a prosthesis, but its blood supply is more tem1ous than that of the latissimus flap, with major necrosis reported in 5% of patients and partial necrosis in as many as 31 o/o of patients.250·251 If these myocutaneous flaps are not available or not suitable for use, it is possible to transfer composite tissues from distant sites and to perform a microvascular anastomosis to nearby vessels. 252 This technique, known as a free flap, requires a skilled microsurgeon and prolonged operating time and is only occasionally chosen for primary rcconstruction.249 The potential benefits and complications of the various recon~tructive procedures are listed in Table 40-15.

Regardless of the technique of reconstruction chosen, the creation of a breast mound is the chief goal in breast reconstruction. Surgery on the contralateral breast, such as reduction or mastopexy, may be required to achieve symmetry. Reconstruction of a nipple-areolar complex is another secondary proc~dure that some patients elect to improve the cosmetic appearance. The patient's own nipple should not be used for this purpose, because recurrent carcinoma due to persistence of breast tissue on the nipple has been re-


ported.253-

255 Microscopic involvement of the nipple is seen in 30% of mastectomy specimens but is frequently not apparent at the time of gross pathologic examination. 256

- 258 The . nipple can be reconstructed using a variety of local flap techniques or by the use of full-thickness skin grafts.'" Tattooing of the grafts produces a color match to the patient's own areola, and allows any site to be used as the donor.259 Tissue from the contralateral nipple should not be used for njpple reconstruction because of the concern of transferring breast tissue to ihe reconstruction site. ·

POSTOPERATIVE IRRADIATION

The first major use of radiation therapy in the primary management of :patients with breast cancer was as an adjuvant to radical mastectomy. There were two rationales for prophylactic radiation therapy. The first was that prophylactic radiation therapy could be used to reduce the risk oflocoregional tumor recurrence. The overall risk of local recurrence after mastectomy is 10% to 15% and is related to whether axillary ppdes are negative or positive. One purpose of postoperative i-adiation therapy is to decrease this risk of local recUrrence. It has been argued that prevention of local recurrence is impo.rtant because these recurrences are often distressing for patients arid because clinically manifested recurrences can be treated effectively in only approximately 50% of patients. The second rationale for postoperative radiation therapy is to improve the likelihood of survival.

Postoperative radiation therapy markedly decreases the risk of locoregional recurrence, but the effect of adjuvant radiation therapy on survival remains uncertain. The survival value of postoperative radiation therapy ideally would be determined by large, properly conducted, prospective, randomized clinical trials. The trials testing the value of postoperative radiation therapy are shown in Table 40-16. It should be stressed that the number of patients entered into these trials is relatively small (i.e., small or moderate benefits may be missed) and that the trials before the late 1960s used techmques of irradiation that did not adequately treat the target volume (i.e., the chest wall anq regional nodal areas) and gave excessive doses to critical structures, such as the heart. In t!ie Stockholm trial, the most modern of the trials using patients without adjuvant systemic therapy, there was a long-term improvement in relapse-free survival after postoperative irradiation, but the improvement in survival was only marginal (p = Q.09) and restricted to the subset of node-positive patients. in nodenegative patients, there was no benefit.

It is possible that postoperative irradiation may be detrimental to survival: An overview of postoperative radiation therapy by Cuzick and associates published in 1987 included the results from the Manchester trials, the Oslo trials, the Stockholm trial, and an unpublished small trial from Heidelberg. 260 No difference was seen in mortality comparing patients treated with or without radiation therapy over the first 10 years after surgery. After 10 years, there was a lower rate of survival associated with the use of radiation therapy (p = 0.005). This decrease in survival was primarily due to an increase in late j::ardiac deaths and was evident in patients treated for left-sic\ed breast cancer in the trials that gave large doses of radiation to the heart. In a recent retrospective analysis of cause-specific mortality in the Stockholm trial, in which cardiac doses were calculated in relation to the patient's lat-

Treatment of Specific Problems in Breast Cancer 1289

era.Uty and the technique employed, a clear relation between cardiac dose and late cardiac death was demonstrated. With a mean follow-up time of 16 years, patients who received low or intermediate cardiac doses had a 20% reduction in breast cancer deaths anc\ no increase in cardiac deaths. This study provides good evidence that late cardiac mortality can be avoided using proper techniques of irradiation.

The use of postoperative radiation therapy needs to be reconsidered in patients who receive adjuvant chemotherapy. It is possible that the chest wall and regional nodes are the sites of greatest tumor burden after mastectomy in certain subgroups, such as patients with larger primary tumors or positive axillary nodes. According to the Goldie-Coldman hypothesis, spontaneous mutations of tumor cells to drug resistance accounts for failures of chemotherapy. and the likelihood that drug-resistant cells will emerge is dire.ctly related to the tumor burden. In theory, by decreasing the local tumor burden, adjuvant radiation therapy may decrease the probability of drug resistance and increase the probability of cure. The use of adjuvant chemotherapy by itself decreases the rate oflocal recurrence, but to a lesser degree than after irradiation. In a randomized trial from sweden that compared the adjuvant use of CMF chemotherapy and irradiation, the local recurrence rate for patients treated with ·irradiation was approximately one half that of patients treated with chemotherapy.••• Table 40-16 shows the results of trials ip which patients were randomized to receive irradiation or to not receive irradiation in conjunction with adjuvant systemic therapy. The largest of these is from the Danish Breast Cancer Cooperative Group, in which patients with high-risk breast cancer (i.e., T3 or T4 or with positive nodes) were randomized to various adjuvmt therapies. Jn an abstract of the results at 7 years, the local recurrence rate was 32% among the 737 premenopausal patients treated with CMF and only 9% among the 736 premenopausal patients treated with irradiation (p < 0.0001), and the corresponding survival rates were 61 % for CMF and 57% for irradiiitfon (p < 0.05).26

'

The role of postoperative irradiation is not established. The modern trials suggest that the treatment can be given safely and is associated with a modest benefit in the survival ofnodepositive patients, particularly if used in conjunction with adjuvant chemotherapy. Further research is required to substantiate. these observations, but many physicians favor the use of postoperative irr~diation to prevent locoregional recurrence, particularly in patients with four or more positive nodes. In patients to be .treated with chemotherapy and irradiation, the optimal sequencing of these modalities is not determined. The results of the Dana-F:µ-ber Cancer Institute/ JCRT trial indicate that radiation therapy can effectively be added at the completion of adjuvant chemotherapy, avoiding any possible interference with the administration of the chemotherapy. 263

TREATMENT OF SPECIFIC PROBLEMS IN BREAST CANCER

AXILLARY METASTASES WITH AN OCCULT BREAST PRIMARY

Breast carcinoma presenting as an axillary metastasis with no appare11t primary tumor is an uncommon problem that was described by Halsted in 1907, and it remains a treatment


\

I

1J 80 of 122 Celltrion, Inc., Exhibit 1006

dilemma today. This entity accounted for 0.35% of 10,014 patients treated for primary breast carcinoma at Memorial Sloan-Kettering Cancer Center from 1975 to 1988.264 A similar incidence of 0.4% of 12,000 breast cancers treated over a 20-year period at the National Cancer Institute in Milan was reported by Vezzoni and associates. 265 The mean age of women with occult primary tumors is between 50 and 54 years, which does not differ significantly from that of women with clinically detectable primary carcinomas.265

-267

Axillary adenopathy due to metastatic adenocarcinoma may be secondary to a variety of primary tumors, but in women, the breast is by far the most common primary site. After a diagnosis of adenocarcinoma is obtained, the radiologic evaluation should be confined to high-quality bilateral mammography, a chest radiograph, and a bone scan to exclude metastases before definitive therapy. There is uniform agreement that additional radiologic studies looking for another primary tumor site are unrewarding if there are no symptoms and should not be done. 264- 269

The sensitivity of mammography in the identification of the occult primary lesion is low. Of 36 women undergoing mammography in Vezzoni's series, only 360/o had tumors detected.265 In the M.D. Anderson Cancer Center experience, only three possible mammographic abnormalities were identified in 40 women with axillary adenopathy due to a presumed occult breast primary. 266 Of 24 patients reviewed by Kemeny, 268 only 1 woman had a mammogram suggesting carcinoma, and Baron and coworkers264 reported a 29% sensitivity for mammography in detecting occult primary tumors. These reports, using modern mamll'1:ographic techniques and equipment, suggest that although abnormalities on mammogram should be aggressively pursued, the absence of a mammographic abnormality does not exclude the presence of a primary breast tumor.

When an axillary node biopsy of an occult primary tumor is done in the clinical setting, the pathologist should be alerted to this fact so that appropriate diagnostic studies can be obtained. Tissue should be sent in a fresh, nonpreserved state to the pathology laboratory and portions retained for hormone receptor determination, immunohistochemistry, and routine light microscopy. Iglehart and coworkers reported the utility of using ele~tron microscopy in a study of 5 patients. 269 However, no features that clearly established the breast as the primary tumor site were identified. Bhatia and coworkers obtained hormone receptor studies for the axillary nodes of 11 women with occult primary breast carcinoma and identified positive estrogen or progesterone receptors in 7 patients, unequivocally establishing the breast as the site of the primary lesion.270 Others have reported receptor positivity in 62%, 50o/o, and 40% of patients. 264

'268

·271 A negative receptor analysis

does not rule out the breast as the site of the primary. Most women with axillary node metastases secondary to a

presumed breast primary have been treated with mastectomy. Table 40-17 summarizes the results of mastectomy in 228 patients, and a primary tumor was identified in 64%. The highest rates of primary tumor identification were seen in some of the older reports from a time when high-quality mammography was not routinely available. Kemeny268 and Ellerbroek,266 reviewing case material in which virtually all patients had mammography, found 45% and 8%, respectively, of mastectomy specimens to contain a primary.


TABLE 40-17. Identification of Primary Tumors in Mastectomy Specimens From Patients With Occult Breast Cancer

Five-Year No. of No. of Cancers/ Survivors

Investigations Patients No. of Mastectomies (%)

Ashikari272 42 22/34 (65%) 79 .Baron264 35 22/33 (67%) 75 Ellerbroek266 42 1/13 (8%) 72 FittS275 13 11/13 (85%) 71 Haagensen128 18 13/14 (93%) 57 Kemeny268 20 5/11 (46%) Not stated Owen571 25 15/25 (60%) 50 Patel267 29 16/29 (55%) 28 Yezzoni265 49 33/44 (75%) 84 Westbrook273 18 9/12 (75%) 61

Total 147/228 (64%)

The size of the carcinomas identified in the mastectomy specimen varies widely. DCIS was the only histologic evidence of malignancy in several cases.264

•267

•268

·272

·273 This finding may

be secondary to sampling error with failure to identify a small invasive carcinoma or invasion through the basement membrane identifiable only by electron microscopy.274 Vezzoni found that 30% of pathologically identified carcinomas were I to 2 cni in diameter.265 Forty-five percent of the 22 cancers reported by Baron and colleagues'" were multifocal, and Rosen271 reported a median tumor size of 1.5 cm (range, 0.1-6.5 cm) in his series. Infiltrating ductal carcinoma is the most comnion histologic tumor type seen. 128

•272

•273

•275

With the increasing use of breast preservation in the management of breast carcinoma, several groups have applied these techniques to occult breast cancer.266

·268

·276

·277 Theoret

ical objections to this approach include the fact that the tumor burden in the breast may be extensive, even if the disease is clinically occult, and that it is not feasible to deliver a boost dose of radiation therapy to the primary tumor site. Thirty patients were treated at the Institut Curie with radiation therapy to the breast and axillary, supraclavicular, and internal mammary fields to a mean midbreast dose of 6200 cGy. 277

Eleven (37%) women developed breast recurrences. Local recurrence occurred as an isolated event in 4 women, in conjunction with nodal recurrence in 4, and as part of systemic failure in 3. The mean time to breast recurrence was 9.3 years, and in another report from this group, only 1 of 3 patients alive at 15 years after treatment retained her breast. 276 Ellerbroek reported 29 patients in which a breast-preserving approach was employed. 266 Sixteen patients had radiation therapy using a mean breast dose of 5170 cGy, producing a 17% 5-year actuarial breast failure rate. An additional 13 woman had an axillary biopsy with irradiation of the axilla only. Seven of the 13 developed clinical primary breast tumors at a mean follow-up of 27 months, and the 5-year actuarial risk of breast failure in this group was 57%. No apparent survival differences between patients treated with mastectomy and breast preservation were observed in these studies, a finding supported by the reports of Kemeny268 and Baron264 for



small numbers of women treated with a breast-preserving approach.

Overall, survival for women with occult ·primary tumors does not differ greatly from the survival of patients with the same extent of axillary involvement and a known primary. This is shown fo Figure 40-10. A difference in survival is not observed between patients in whom the histologic tumor is identified in the breast and those in whom tumor can not be documented, suggesting that the breast is the primary tumor site in both cases. Adjuvant chemotherapy should be given postoperatively to these patients, using the same indications and therapeutic selection criteria employed for any stage II breast cancer patients.

LOCALLY ADVANCED BREAST CANCER

The designation of locally advanced breast carcinoma (LABC) encompasses a heterogeneous group of tumors ranging from neglected, relatively slow-growing, large primary tumors to small breast ttimors presenting with extensive nodal metastases. Locally advanced breast cancer corresponds to stage III as defined by the UICC-AJCC system. Patients may be stage Ill by virtue of a T3 or T4 tumor or N2 or N3 disease. Inflammatory carcinoma, characterized clinically by a diffuse edema and erythema of the breast, frequently without a palpable breast mass, and defined pathologically by the presence of tumor emboli in the subdermal lymphatics, is a particularly virulent form of locally advanced breast cancer. The biologic diversity seen in locally advanced tumors makes a comparison of different treatment results or the formation of a single treatment recommendation for all women difficult.

Despite increasing recognition by physicians and the public of the importance of screening and early detection to decrease breast cancer mortality, 10% to 20% of women with breast cancer have locally advanced disease at diagnosis. 278

•279 Al

though clinical features at presentation and prognosis among women with locally advanced cancer may vary, there are two common problems in the treatment of these patients: obtaining local control and prolonging survival by preventing or delaying distant metastases.

Historically, treatment of LABC was surgical. By 1942,

PROPORTION SURVIVING 1 .

0.9

0,8

0.7

0.6

0.5

~ Occult Breast Csnc:er

-0.· T1N1 Historic:

-e- T2N1 Historic:

--

o.•L'======::;:'._~-~-~~-~_J o n u ~ ~ ~ n M % ~ w

TIME (mo}

FIGURE 40-10. Survival among patients with occult breast cancer compared with patients with an identified breast prtmary. (Baron P, Moore M, Kinne D. Occult prtmary cancer presenting with axillary metastases. Arch Surg 1990;125:210-214)

Haagensen and Stout developed the Columbia Clinical ClasSification of breast cancer based on 1135 carcinomas treated at Columbia Presbyterian Hospital from 1915 through 1942. 280

This classification identified clinical features associated with a high local recurrence rate and poor survival after surgery. These are listed in the earlier section on staging. Radical mas.tectomy in these patients was associated with a 53% local failure rate and a 0% 5-year disease-free survival.

Haagensen identified five grave signs, including edema of less than one third of the skin of the breast, skin ulceration, fixation of the tumor to the chest wall. fixed axillary lymph nodes, or an axillary node larger than 2.5 cm in diameter, that were associated with a high likelihood oflocal recurrence, although a single grave· sign was not thought to connote inoperability.

The poor results seen with radical surgery resulted in the use of radiation therapy as local treatment in wOmen with locally advanced breast cancer. Early studies demonstrated that the ability of radiation therapy to control LABC depends on dose. Fletcher and Montague reviewed the M.D. Anderson Hospital experience from 1948 to 1962 with the treatment of locally advanced breast cancer by radiation alone. 281 Fortyfour (28%) ofl59 women with tumors larger than 7 cm and 19 (17%) of 114 patients with tumors smaller than 7 cm failed. locally after radiation therapy. Fletcher emphasized the concept that, although microscopic disease can be sterilized by 4500 cGy in 5 weeks in 90% of all instances, large tumors require large doses (7500-8000 cGy) to control gross disease. Sheldon and associates corroborated the importance of dose in a report' of 192 stage III patients.'" At 5 years, women receiving 6000 cGy or higher doses to the primary site had an 83% local control rate, compared with 70% for those who received less than 6000 cGy. However, local control rates over 70% are not reported in all groups of stage Ill patients treated with radiation therapy. A 5-year actuarial risk of local recurrence of 59o/o was found in a reView of patients with inflammatory cancer treated from 1968 to 1986 at the JCRT.283 Two thirds of the local failures occurred before the appearance of distant metastases, and 26% were associated with moderate to severe symptoms. Bedwinek observed that 65% of 43 women with noninflammatory T4 lesions failed locally after treatment doses of 4000 to 7000 cGy.284

Increasing.the dose of radiation therapy may increase the likelihood of local tumor control and increase the risk of complications. Spanos and associates reviewed treatment Sequelae in patients at the M.D. Anderson Cancer Center and found that the most severe consequences of high-dose radiation were fibrosis and necrosis.285 The incidence of severe complications ranged from 17% at 5100 to 6000 cGy to 28% at 8100 cGy, but all patients were reported to have some component of fibrosis. Bedwinek and colleagues reported 4 cases of severe breast fibrosis among 83 patients. 284 Twenty moderately severe complications were reported, including breast fibrosis, breast and arm edema, stiff shoulder, and one brachia! plexus injury. Sheldon and coworkers observed 18 severe or moderate complications in 15 of 192 patients (197 breasts treated).282

These included moderate or severe arm edema in 6 patients and brachia! plexopathy in 4 patients. A cosmetic evaluation of these patients revealed that satisfactory cosmetic results were obtained in only 56% of patients, and 19% of the patients had poor cosmetic results.

r

J 82 of 122 Celltrion, Inc., Exhibit 1006

The poor results using local treatment in patients with LABC led to use of initial treatment with chemotherapy. It was actually in these stage III noninflammatory breast cancer patients that primary or neoadjuvant chemotherapy containing doxorubicin (Adriamycin) was first tested. These early results demonstrated prompt tumor shrinkage in many patients that facilitated subsequent radiation therapy or mastectomy. 286

Treating the tumor with primary chemotherapy allows the clinician to assess the efficacy of a particular drug regimen using the local disease as a marker. This approach may help in the design of systemic adjuvant treatment. 279·287- 289 The use of primary chemotherapy in patients with LABC is discussed later.

Two randomized trials examined whether mastectomy or radiation therapy is the more effective local treatment modality if combined with chemotherapy. 290

•2

" De Lena and colleagues randomized 132 women to mastectomy or radiation therapy after three cycles of doxorubicin and vincristine.290

Additional chemotherapy was given after local treatment. No significant differences in duration of response, relapse-free survival, or overall survival were observed. Similar findings were reported by the Cancer and Leukemia Group B in a study of radiation therapy versus surgery after three cycles of cyclophosphamide, doxorubicin, fluorouracil, vincristine, and prednisone. 291 In the 8 7 patients eligible for randomization after induction, no differences in relapse rates or survival were found.

In an attempt to improve local control rates and to avoid the.complications ofhigh-dose radiation therapy, several in~titutions combined preoperative or postoperative radiation therapy with mastectomy. Although excellent local control .rates were reported with the combination therapy, it is difficult to compare the results in patients who were able to undergo surgery with those in patients who are technically inoperable, frequently due to more diffuse disease, and receive radiation alone.284·292-294 High-dose radiation therapy or combinations of surgery and radiation therapy may result in excellent local tumor control in stage III breast carcinoma, but survival remains poor because of frequent distant recurrences. Bedwinek and associates observed that, although women treated with surgery and radiation therapy had a local failure rate of only 13% compared with 61 % after radiation therapy alone, a high frequency of distant metastases was seen after either treatment approach. 284

Few randomized studies have been performed to evaluate the addition of chemotherapy to local treatment in locally advanced breast cancer. These trials can be criticized because the chemotherapy used was not sufficiently aggressive. In a study from Finland, 120 patients with stage III cancer underwent modified radical mastectomy and were randomized to receive radiation therapy alone, chemotherapy alone, or both treatments.295 Disease-free survival after combined chemotherapy and radiation therapy was significantly better than after either modality alone, but a significant benefit in overall survival for the combined-treatment group was seen only in comparison with the radiation therapy group. Schaake-Koning and coworkers reported a randomized study of 118 patients with locally advanced breast cancer, including inflammatory cancer, comparing radiation therapy alone, radiation therapy followed by CMF chemotherapy, and radiation therapy preceded and followed by alternating cycles of CMF and doxo-


rubicin plus vincristine.261 Although there were trends toward improved relapse-free survival in the chemotherapy arms, these differences were not statistically significant. The projected 5-year survival rate was 37%, with no differences observed between treatment groups. The negative results of this study should be viewed in light of problems with its design. The technique of radiation therapy differed in the third arm of the study, and 6 of the 8 patients with Tl tumors were assigned to the radiation therapy-alone arm.

A European Organization for Research and Treatment of Cancer randomized study assigned 363 evaluable patients (13% with inflammatory cancer) to receive radiation therapy alone, radiation therapy plus endocrine therapy, radiation therapy plus CMF, or radiation therapy plus endocrine therapy plus CMF. 296 The systemic treatments prolonged the diseasefree interval if compared with radiation therapy alone, and the prolongation of disease-free survival was most notable for the combined endocrine and chemotherapy group. However, this improvement was accounted for primarily by a delay in locoregional relapse and no significant improvement in distant disease-free survival was found. Differences in overall survival, although appearing to favor the combined modality therapy, did not reach statistical significance.

There has been an increasing trend toward the use of induction chemotherapy before undertaking local treatment in the management of LABC. This approach has several potential benefits, including the prompt treatment of presumed systemic disease, the reduction of tumor burden before definitive local therapy, and the use of the response of the primary tumor as an in· vivo chemosensitivity assay. Clinical response rates of 70% to 90% have been reported for a variety of induction therapies.297-aoo Pathologic studies, however, indicate that the clinical assessment of response is relatively inaccurate. Feldman300 and Morrow299 each observed a 60o/o error rate in the clinical assessment ofresponse compared with the pathologic findings at mastectomy. The amount of residual tumor (microscopic versus gross) in the mastectomy after induction therapy was found to be the most significant predictor of disease-free and overall survival in the experience of Feldman and coworkers. 300 No increase in surgical morbidity has been reported for patients undergoing mastectomy after induction chemotherapy. 301

Representative results of combined-modality treatment are those reported from the M.D. Anderson Cancer Center.302 Of 174 patients with LABC treated with three cycles of fluorouracil, doxorubicin, and cyclophosphamide followed by surgery, radiation therapy, or both and additional chemotherapy, disease-free and overall survivals at 5 years were 71 % and 84% for stage IIIA patients and 33% and 44% for IIIB patients. Seventy percent of treatment failures in this study occurred at distant sites. The results appear related to tumor size and clinical nodal status, as shown in the experience from the National Cancer Institute of Italy in Milan. In their iO-year analysis of 277 patients with stage IIIB disease treated with a combined-modality approach, the survival rate was 48% for patients with primary tumors smaller than 5 cm, 20% for 5 to 10 cm, and 4% for tumors larger than 10 cm.303 The relation between nodal status and survival followed a similar pattern (NO, 27%; Nl, 28%; N2, 12%). Premenopausal or postmenopausal status had no prognostic influence (24% versus 23%). In a review of trials using a multidisciplinary approach,



a benefit in survival was observed when compared with historic controls.289

In most institutions, it is standard to treat patients with LABC with combined-modality treatment consisting of induction or primary chemotherapy and a combination of mastectomy and radiation therapy. Typically, primary chemotherapy is started using a doxorubicin-containing regimen (e.g., FAC or AC regimens) and given at full dose for four to six cycles, depending on the magnitude of response. After maximal tumor shrinkage has been obtained, it is still uncertain whether the next step should be surgery followed by radiation therapy or the opposite sequence. Usually, surgery is performed first if the local disease has become technically operable, and postoperative irradiation to the chest wall and appropriate nodal areas is reserved until the completion of all chemotherapy. The experiences at the National Cancer Institute ofltaly and M.D. Anderson Cancer Center strongly suggest maintenance chemotherapy is useful after induction chemotherapy and surgery to improve treatment outcome.289

·303 Based on the results from

the National Cancer Institute of Italy, six cycles of CMF is a reasonable approach to adjuvant therapy. 304

•305 In steroid re

ceptor-positive tumors, tamoxifen (20 mg/day given orally) can be added for 3 or more years after adjuvant chemotherapy is completed.

The trials using combined-modality treatment appear to show an improvement in outcome, but this approach is still associated with a high rate of distant metastases. Further progress in treating women with LABC awaits improvements in systemic therapy.

INFLAMMATORY BREAST CANCER

Inflammatory breast cancer is a rare type of breast cancer accounting for only 1 % to 4% of all mammary carcinomas.279

Inflammatory breast cancer carries an invariably poor prognosis if treated with radiation therapy or surgery alone. Almost 20 years ago, a combined-modality approach with primary chemotherapy (doxorubicin plus vincristine [AV]) was first attempted. 306 Currently, treatment consists of primfil-y chemotherapy followed by local treatment and additional chemotherapy, as for patients with LABC. Using four to six cycles of a doxorubicin-containing regimen (e.g., FAC, AVCMF) as induction chemotherapy, the objective response rate is about 70% with a 10% to 15% complete tumor response.306

-311 In

patients with inflammatory breast cancer, radiation therapy rather than mastectomy follows chemotherapy at many institutions because of concerns about ''cutting through'' dermal lymphatic involvement. Later, if feasible, mastectomy is performed.312

The use of combined modality treatment in patients with inflammatory breast cancer has resulted in a large improvement in the 5-year survival rate compared with the results in historic series in which local therapy alone was used. As for patients with LABC, there continues to be a high rate of metastatic disease with tiffie, and improvements in systemic therapy are required for this patient subset.

PAGET'S DISEASE OF THE NIPPLE

Paget's disease of the nipple is a rare form of breast cancer that is characterized clinically by eczematoid changes of the

nipple. Associated symptoms include itching, erythema, and nipple discharge.313-si 5 Histologically, Paget's disease is diagnosed by the presence of large cells with pale cytoplasm and prominent nucleoli known as Paget's cells. Sir James Paget found that this condition was invariably followed by cancer of the breast, which usually occurred within I year of diagnosis. 316 In approximately 54% of women with Paget's disease, a breast mass is detected at presentation, and in most of the remainder, infiltrating or intraductal carcinoma is identified in the mastectomy specimen.317 The average age of women with Paget's disease does not differ from that of women with other forms of breast cancer, but symptoms are frequently present for 6 months or more before diagnosis.313,31a

The relation between the changes observed in the nipple and the underlying breast cancer remains a matter of controversy. One theory suggests that the nipple involvement represents the migration of malignant cells from the underlying breast tumor. This view is based primarily on older histologic studies in which ducts containing malignant cells were connected to the overlying nipple containing Paget's cells.319·320

The alternative hypothesis proposes that the Paget's cells are a separate disease process and in fact are not neoplastic.315·321

If changes indicate Paget's disease, a mammogram may be helpful in identifying nonpalpable masses or microcalcifications indicating intraductal carcinoma. Regardless of the findings of the mammogram, all suspicious lesions of the nipple should be promptly biopsied.

Paget's disease has traditionally been treated with mastectomy. The rationales for this approach are the need to sacrifice the nipple-areola complex, the fact that the subareolar ducts may be diffusely involved with tumor, and the observation that carcinoma may be found at a considerable distance from the nipple. 315 The prognosis in Paget's disease is related to · the stage of the disease and appears to be similar to that of women with other types of breast carcinoma. A limited experience with breast-conserving procedures in the management of Paget' s disease has been analyzed. Paone reported 5 patients who underwent excision of the nipple with a wedge resection of underlying breast tissue who remained free of disea5e at IO-year follow-up.315 Lagios and coworkers reported 5 patients with no palpable breast mass and negative mammograms treated by excision of the nipple-areola complex. 322

One patient, treated with only partial nipple excision, developed recurrent Paget's disease at 12 months, which was resected. At a mean follow-up of 36 months, no patient had developed parenchymal breast recurrence. Twenty selected patients with Paget's disease without clinical or radiologic evidence of tumor were treated with radiation therapy alone or excision plus radiation therapy at the Institut Curie from 1960 to 1984. 323 At a median follow-up of 7 .5 years, 3 patients had recurrent disease in the nipple-areolar region and were treated with mastectomy. The 7-year actuarial probability of survival with the breast preserved was 81 %. Osteen collected a total of 79 patients treated by local excision with or without radiation therapy, with 9 local recurrences.317

When considering therapeutic options .in Paget's disease, it is helpful to think of the condition as an intraductal carcinoma involving the nipple that usually is associated with additional intraductal or invasive carcinoma in the underlying breast parenchyma. Mastectomy is considered the standard

T


therapy. In patients with apparently localized Paget's disease after clinical and radiographic evaluation, breast-conserving therapy can be considered. The selection criteria and treatment techniques applied to other women with breast cancer should be applied in this setting, with the understanding that the nipple-areola complex must be sacrificed and the risk of local recurrence may be higher than after mastectomy.

MALE BREAST CANCER

Cancer of the male breast is an uncommon disease, with an estimated 900 cases diagnosed in the United States in 1991 and 300 deaths in the same period. 324 Rare families have been reported with more than one man with breast cancer, and in some reports, as many as 60% of men with the disease have affected female relatives. 325 Other factors that increase the risk of male breast cance.r include Klinefelter's syndrome, schistosomiasis, and radiation exposure. 326- 328 The presence of gynecomastia does not seem to be associated with an increased risk of male breast carcinoma, although the microscopic changes of gynecomastia were associated with male breast cancer in 40% of patients reported from Memorial Hospital.329-331

Clinically, male breast cancer presents as a mass beneath the nipple-areolar complex in most cases. 329

•330•332•333 Ulceration of the nipple is a frequent sign, although isolated nipple discharge is uncommon. 334 The mean age of men with breast carcinoma is between 60 and 70, approximately 10 years older than that of women with the disease .327

•330

•335 Infiltrating ductal

carcinoma is the most common tumor type, but Paget's disease of the nipple and inflammatory carcinoma have been reported in men. Lobular carcinoma in situ is not seen in the male breast. 335

•336 As many as 80% of male breast carcinomas are

ER positive.337·338 An inverse correlation exists between re

ceptor positivity and age, similar to that seen in women.338

The standard treatment for male breast carcinoma has been mastectomy. If the tumor is not fixed to the pectoral muscle, a modified radical mastectomy can be done, or if muscle involvement is limited, a portion of this structure can be removed. for patients with extensive involvement of the pectoral muscle, a radical mastectomy may be required.339

Postoperative radiation is considered in patients with close surgical margins or extensive nodal involvement. Reports of primary radiation therapy for male breast carcinoma are scarce, but Ramantanis' series included 22 men treated by tumor excision and radiation therapy, producing a 35% 5-year survival rate. 329

There is some controversy about the prognosis of male breast cancer. Most oncologists think that, after accounting for differences in stage between male and female breast cancer, survival rates do not differ. 331

·335 Axillary node involve

ment has the same prognostic significance in men as it does in women. Heller and associates found a 90% 5-year survival rate for node-negative men and a 590/o survival rate for nodepositive men, which was similar to the survival of women with breast cancer treated in the same time period. 331

Systemic therapy of metastatic male breast cancer has usually involved hormonal manipulations to eliminate sources of androgen. Response rates of 50% to 80% are reported, and responses to second-line therapy are commonly seen. 338- 340

The traditional method of hormonal manipulation has been


orchiectomy. A recent literature review found a 53% response rate to this treatment. 339 Tamoxifen has a similar response rate and has become increasingly popular as a first-line hormonal manipulation, with orchiectomy reserved for patients who have failed multiple other therapies.339•340 Responses to progestins, antiandrogens, luteinizing hormonereleasing hormone analogs, and aminoglutethimide have been reported. 339

The available data on the use of chemotherapy in metastatic male breast cancer is limited. The drugs reported are those used to treat women, and the approach to chemotherapy for these patients should be the same as that employed for women.339

·341

·342 Two small studies of adjuvant chemotherapy

in node-positive men project 5-year survivals greater than 80%, suggesting that adjuvant chemotherapy may be of hen- . efit. 343

•344 Without definitive data, the guidelines for the use

of adjuvant therapy in male breast cancer commonly follow those used in female breast cancer.

BREAST CANCER DURING PREGNANCY

Breast cancer occurring during pregnancy is relatively uncommon, accounting for 2.8% of 45,881 patients reviewed by White. 345 If only breast cancer patients in their childbearing years are considered, 7% to 14% have breast cancer during pregnancy."'·"' A review of 416,441 pregnancies found an incidence of 2.2 breast cancers per 10,000 pregnancies. 346

The clinical presentation of breast cancer during pregnancy is the same as in the nonpregnant patient, and a palpable mass is the most common symptom. Nipple discharge, including a thin, bloody discharge from multiple ducts, may be a physiologic accompaniment of pregnancy. However, a persistent, pathologic, unilateral, bloody discharge during pregnancy requires surgical investigation. Man1mography is not particularly helpful in the pregnant woman due to the increased density of the breast during pregnancy. Decisions about the need for breast biopsy should be made on the basis of clinical examination. Delays in the diagnosis of breast masses in pregnant women are common, and most of this delay is due to physicians.346·347·349 Breast biopsy under local anesthesia is safe at any time during pregnancy and should be done for any suspicious mass.

The options for the local treatment of breast cancer during pregnancy are limited for the woman who wishes to continue her pregnancy. The use ofradiation therapy is contraindicated, due to the inability to shield the fetus from the internal scatter of radiation. If cancer is diagnosed in the third trimester, lumpectomy and axillary dissection can be performed and radiation therapy delayed until after delivery. Delays for longer periods to allow breast preservation may be detrimental.

Immediate reconstruction is contraindicated during pregnancy. It would be extremely difficult to obtain symmetry with the postpartum breast, and the risk to the fetus of a more prolonged anesthesia and increased blood loss is not warranted. Therapeutic abortion does not play a role in the treatment of nonmetastatic breast carcinoma. Donegan, reviewing the literature, found a 5-year survival rate of 62% for women undergoing therapeutic abortion compared with 54% for those whose pregnancy was completed. 35° King and colleagues found no survival differences between women whose pregnancies



were interrupted and those who delivered after stratification for stage of disease. 351

Breast cancer during pregnancy was thought to be a particularly virulent disease. Haagensen and Stout originally thought that breast cancer during pregnancy was categorically incurable.280 Much of the poor prognosis seems to be due to advanced stage of disease at diagnosis in pregnant women. Nugent352 found that 7 4% of women in his series had positive axillary nodes, a figure similar to the 65% reported by King,351

and 72% reported by Holleb.353 Petrek compared 56 pregnant breast cancer patients treated at Memorial Hospital from 1960 to 1980with 166 nonpregnant women of the same age treated in the same period and found that 61 % of the pregnant women and 38% of the nonpregnant women had positive lymph nodes (p < 0.05) and that 31 % of the pregnant women and 50% of their counterparts had Tl tumors (p < 0.05).354

After correction for tumor stage, survival in women treated during pregnancy is similar to that seen in nonpregnant women. Ribeiro and Palmer compared the survival of 40 pregnant patients with 120 controls matched for age, stage, and year of treatment and found no differences between groups.355 Petrek reported a 77% 10-year survival rate for node-negative women treated during pregnancy and a rate of 75% for nonpregnant women with negative nodes.354 j(ing observed a 71 % 10-year survival rate for node-negative pregnant patients.351 It appears that delay in diagnosis and treatment, rather than pregnancy itself, is responsible for the poor prognosis observed in some series of women treated for breast cancer during pregnancy. Subsequent pregnancy does not seem to have a detrimental effect on prognosis. 352

·354

•356

The use of chemotherapeutic agents in pregnant women is controversial. The effect of chemotherapy on the fetus is influenced by gestational age, the drugs used, and drug dosage. Schapira reviewed 71 patients treated with chemotherapeutic agents during the first trimester and found a 12.7% fetal malformation rate. 357 Another review found a low incidence of fetal malformations in women receiving antineoplastic drugs during pregnancy, although 40% of resulting offspring had low birth weights. 358 The decision to treat an individual patient with chemotherapy during pregnancy depends on the strength of the indication for treatment and the woman's desire to continue her pregnancy after being informed of the potential risks and benefits of treatment. In the third trimester, chemotherapy can usually be postponed until fetal maturity, when delivery can be induced.

LOCAL RECURRENCE AFTER MASTECTOMY OR CONSERVATIVE SURGERY AND RADIATION THERAPY

Local recurrence refers to the reappearance of breast cancer in the surgical or radiation field. The term local recurrence includes tumor in the chest wall, the overlying skin, residual breast tissue, ipsilateral axillary lymph nodes, supraclavicular nodes, or internal mammary lymph nodes. Local failure after mastectomy differs from local failure after conservative surgery in its time course, treatment, and prognosis.

The axillary node status is the best predictor of local recurrence after mastectomy. Local recu~rences are seen in fewer than 10% of women with negative axillary nodes, but 12% to 27% of node-positive women fail locally. 136

·'69•359 The

greater the number of involved axillary nodes, the higher is the risk of local failure. High local recurrence rates are also seen in women with locally advanced breast cancer treated with surgery alone. Sixty percent to 80% oflocal failures after mastectomy occur in the first 2 years after surgery, and failures are uncommon after 5 years. 136

•360

•361 The mechanism of local

recurrence is uncertain. Although the number of involved axillary nodes .is highly predictive oflocal failure, relapse in the axilla alone is uncommon. The chest wall and the overlying skin are the sole sites of local recurrence in more than 50% of patients. 360

•361

Approximately one third of patients with local recurrences after mastectomy have concurrent metastatic disease, and another 25% develop metastases shortly after the diagnosis of local failure.361 - 363 The median survival for patients with isolated local recurrence is 2 to 3 years. 364 A review of patients treated for local recurrence at the JCRT revealed that only 7o/o were free of metastases at 10 years.365

Surgery and radiation therapy have been used as treatment modalities for local recurrence. Local excision is applicable only for patients with limited amounts of disease. Simple excision is a relatively poor way of maintaining long-term local control, with fewer than 30% of women treated by this method remaining free of further local relapses. " 0·366·367 Radical chest wall resection has been employed for highly selected women with locally recurrent breast carcinoma. Local control has been achieved in approximately 90% of patients with this technique, and 5-year survivors are reported. 368

·369 This may

be secondary to the selection of women with long diseasefree intervals and relatively limited disease for this procedure. Hospitalization ranges from 14 to 21 days, and operative mortality is less than 5%.

Radiation therapy has been employed more frequently than surgery in the management oflocal failure after mastectomy. Iilitial complete tumor regression is seen in most patients, with reported complete remission rates ranging from 63% to 97%.365- 367•370 Unfortunately, additional local relapse~ occur in 36% to 61 % of women. The removal of all gross disease before radiation is associated with more favorable long-term local control rates in some series. 365

The routine use of systemic therapy after local relapse remains controversial. Although it is recognized that most of these patients fail distantly, the optimal timing of systemic therapy is unknown.

Local failure after breast-conserving therapy is defined as recurrence in the breast parenchyma, skin of the breast, or nodal areas. About 95% of local recurrences involve the breast parenchyma alone, and most local recurrences occur at or near the primary tumor site.311-

374 However, as the interval from initial therapy to local recurrence increases, more recurrences are seen elsewhere in the breast, and these recurrences probably represent new primary tumors. Kurtz and associates, reviewing 178 local recurrences in 1593 patients, found that only 21 % were distant from the primary tumor site, but 64% of recurrences seen after 10 years were in other quadrants of the breast.374 Recht and coworkers observed that the hazard rate for local failure at or near the primary site increases over the first 2. 5 postoperative years to a rate of 2% per year until year 5 and then decreases to about 0.5% per year by 8 years after treatment. 198 In contrast, the risk of failure elsewhere in the breast is about I% per year at 5 years,

r

1J 86 of 122 Celltrion, Inc., Exhibit 1006

with little change after further follow-up. The time course of local recurrence after breast-preserving therapy is much more protracted than that after mastectomy.

Approximately o;ne third of breast recurrences are detected solely by mammography, one third by physical examination

· alone, and one third by both modalities.375·376 Most recurrences are of the same histologic type as the primafy cancer, although recurrences detected solely by mammography have a higher likelihood of being purely intraductal. 375

·376

Unlike local recurrence after mastectomy, distant metastases are rarely seen in conjunction with local recurrence after breast-preserving therapy. Ninety-five percent ofrecurrences are described as operable, and the main reason for unresectability is diffuse skin involvement (i.e., inflammatory recurrence).372

·374 Total mastectomy has been the mainstay

of surgical therapy. Five-year survival rate$ after salvage surgery for breast recurrence range from 48o/o to 84%.373

·374

·377

If reconstruction is performed with salvage mastectomy, autogenous tissue should be Used. A small experience with wide excision as a treatment for selected breast recurrences has been reported.214.378 At a median follow-up of 5 years, 23% of patients had further local failure in the breast, suggesting that mastectomy is the treatment of choice. The role of systemic therapy after breast recurrence is controversial.

IN SITU CARCINOMA

In 1932, Broders defined carcinoma in situ as a condition in which malignant epithelial cells and their progeny are found in or bear positions occupied by their ancestors before the ancestors underwent maligriant transformation. In the breast, carcinoma in situ has traditionally been categorized as lobular or ductal, depending on the cytologic features and the pattern of growth.

DUCTAL CARCINOMA IN SITU

DCIS (i.e., noninfiltrating or intraductai carcinoma) is a proliferation of presumably malignant epithelial cells confined to the mammary ducts and lobules without demonstrable evidence of invasion through the basement membrane into the surrounding stroma. The management of ductal carcinoma in situ has become one of the most important and controversial topics in breast diseases. In this section, we review the available information about the incidence, diagnosis, pathologic and mammographic features, natural history, workup, and treaiment options for patients with DCIS.

The incidence of DCIS has increased dramatically since approximately 1983. In the period between 1983 and 1985, the incidence of DCIS doubled for premenopausal and postmenopausal wonien. The reason.$ for this increase are not certain, but a major factor is believed to be the increased used of high-quality screening mammography combined with biopsy for suspicious lesions. However, the incidence of breast cancer has been increasing at about 1 o/o per year for several decades before the use of screening mammography, and the incidence of breast cancer has increased approximately 2% per year since 1970 in women younger than 35 in whom screening mammography is rarely used. If the recent in-

In Situ Carcinoma 1297

creased incidence of DCIS is primarily due to earlier detection, the mortality due to breast cancer should decrease with time.

The clinical spectrum of DCIS is broad and includes lesions discovered incidentally during microscopic examination of breast tissue removed because of another abnormality (e.g., fibroadenoma or an area of fibrocy,stic change); small foci detected by mammography; nipple discharge with or without an associated mass; localized, sometimes large, palpable tumors; and large areas of abnormality found on mammography. The frequency with which these various patterns of presentation are observed depends on the population under study. In series reported before the advent of screening mammography, most patients with DCIS presented with a palpable mass, nipple discharge, or both. With the use of mammographic screening, most in situ cancers are detected exclusively by mammography.

The most common mammographic abnormality associated with DCIS is clustered microcalcifications. The probability of finding malignant disease is related to the number of clustered microcalcifications, and with their appearance. Of particular concern are branching, irregularly shaped calcifications. However, mammographically detected calcifications, even if clustered, are not specific, and pathologic examination reveals carcinoma, most often DCIS, for only approximately 25% of patients for whom the mammogram shows suspicious calcifications. Less frequently, DCIS may have the mammographic appearance of a soft tissue mass with or without calcifications. In one series of 100 consecutive cases of DCIS seen at the Brigham and Women's Hospital and Dana-Farber Cancer Institute, 10% presented on mammography with only soft tissue abnormality and another 12% with soft tissue abnormality and microcalcifications.379 The soft tissue abnormality was associated with gross involvement and dilation of ducts or lobules or periductal fibrosis on pathologic correlation of these lesions.

There are three histopathoio!';ic entities that can be confused with DCIS: atypical ductal hyperplasia, lobular carcinoma in situ, and ductal carcinoma in situ with minimal stromal invasion.

Atypical dUcial hyperplasia is a nonmalignant intraductal epithelial proliferation closely resembling DCIS, although not cdtnpletely fulfilling its histologic criteria. No special techniques are available that enable a clear and objective differentiation of the two lesions. Atypical ductal hyperplasia carries a risk of the development of a subsequent cancer and is believed to be the precursor of noncomedo DCIS. 69 In most cases, the distinction between DCIS and atypical ductal hyperplasia is obvious. In a few borderline lesions, particularly if small, differentiation may be difficult. In a recent study of concordance among a group of five experts in breast cancer pathology, little agreement was seen in the diagnosis of borderline lesions.380

Lobular carcinoma in situ is often difficult to differentiate from the solid type of the noncomedo subtype of DCIS. In some instances, both may be present in the same lesion, and it has been conjectured that this type of DCIS may be a manifestation of lobular carcinoma in situ in ductal structures.381

The ability to identify small clusters of tumor cells close to the intraductal foci as evidence of minimal stromal invasion may be difficult for several reasons. Sampling error may hamper the detection of microinvasion. DCIS may extend to in-



volve adjacent lobular units and mimic the appearance of invasive cancer. In cases of DCIS with severe periductal fibrosis, a distorted small duct may be mistaken for invasion. Because of these difficulties and the clinical implication of diagnosing invasive cancer, most experts in breast pathology restrict the diagnosis of minimal stromal invasion to cases in which invasion is straightforward. Muscle-specific antibody to actin or one of the components of the basement membrane, which would be expected to be positive in the myoepithelial cells around in situ lesions and lacking in the. invasive foci, may help in this differential diagnosis, but is not used in routine cases.382

DCIS is heterogeneous in terms of its histopathologic growth pattern, cell type, extent, and biologic behavior. Based on the growth pattern and cell type (e.g., cribriform, micropapillary, clinging, solid), DCIS can be classified into comedo and noncomedo subtypes, and there may be a combination of these types. The comedo type is characterized by nuclear polymorphism, numerous mitoses, and significant necrosis. The noncomedo subtype is characterized by monomorphic nuclei, few or no mitoses, and the absence of significant necrosis. Many investigators believe that it is useful to divide DCIS into the comedo and noncomedo types, because the comedo type appears more malignant cytologically, has a higher proliferative rate as determined by thymidine-labeling studies, more frequently expresses the ERBB2 oncogene, and is more likely to be associated with areas of microinvasion than the other types.383•384 The mammographic presentation of the two types of DCIS is different. 211 The comedo type is associated with the so-called casting (i.e., linear) or coarse granular microcalcifications that usually develop in the necrotic debris of the tumor. The noncomedo type is associated with the fine granular microcalcifications that appear to develop in the secretion produced by the lesion. In most cases, the mammographic detection of the comedo DCIS is straightforward, because few benign lesions have a similar type and pattern of microcalcifications. Benign lesions often have microcalcifications similar to those of the noncomedo DCIS. There is a difference in the reliability of mammography in predicting the histologically assessed size of the two types of DCIS. Mammographic estimates of the comedo type Closely approximate the histologic size of the tumor, and the estimates of size of the noncomedo type based on the extent of microcalcifications may be considerably s.maller than the histologic size.211 The distinction between the comedo and noncomedo subtypes is limited by the lack of a clear and reproducible definition, the fact that many lesions contain both subtypes, and the lack of a clear association between subtype and clinical behavior.

DCIS has been considered by some to be a multicentric process with separate areas of involved ducts. This concept is based on studies using the standard pathologic examination of surgical biopsies and subsequent mastectomy specimens. Residual tumor foci in the latter were considered as separate foci and therefore evidence of multicentricity. However, in a study using a correlated radiologic-pathologic technique in combination with a subgross sectioning and extensive sampling of a series of 82 mastectomies harboring DCIS, only one was found to have a multicentric distribution. 211 The tumor foci are more likely to be evenly distributed within a given region (typically corresponding to a breast segment) without

intervening areas of uninvolved breast tissue. This concept is supported by clinical data showing that local tumor recurrence after breast-conserving treatment of DCIS typically appears in the vicinity of the biopsy s.ite. The finding ihat most DCIS has a unicentric distribution is important in allowing the possibility of a complete resection of the lesion and implies that the extent of the lesion is the main factor influencing the feasibility of breast-conserving treatment.

Reports on the size distribution of DCIS are conflicting. In a series of 115 patients, Lagios and colleagues reported 52% of patients to have tumors that were 25 mm or smaller. 385 In this series, 79 patients with mammographically detected DCIS with a reported average size of 7 mm were treated by excision alone. The resection margins were carefully assessed for completeness and considered negative. In the series of Silverstein and associates, only about 20o/o of lesions were smaller than 2 cm.386 In about half of the patients with small lesions who were selected for breast-conserving treatment, the margins of the initial biopsy were considered involved, suggesting that the extent of the lesion was even larger than originally· assessed. In the series reported by Holland, only 15% of the DCIS lesions were smaller than 2 cm and 51 % were larger than 5 cm.211 The size distribution was not affected by the mode of detection, with a similar distribution found for tumors detected by mammography and by clinical examination. Sixtysix percent of the lesions were confined to one breast quadrant, 23% extended over more than one quadrant, and 11 % were centrally located. There was a 52% rate of occult involvement in the nipple-subareolar region. For many cases of DCIS, adequate resection requires wide excision and may be feasible in only a few patients.

The chief issue in treating DCIS is its risk of progression to invasive cancer. The available information on this subject is extremely limited, because most patients with DCIS have been treated by mastectomy. The only long-term studies to address this issue are those in which patients with DCIS were inadvertently treated by biopsy alone. Two studies have identified patients with DCIS during histologic review of biopsies originally categorized as benign. However, in both of these studies, a complete excision of the lesion was not attempted and the status of the resection margins was unknown. In one of these studies, Page and coworkers found subsequent invasive carcinoma in the ipsilateral breast in 7 (28%) of 25 patients at intervals of 3 to 10 years (mean, 6.1 years) after the initial biopsy.387 All seven invasive cancers occurred in the vicinity of the original biopsy. In the other study by Rosen and colleagues, 8(53%) of 15 patients with follow-up data or 8 (27%) of 30 total patients subsequently developed an invasive carcinoma in the same breast an average of 9.7 years after the initial biopsy showing DCIS. 388 In most of these pati'ents, the invasive tumor occurred at or near the original biopsy site. In both series, the cases of DCIS were all micropapillary or cribriform (noncomedo) types. These studies suggest that some patients with noncomedo DCIS treated by biopsy alone develop invasive cancer in the ipsilateral breast, usually in the region of the initial lesion. The invasive carcinoma commonly occurs a long time after the initial biopsy.

Insight into the biologic significance of DCIS can be obtained from studies that indicate that foci of DCIS are frequently detected in the contralateral breast of women with invasive breast cancer. However, there is a discrepancy between this


incidence of contralateral DCIS and the risk of developing a subsequent, clinically evident, opposite breast cancer. Alpers and Wellings found DCIS in 48% of breasts contralateral to cancer-containing breasts, yet the cumulative 20-year risk of Opposite breast cancer in patients with an ipsilateral breast cancer .is only about 10%. 389 These data suggest that not all examples of histologically detectable DCIS progress to clinically significant cancers.

It may be possible to discern the biologic significance of DCIS based on its frequency in autopsy series. Unfortunately, there are few studies, because routine systematic examination of breast tissue at autopsy is not performed, and the available studies are not consistent. Alpers and Wellings reported 185 randomly selected breasts from autopsies in 101 patients. 389

Each breast was examined in toto by a subgross sampling technique. Of these 185 breasts, one or more foci of DCIS were found in 11 (6%). The average number of foci per involved breast was 1.5. The likelihood of finding DCIS was not directly related to age. DCIS was found in 3 (5%) of 56 patients 49 years ofage or younger, in 7 (10%) of 70 patients between the ages of 50 and 69, and in 1 of 59 patients older than 70. Bartow and associates studied the breasts from autopsies of 519 female patients who were 14 years or older. 390 All breasts were studied by examination in toto by subgross sampling technique. Only 1 (0.2%) patient (40 years old) was found to have DCIS, and 5 patients were found to have invasive cancers. These studies from the United States indicate that DCIS is Uncommonly found on routine autopsy which suggests that the presence of DCIS is of potential significance. A study from Denmark reported by Anderson and coworkers suggested a higher incidence of DCIS in autopsies.381 They examined 83 patients and found DCIS (with or without lobular carcinoma in situ (LCIS]) in 11 (13%). It is possible that variations in diagnostic criteria may account for these differences.

It is not possible by routine pathologic examination to differentiate examples of DCIS that are likely to progress to invasive cancer fro~ lesions that are biologically innocuous. The various histologic subtypes may have different likelihoods of progression to invasive disease, but the differences are incompletely understood, because most clinical series have not differentiated subtypes in reporting outcome. It is possible

In Situ Carcinoma 1299

that recently developed techniques for studying oncogenes and their products, growth factors, and markers of cell proliferation and DNA content may ultimately determine the effects of the types of DCIS. In one study of 69 patients with DCIS treated by excision, only 1 of 31 diploid noncomedo lesions recurred, compared with about 40% of the other 38 lesions (p < 0.05).391

The results of treatment with mastectomy indicate that about 99% of patients with DCIS are cured. 386

•392 However,

patients with an initial biopsy that showed DCIS but who later had invasive cancer in the mastectomy specimen were generally excluded from the series. In the study by Silverstein and colleagues of 75 patients with DCIS (average tumor size of 4.1 cm) treated by mastectomy, 51 (68%) had residual disease at the time of mastectomy, including 10 (13%) with microinvasive disease.386 For this reason, it may be difficult to compare outcome after mastectomy with outcome in series of patients treated by breast-conserving approaches.

The results of the use of excision alone for DCIS (listed in ascending order ·of follow-up time) are given in Table 40-18.393-398 Local recurrence rates after excision alone tend to increase with time of follow-up; in those studies with longer follow-up, the local recurrence rates are close to 50%. This most likely reflects the propensity of this lesion for late recurrences, although it is possible that patients in the series with longer follow-up had less adequate excisions: Approximately half of local.recurrences contain invasive cancer. Although these recurrences have generally been detected at an early stage, the salvage of patients with these recurrences as determined by 5- and 10-year relapse-free and overall survival rates has not yet been established.

Results of the use of excision combined with breast irradiation are given in Table 40-19.396

•399

-404 Local recurrence

rates after excision and breast irradiation appear to be lower than those after excision alone (by at least 50%), although the results in some of the series are less favorable than in others, and as in the studies using excision alone, approximately half of local recurrences contain invasive cancer. 399

•402

Only short-term information is available about the salvage of recurrences after excision combined with breast irradiation.

Several trials comparing excision with and without breast

TABLE 40-18. Results of Treatment With Excision Alone for Ductal Carcinoma In Situ

Median No. of Follow-up Percent No. With

Investigations Patients (mo) With Recurrence Invasion

Carpenter, 1979-86393 28 38' 18 1/5 Arnesson, 1978-84394 38 60 13 2/5 Lagios, 1972-87395 79 68 13 5/10 Fisher, 1976-84396 22 85' 43 5/9 Graham, 1969-88397 53 97 38 7/14 Gallagher, 1944-81398 13 100 38 3/5 Price, 1972-82572 35 108 63 12/22

Total 35/70 (50%)

• Mean follow-up (mo).



TABLE 40-19. Results of Excision Combined With Radiation Therapy for Ductal Carcinoma In Situ

Median No. of Follow-up Percent No. With

Investigations Patients (mo) With Recurrence Invasion

McCormick, 1977-883911 54 Kurtz, 1975-85400 44 Salin, 1978-85401 51 Bornstein, 1976-85402 38 Fisher, 1976-84396 29 Stotter, 1958-87403 42 Fourquet, 1967-85404 67

Total

irradiation for patients with DCIS have been undertaken, but published results from these trials are not yet available. There were a few patients with DCIS inadvertently entered onto the NSABP trial B-06, suggesting that irradiation is useful in avoiding recurrence, but this number: is too small to provide meaningful conclusions (see Fisher in Tables 40-18 and 40-19).396 There are no firm guidelines for the use, of irradiation after breast-conserving surgery for DCIS. Breast irradiation is commonly recommended in patients electing breast conservation if the lesion is greater than 1 to 2 cm or if the lesion approaches the margins of resection.

Careful mammographic and pathologic evaluations are essential for patients with DCIS considered for breast-conserving treatment. A careful mammographic evaluation consists of the standard mammographic views, but it is extended by magnification views for better evaluation of the form, distribution, and extent of the microcalcifications. The area between the lesion and the nipple should receive special attention. Larger lesions should preferably be localized at their margins using two hook-wires to ensure an adequate excision. A postoperative mammogram may be required to confirm the completeness of resection if there is any doubt based on review of the specimen mammogram. A careful pathologic evaluation requires a proper orientation of the specimen. The surgeon should label the specimen at two points: for example, at the closest point to the nipple and ,at 12 o'clock. The entire surface of the specimen should be inked to facilitate recognition of the surgical edges on microscopy. Radiographs should be obtained of the specimen while it is intact to permit comparison of the lesion with that of the preoperative mammograms and again after sectioning at 4- to 5-mm intervals to allow recognition of the exact site of the microcalcifications.

Generous sampling of the area of microcalcifications and the margins is recommended .to facilitate the detection of possible microinvasion and the assessment of the margins. Adequate orientation and sampling of the specimen are essential to assess the extent of the lesion and the adequacy of the resection. A portion of the lesion should not be sent for biochemical assay of hormone receptors or other assays, because of the concern that the portion sent might contain an undiagnosed area of invasive cancer. A recent consensus committee on DCIS recommended that the following should be described routinely: tumor size; margin involvement; nuclear features including size, pleomorphism, and mitoses; necrosis; and architecture. •05 It is useful for the pathology report

35 18% 3/10 61 7% 3/3 68 10% 2/5 81 21% 5/8 85 7% 1/2 92 9% 414

104 10% 5/7

25/42 (60%)

to include a statement about the type of the microcalcifications and their relation to the malignant ducts.

The histologic evaluation of the surgical margins is considered an important part of ·the assessment in any patient being considered for breast-conserving treatment. Unfortunately, there is no established definition of what constitutes a negative margin. The minimal definition of an negative margin is the absence of foci <>f DCIS at the surgical edge. It is not known if a certain distance is required between any foci of DCIS and the margin to be certain about the adequacy of the resection. It has been suggested that a negative margin should include uninvolved glandular structures beyond the lesion at the margin.of the specimen.405a

With the increased risk for local recurrence with breastconserving treatment, the frequency of invasive cancer in these recurrences, and the uncertain effectiveness of salvage for recurrence, mastectomy is viewed as the safest option. This is always true for a patient with a locally extensive lesion. The greater difficulty arises in treating a patient with an apparently localized lesion that has been adequately resected and for whom breast conservation is an important issue. It is useful to attempt to estimate for the patient the degree ofrisk associated with breast conservation. With the information currently available and after a careful mammographic and pathologic assessment, breast conservation may be associated with a 15% to 20% risk ofa recurrence over a 10-year period. Because approximately half of these contain invasive cancer, there is a 7% to 10% risk of an invasive cancer over a 10-year period. If approximately 50% of invasive cancers can be cured, the breast cancer mortality associated with breast conservation is in the range of 3% to 5%. (These estimates are given to illustrate the level of risk and may need to be modified as additional information is obtained.) Patients show a wide variety of reactions to a discussion of these estimates of risk. For some patients, this level of risk is unacceptable, but for others, it is viewed as a reasonable risk to assume to avoid · breast amputation.

Is it reasonable to allow a patient to assume this level of risk? There is not a consistent policy of prophylactic mastectomy for high-risk patients. Because DCIS is, strictly speaking, not a cancer, it is of interest to view the policy for DCIS in relation to the policy for other high-risk conditions. The risk of developing an invasive breast cancer over a 10-year period for a 50-year-old woman in the United States is approximately 1.8%. For a 50-year-old woman identified as high risk on the


basis of family history, hormonal and reproductive factors, or a prior breast biopsy revealing atypical epithelial hyperplasia, the risk is in the range of 2% to 10% or even greater. For a woman with a diagnosis of lobular carcinoma in situ (discussed later), the risk is approximately 1 % per year or 10% over a 10-year period. Prophylactic mastectomy is not routinely recommended for patients with these high-risk conditions. The 10-year risk of an invasive cancer for a woman with DCIS is in the range of 7% to 10% or similar to that for other highrisk conditions for which prophylactic mastectomy is not recommended. Based on all these considerations, the option of breast conservation is reasonable assuming that the lesion is apparently localized, has been adequately resected, and that the patient comprehends and accepts the risk assumed by breast-conserving treatment.

DCIS represents a challenge and an opportunity for the 1990s. Because DCIS probably represents breast cancer at its earliest stage, it is an opportunity to identify the earliest genetic abnormalities associated with the disease. It is important to determine whether the real incidence of the lesion is increasing and what factors (e.g., dietary and hormonal) may be responsible for this. The identification of markers for progression to invasive breast cancer would greatly improve the management of patients with DCIS. Given the importance of breast conservation for many women, it is important to develop rules for the successful use of breast-conserving treatment. Clinical trials currently in progress may help to define guidelines for the use of breast irradiation. It may be useful to consider some patients with DCIS (especially those with the noncomedo type) as being at high risk for invasive cancer and candidates for trials testing the value of prevention (e.g., tamoxifen). The use oftamoxifen is being tested in the current NSABP trial for DCIS in which patients are randomized after excision and irradiation.

LOBULAR C.ARCINOMA IN SITU In 1941, Foote and Stewart described a noninvasive form of mammary carcinoma arising from the lobules and terminal' ducts that they called lobular carcinoma in situ (LCIS). 406

Their initial report included three important features of LCIS. The lesion is an incidental microscopic finding that cannot be identified clinically or by gross pathologic examination, invasive cancers that develop after LCIS may be infiltrating ductal or infiltrating lobular tumors, and the lesion is multifocal in the breast. These observations led Foote and Stewart to conclude that simple mastectomy was the appropriate treatment for LCIS. Additional information about LCIS has . been accrued since the report of Foote and Stewart, but the appropriate clinical management of this lesion remains controversial.

The true incidence of LCIS in the general population is unknown because it lacks clinical or mammographic signs. LCIS is found in from 0.8% to 8.0% of breast biopsies.•07

-4o9

In all reports, LCIS is more frequent in premenopausal women, with a mean age of 44 years in one series of 98 patients. 408-410 This age distribution may reflect the fact that the benign breast abnormalities that require biopsy are pril)'larily seen in premenopausal women. The frequency with which LCIS is diagnosed appears to be increasing, with a 15% rise in the number of cases seen from 1973 to 1988 reported in one series.411 This increase in the incidence of LCIS may be due to the increased numb"er of breast biopsies done for mam-

Systemic Adjuvant Therapy 1301

mographic abnormalities. Although LCIS lacks mammographic signs, it accounted for 14% of mammographically detected cancers in one large series. 412

•413

The major issue in the management of LCIS is whether the LCIS is a premalignant lesion or a marker of increased risk for the development of breast carcinoma. Support for the idea that LCIS is a premalignant lesion came from the report of McDivitt and associates,414 which was subsequently updated by Rosen."' Of 40 women with LCIS treated with biopsy alone, a cumulative risk of ipsilateral breast cancer of 10% at 5 years, 15% at 10 years, and 30% at 15. years was found. The risk of contralateral breast cancer v.:as 15% at 15 years. Rosen and coworkers found a similar degree of risk that continued to increase with the length of follow-up through 24 years. 415 Rosen recommended ipsilateral mastectomy with a large contralateral biopsy as the appropriate treatment for LCIS. However, a significant body of data supports the idea that LCIS is a marker of increased risk for the development of cancer in either breast rather than a precancerous lesion.408

·410

·416 An

derson found a 19.5% incidence of ipsilateral breast cancer and an 8.5o/o incidence of contralateral carcinoma in 52 women followed after a biopsy demonstrating LCIS.417 Haagensen followed 211 women and found a 1 7% incidence of

· invasive carcinoma that was equally distributed between the index breast and the contralateral breast~ 410 In a review of 228 published cases of LCIS treated by biopsy alone, 15.5% developed invasive carcinoma of the ipsilateral breast and 9.3% had a contralateral carcinoma.416

It appears that most women with LCIS do not develop breast cancer, !llld the risk of breast cancer development is approximately equal in both breasts. Treatment strategies addressing one breast, such as unilateral mastectomy with contralateral biopsy, appear to be illogical, because the risk of LCIS is bilateral regardless of the findings of the contralateral biopsy. One treatment option for the woman with LCIS is careful observation, as would be carried out for any woman known to be at increased risk for the development of breast cancer due to a positive family history or prior history of breast cancer. In women unwilling to accept the 20% to 30% risk of the development of breast cancer associated with this policy, bilateral simple mastectomy, usually with immediate breast reconstruction, is another therapeutic option. Radiation therapy has no role in the management of LCIS. It is unnecessary to obtain histologically negative margins in women who will be followed expectantly, because LCIS is known to be a diffuse lesion. If observation is elected, it must last for the patient's lifetime, because the increased risk of breast cancer persists indefinitely. 415 Efforts to identify features of LCIS associated with a higher likelihood of the development of malignancy have been largely unsuccessful. Haagensen found that the ratio of observed to expected cases of breast cancer increased from 5. 7 for women with LCIS alone to 8.5 in women with a positive family history and LCIS.418 Histologic features, including the amount ofLCIS, have not predicted the subsequent development of invasive carcinoma.415

SYSTEMIC ADJUV ANT THERAPY

SCIENTIFIC BACKGROUND Many of the experimental foundations of systemic adjuvant therapy were derived from laboratory studies performed more than 25 years ago by investigators in North America, and these



are summarized in Table 40-20. 304•419

-423 This evidence in

dicated that the lymphatic and blood vascular systems were so interrelated that it was impr~ctical to consider them as independent routes of neoplastic cell dissemination. This meant that the regional lymph nodes were of prognostic, rather than anatomic, importance in cancer survival. This provided a basis on which an alternative hypothesis (i.e., biologic, rather than the anatomic and mechanistic one proposed by Halsted) could be formulated. Observations on the survival of mice with transplantable solid tumors of various sizes subjected to resection revealed that the surgical cure rates correlated with tumor size, although there was great variability in survival and the metastatic· tumor burden at the time of surgery among the animals. There were breakpoints in the survival curves, after which time no further recurrences were found. These breakpoints appeared at similar times after surgery, although the percentage of survival in the various groups was different. This observation suggested that the biology of the tumor was similar in the various animals but that the extent of the tumor cell burden determined the cure rates. A corollary to this observation is that definitive conclusions about curative effects of any treatment must be withheld until this breakpoint is reached.

Empiric and practical considerations were used in the selection of chemotherapeutic agents, the treatment schedule, and the duration of treatment. The selection of drugs and the treatment schedule (typically given in cycles) must achieve a net reduction in the cell burden by the time to deliver the next cycle, and the optimal treatment intensity must be at least as intense as that which yields the optimal results in advanced disease. In mice bearing mammary tumors, survival was invariably superior after a combination of drugs compared

TABLE 40-20. . Laboratory Findings, Clinical Observations, and Biologic Hypotheses Influencing the Beginning of Modern Adjuvant Chemotherapy for High-Risk Operable Breast Cancer

• By the time cancer becomes clinically detectable, it is advanced (about 30 doublings) and has had ample opportunity to establish distant micrometastases.

• There is no orderly pattern of tumor cell dissemination, and the blood stream is of considerable importance in tumor spread.

• The likelihood of metastatic disease is directly related to the size of the tumor mass; surgical "c~re" rates drop as tumor volume at surgery increases.

• Tumor growth fraction is inversely related to population size. • In homogenous tumor cell populations, effective drug kill follows

first-order reaction kinetics. • The use of a combination of drugs is superior to the use of a

single agent and can eradicate 10 to 100 times as many cells. • In transplantable tu):11ors, surgical adjuvant chemotherapy

increases the long-term cure_rates. • The optimal kinetic conditions to achieve cure exist if

microscopic foci of disease are present after curative surgery and/or radiation therapy.

• The efficacy of chemotherapy is dose dependent, related to the tumor cell burden at the time of drug treatment and to the presence of primary resistant tumor cells.

• Operable breast cancer is often a systemic disease and variations in local-regional therapy are unlikely to substantially affect survival. Only by control of distant disease can there be an improvement in the outcome of breast cancer patients.

with single-agent chemotherapy. The animal models showed that the emergence of drug resistance increases with the exten't of the disease and that this constitutes a major element in the failure of adjuvant therapy. The experimental systems demonstrated that surgery followed by adjuvant chemotherapy yielded superior survival results to either modality alone.

It was the evaluation of two surgical trials that challenged the principles of the Halstedian hypothesis and supported the laboratory investigations. 136·138 The high recurrence rate, especially during the first 3 years after the Halsted radical and extended radical mastectomy, could have only one reasonable explanation: micrometastases may exist at the time of surgery, and they are unaffected by local treatment. This. observation helped to challenge the old anatomic and mechanistic dogma, opened the pathway to biologic hypotheses, and paved the way for breast-conserving treatment and adjuvant chemotherapy. During the mid-1970s, hormones began to receive renewed attention as potential adjuvant treatments.424 In a carcinogen-induced mammary tumor model in rats, it \vas demonstrated that the antiestrogen compound, tamoxifen, could delay the appearance of tumors and decrease their overall frequency. This was of great interest because of its relevance in the adjuvant treatment of breast cancer and because of its possible inhibition of a carcinogenic process.

ADJUV ANT CHEMOTHERAPY

The first randomized trials based on the modern concepts of adjuvant chemotherapy were performed by the NSABP and by the National Cancer Institute of Italy, using patients with positive axillary nodes. 425

'426 The first NSABP study used

single-agent chemotherapy (melphalan or L-phenylalanine mustard [L-PAM]), and the Milan trial used a multidrug regimen that included cyclophosphamide, methotrexate, and fluorouracil (CMF). Eventually, the strategy of using combination chemotherapy proved superior in the adjuvant situation as it had in animals, and it became the standard approach. The long-term results of these trials established the effectiveness of adjuvant chemotherapy in node-positive breast cancer and provided evidence or confirmation of important biologic and pharmacologic concepts: the prognostic importance of the tumor burden in the ipsilateral axilla, the doseresponse relation for adjuvant chemotherapy, and the clinical relevance of drug resistance. 286

•427

Long-term results are now available from the initial Milan trial.304 Figure 40-11 displays the 15-year results of CMF administered for 12 monthly cycles starting within 4 weeks after surgery and compared with a no-treatment control group. The benefit of adjuvant chemotherapy occurred during the first 3 years after radical mastectomy, and the difference between controls and CMF patients remained about the same for the subsequent 12 years. This observation suggests that early findings can predict the late outcome with sufficient accuracy and that, in some patients, micrometastases include aggressive tumor clones that are resistant to chemotherapy. The results of the CMF trial confirmed the prognostic importance of the extent of nodal involvement in the axilla; there was a consistent inverse relation between the number ofhistologically involved lymph nodes and the treatment outcome (Table 40-2I). In the first CMF trial, the significant survival benefit observed in premenopausal women (Fig. 40-


p E R c

100

E 50 N T

A

\ \,\

......

3

... __

LDG-RANK: P • 0.002 WILCOXON: P ,. 0.0001

...... ...... ...... ......... CMF

.... __ ..... .... _________ , SURGERY ALONE

5 7

YEARS

9 11 13 15

p E R c

100

E 50 N T

B

', ,, '


', LOG-RANK: P • 0.08 ',, WILCOXON: P .., 0.08

......

3 5

...... ...... ' ', CMF ......

7

YEARS

... _ ...

9

... ... ...... ... ... ....--.... .._ .. SURGERY ALONE

11 13 15

FIGURE 40-1 t. First Milan i:rial comparingadjuVant cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) with surgery alone (control). 15-year results. (A) Comparative relapse-free survival. (B) Comparative total survival. (Bonadonrta G. Evolving concepts in the systemic adjuvant treatment of breast cancer. Cancer Res 1992;52[8],2127-2137)

12) was not duplicated in postmenopausal patients. However, as a consequence of trial design and protocol deviations, the drug doses employed in postmenopausal patients were initially low and were often arbitrarily further reduced during most cycles. 421 ,42a

The second CMF adjuvant program attempted to determine whether six cycles of combination chemotherapy was as effective as 12 cycles.286

·'0

"427 After 14 years, relapse-free (Fig.

40-13) and total survival rates are equal in the two treaiment groups. In this trial in which full doses of drug were frequently used, there was equal benefit for premenopausal and postmenopausal patients. Comparable findings with the same theoretical and practical considerations were rec~ntly reported

TABLE 40-21. Fifteen-Year Results of the First CMF Program Carried Out at the Milan Cancer Institute

Control CMF Patient Group (%) (%) p

Relapse-Free Survival

All patients 26 36 0.002 Premenopausal 28 42 0.002 Postmenopausal 25 31 0.22 1-3 nodes positive 31 42 0.009 >3 nodes positive 15 24 0.05

Total Survival All patients 33 42 0.08 Premenopausal 35 51 0.02 Postmenopausal 32 35 0.85 1-3 nodes positive 37 48 0.08 >3 nodes positive 24 31 0.31

by the Dana-Farber Cancer Institute429 and the NSABP group430 using doxorubicin-containlng regimens. Although the optimal treatment duration has not been firmly established, six cycles offull-dose CMF or one of the other effective polydrug regimens (Table 40-22) are currently recommended in the adjuvant treatment of node-positive breast cancer if given outside the context of a clinical trial. Although other effective polydrug regimens are delivered in various centers of the world, those reported in the table. have been exhaustively tested through large prospective trials and found to be reproducible in terms of benefit and toxicity. If chemotherapy is indicated, these regimens can be administered in an adjuvant situation as primary (neoadjuvant) treatment or in the management of locally advanced or clinically metastatic disease.

The first CMF schedule is the so-called standard regimen and includes oral cyclophosphamide. Whether the addition of prednisone improves treatnient outcome remains unsettled. The schedules with intravenous cyclophosphamide were devised at the Milan Cancer Institute to improve patients' coinpliance. Although the intravenous administration on days 1 and 8 of all three drugs provides a greater dose intensity, a recent trial of administration every 3 weeks in women with one to three positive nodes indicated that the 5-year results are equivalent to those achieved with standard CMF. 431 In clinical practice, the well-known CMFVP ( cyclophosphamide, methotrexate, fluorouracil, vincristine, predniSone) regimen has been used less frequently during the past few years because of severe myelosuppression and neurotoxicity without clearcut evidence of a superior long-term benefit. Other drug regimens, including cisplatin, paraplatin, mitomycin C, mitoxantrone, vinorelbine (Navelbine), taxol, growth factor-supported, high-dose chemotherapy with or without autologous bone marrow or cryopreserved circulating hematopoietic progenitor cells, are effective, but they are still considered experimental.


1304

p E R c

100·

E 50 N T

A


\

\ \ ,,

' ......

3

'-- ....

5

LDG-RANK: P • 0.002 WILCOXON: P • 0.0001

~ ... ...... ...... ----------------

7

YEARS

9

SURGERY ALONE

11 13 15

p E R c

100

E 50 N T

B 3 5

LOG-RANK: P = 0.02 WILCOXON: P = 0.02

CMF

SURGERY ALONE

7

YEARS

9 11 13 15

FIGURE 40-12. First Milan trial comparing adjuvant chemotherapy (CMF) with surgery alone (control): 15-year results in premenopausal women. (A) Compartitive relapse-free survival. (B) Comparative total survival. (Bohadonna G. Evolving concepts in the systemic adjuvant treatment of breast cancer. Cancer Res 1992;52[8]:2127-2137)

DOXORUBICIN COMBINATIONS

Twenty years after its first reported clinical use, the anthracycline doxorubicin remains the single most effective dnig in the treatment of advanced breast cancer. 432 Combination chemotherapy regimens that include doxorubicin or its analog epirubicin have repeatedly been shown to induce a higher response rate in patients with locally advanced or clinically disseminated breast carcinoma th:in regimens that do not contain an anthracycline. However, the median duration of response and survival have not been unquestionably superior using an anthracycline. The initial reluctance to include

p E R c E 50 N T

"-.

3

...... ...... ...... ......

5

~,

.......... cMFx6 ............. __ _ ""--........ ____ _

CMF x 12

7 9 11 13

YEARS

FIGURE 40:-13. Second Milan trial comparing adjuvant chemo' therapy (CMF) for 6 and 12 cycles: comparative relapse-free survival at 14 years. (Bonadonna G. Evolving concepts in the systemic adjuvartt treatment of breast cancer. Cancer Res 1992;52[8]:2i27-2137)

doxorubicin in an adjuvant situation was due to its potential for myocardial damage after prolonged treatment.

The M.D. Anderson Cancer Center and the University of Arizona tested the potential efficacy of adjuvant doxorubicin in nonrandbmized studies but did not document a clear superiority for doxorubicin-containing regimens over CMF or CMFP. 43

"434 After it became clear that adjuvant chemotherapy

could be limited. to about 6 months, randomized trials with doxorubicin-containing regimens were started in Europe and in the United States.286·305·427·429- 431 •435 Possibly due to the low dose (30 mg/m2 every 3 weeks) of doxorubicin administered in combination with PF (L-PAM and fluorouracil) by the NSABP group, patients treated with this anthracycline showed only a marginal advantage at 5 years compared with patients

'given PF alone and similar outcome if adjuvant tamoxifen was added to both drug combinations. 435 In a subsequent NSABP trial (B-15) consisting of tamoxifen-nonresponsive tumors, node-positive patients were randomized to one of three arms: doxorubicin (60 mg/m2) plus cyclophosjlhamide (600 mg/m2

) given every 3 weeks for four cycles (AC regimen); AC followed 6 months later by three monthly cycles of a slightly modified intravenous CMF; or standard CMF for six monthly cycles. The 3-year results did not reveal any significant difference in outcome among the three arms; the relapse-free survival rate was 62% for AC; 6So/o for AC followed by CMF; and 63% for CMF. The investigators concluded that AC chemotherapy should be preferred to CMF because of the shorter treatment duration and lesser toxicity. 430 In node-positive patients with a median of seven positive riodes treated at the Dana-Farber Cancer Institute, 15 and 30 weeks of the AC regimen consisting of doxorubicin ( 45 mg/m2

) plus cyclophosphamide (500 mg/m2

) given intravenously every 3 weeks were compared. At the 8-year analysis, the diseasefree survival rate was comparable in the two arms, indicating that there was no significant advantage for the use of longer durations of AC.429


TABLE. 40-22. Effective Combination Chemotherapy Regimens Commonly Used to Treat Brea.st Cancer

Dose Day(s) of .. ,RecycleC Regimen (mglm') Route Treatment

CMF (P) Cyclophosphamide 100 PO I to 14 Methotrexate 40 (60) I.V. I and 8 4wk Fluorouracil 600 (700) LV. I and 8 Prednisone (40) (PO) (I to 14)

CMF.

Cyclophosphamide 600 I.V. Methotrexate 40 LV. 3wk Fluorouracil 600 LV. I

CMF Cyclophosphamide 600 I.V. I and 8 Methotrexate 40 LY. I and 8 4wk Fluorouracil 600 I.V. I and 8

CA Cyclophosphamide 200 PO 3 to 6 3-4wk Doxorubicin 40 I.V. I

AC Doxorubicin 45 I.V. I 3wk Cyclophosphamide 500 I.V. I

FAC Fluorouracil 500 LV. I and 8 Doxorubicin 50 LV. I 4wk Cyclophosphamide 500 I.V. I

Recent results oh the use of adjuvant doxorubicin are available froin the :Milan Cancer Iristitute.304

·305

·431 Patients with

one to three positive nodes were randomized to receive 12 courses of intravenous CMF every 3 weeks (see Table 40'-22) or intravenous CMF for eight courses followed by four courses of doxorubicin (75 mg/m2) given every 3 weeks. The 5-year results did not show improved outcome with the use of doxorubicin compared with CMF alone (relapse-free survival rates: 72% versus 74%). In another trial involving patients with more than three positive nodes, sequential administration of doxorubicin (as previously described) followed by eight qourses of CMF (as described) yielded significantly improved treatment outcome at 5 years (relapse-free survival rates: 61 % versus 38%) in both inehopausal groups compared with the alternating administration of the same drug regimens. The superiority of the sequential to the alternating chemotherapy is stiil observed at the 6-year analysis in this subset of patients (Fig. 40-14).304 These findings suggest that the use of doxorubicin and this scheduling are responsible for this improved outcome. If the Milan data with doxorubicin followed by CMF are confirined by other research groups, this sequence (i.e., four cycles of doxorubicin followed by four cycles Of intravenous CMF on days 1 and 8; see Table 40-22) could

· replace CMF as the standard adjuvant treatment in patients with node-positive breast cancer, regardless of inenopausal status.


PERI OPERATIVE CHEMOTHERAPY

Perioperative chemotherapy refers to the administration of a single course of chemotherapy (sirigle agent or drug combination) within a few days after surgery. Studies testing the value of a single dose of perioperative chemoiherapy in nodepositive and node-negative breast cancer patients have not shown any consistent benefit. The Ludwig Trial V, testing the value of a single cycle of chemotherapy given perioperatively, included a total of 2504 evaluable patients. In 1229 nodepositive patients, the study demonstrated that one cbutse of CMF was inferior to the same perioperative chemotherapy followed by six monthly cycles of standard adjuvant combination chemotherapy or standard combination chemotherapy without perioperative treatment."' Jn 1275 node-negative p:itients, the use of perioperative CMF was associated with a marginal benefit at 5 years compared with no treatment (7 4% versus 68% for controls). 437 This benefit was significant in ERnegative tumors, particularly in the postmetiopausal subset (79% for CMF, 56% for controls).

Investigators at the Cancer Institute in Genoa designed a siUdy protocol using randomization to one cotirse of CEF ( cyclophosphamide, epidoxorubicin, and fiuorouracil) or to no perioperative treattnent.438 Node-negative patients did riot receive further therapy; and node-positive women were given CEF altetnatingwith.CMF for a total of 12 cycles given every 3 weeks. This trial, which included 467 patients, failed to reveal any significant difference in treatment outcome at 5 years. The findings do nbt support the hypothesis that perioperative Cheffiotherapy is an important strat~gy in the management of breast cancer. These clinical findings are in agreement with the experiniehtal studies that demonstrated that tumor cell killing by drugs follows first-order reaction kinetics. Unless perioperative chemotherapy involves the use of very -high doses, one treatment cycle is insufficient to yield a consistent benefit.

DRUG DOSE INTENSITY AND HIGH-DOSE CHEMOTHERAPY

Except for the initial studies with polydrug regimens in acute leukemia and Hodgkin's disease, it was only during the past decade that clinicians began to appreciate that treatment failure could be the consequence of instlfficient dose int~nsity. 439

In the area of adjuvant therapy of breast cancer, the Milan CMF program was the first to provide retrospective evidence that the outcome at 5 and 10 years was related to the use of full doses of drugs.'""" Despite the intrinsic limitations of retrospective evaluations, these findings generated the irtitial interest in the subject of dose intensity.

Dose intensity is a concept developed in 1984 by Hryniuk and coworkers, who emphasized received dose rate, rather than total ddse received. 440 The concept is complex, because it is a inathematical combination of the total amount of drug received and the amount of drug received per unit time (e.g., mg/m2/week). Hryniuk and colleagues evaluated the relation between reiapse-free survival and the dose intensity of CMF adjuvant chemotherapy in node-positive breast cancer (Fig. 40-15).441 This analysis showed a highly sighificant relation between projected dose intensity and the 3-year relapse-free survival within all four subsets of patients: one to three positive


1306

p E A c

100

E 50 N T

A


LOG-RANK: P =.0.002

3

YEARS

1001---~

CMF/ADM

5

B

~----------............ ADM +-CMF --

LOG-RANK: P "' 0.002

3

YEARS

--------. CMF/ADM

5

FIGURE 40-14. Milan trtal comparing doxorubicin-containingregimens: 6-year results in patients \Vith more than 3 positive axillary lymph nodes. (A) Comparative relapse-free survival. (B) Comparative total survival. (Bonadonna G. Evolving concepts in the systemic adjuvant treatment of breast cancer. Cancer Res 1992;52[8],2127-2137)

ALL PATIENTS 100

90

80 * •Control

\b c>'" A CMFVP

:g- X CMFP

0 70 • Cj;F ~ C. cyclophosph•mld• UJ * * u.

c;;~ ::{)

V CMFV

a: 0 CMF

a: 60 ~Cj;MF q, Phen}'lal1nlne

>- mustard

o; llq; •• TrW with radio-u .. ,..,., llddecl

50 * Levell of CMF chemolher.py

r:0.76 .ccordlng to Bonadonna"

40 P<.01

30 f I I I

0 0.2 0.6 1.0 DOSE INTENSITY

FIGURE 40-15. Three-year relapse-free survival in stage II breast cancer related to the average relative dose intensity of the adjuvant chemotherapy used. The size of the symbols is proportional to the number of cases at each dose intensity. (Hryniuk WM, Bonadonna G, Valagussa P. The effect of dose intensity in adjuvant chemotherapy. In: Salmon SE, ed. Adjuvant therapy of cancer V. Philadelphia: Grune & Stratton, 1987,13-23)

nodes or more than three positive nodes in premenopausal and postmenopausal women. Henderson and associates summarized-the. results of retrospective analyses of chemotherapy dose and treatment outcome in adjuvant chemotherapy trials. 442 Although conclusions from data-derived results may be biased, it appears that the trials refuting the dose-response hypothesis mostly employed single-agent or low-dose combination chemotherapy. Recent prospective randomized trials carried out using patients with advanced breast cancer have shown that the use of higher doses of CMF is superior to lower doses.443.444 Medical oncologists must be aware that frequent dose reductions or treatment delays may compromise treatment results. It is considered preferable to delay treatment for a few days for myelosuppression rather than to reduce the drug dose by 50%.

Dose-intensive adjuvant chemotherapy is being evaluated in patients with 10 or more positive axillary nodes. The Johns Hopkins Cancer Center devised a five-drug regimen using weekly chemotherapy, sequential administration of antimetabolites, and continuous infusion of fluorouracil to minimize dose reduction and treatment delay. 445 At a median follow-up of 17 months, only 8 of the 53 treated patients relapsed. This encouraging preliminary outcome, the manageable side effects of the regimen, and the demonstrated ability to deliver more than 90% of the planned doses provide the rationale for a phase III comparison of this new dose-intense regimen to standard chemotherapy.

During the 1980s, the use of high-dose chemotherapy and autologous bone marrow transplantation was shown to produce rapid and complete responses in patients with advanced breast cancer. 446 With the exception of a few women presenting with small tumor burdens, sustained complete remissions were limited. The limited success can be attributed to drug-resistant


tumor cells in these relapsed patients, who usually had large tumor burdens. The current strategy of bone marrow transplantation focuses on the adjuvant setting for patients with 10 or more positive axillary nodes. Without absolute drug resistance, increasing the dose intensity should increase the log kill and improve the results.

Table 40-23 presents the treatment program of the Milan Cancer Institute, which consists of the sequential administration of cyclophosphamide, vincristine, methotrexate plus folinic acid, cisplatin, and melphalan. 304 It is important to point out that patients were also treated with recombinant human granulocyte-macrophage colony-stimulating factor (rhGMCSF) or interleukin-3 (rhlL-3) as a means of harvesting by leukapheresis large quantities of peripheral blood precursors capable of accelerating hematopoietic recovery after subsequent myeloablative chemotherapy (Fig. 40-16).447 To minimize hematologic toxicity_ after high-dose melphalan, cryopreserved circulating hematopoietic cells, harvested during the leukocyte rebound phase after the postcyclophosphamide nadir, ':Vere autotransplanted in combination with bone marrow cells. Although clinical data are still premature (i.e., only 48 patients were treated, with a median follow-up period of 21 months at the end of 1991), the 2-yearrelapse-free survival was 93%. 448 Patient tolerance of this therapy was excellent, and there were no life-threatening complications. These findings, which compare favorably with those from Duke University (i.e., 53 patients, 80% 3-year relapse-free survival rate), appear superior to those achieved with the best available adjuvant therapy (i.e., doxorubicin followed by CMF), for Which the 2-year relapse-free survival rate for the subset with more than 1 O positive nodes is only 60%. 305

•446 A longer follow

up is necessary to demonstrate that the approach using growth factor-supported high-dose sequential adjuvant chemotherapy is superior and can be used effectiv~ly in properly selected patients outside a research center.

TABLE 40-23. The Milan High-Dose Chemotherapy Regimen for the Adjuvant Treatment of Patients Having More Than 10 Positive Axillary Nodes

Intravenous Approximate Dose Scheduled

Regimen (mglm') Day

Surgery -30 to -45 Cyclophosphamide 7000 0 rhGM-CSF (µg/kg/day)• 5 1 to 14-16 Leukaphereses (CPC harvest) 13-16 Bone marrow harvest 17 Vincristine 1.4 18 Methotrexate plus rescue 8000 18 Cisplatin 60 23 and 30 Melphalan 200 40 Marrow plus CPC autoinfusion 41 Radiation therapy ( 4500 cGy) 6.0

• rhGM-CSF, recombinant human granulocyte-macrophage colonystimulating factor; CPC, circulating progenitor cells.

100,000

:;_ --!!? c: 10,000

8

:i. -"' -c: " 8 ;; ;; -"'

1,000

100

10

300,000

250,000

200,000

150,000

a. 100,000 c:

"' .. :;: 50,000


--BMT+CPAT(rhGM-CSF) , (with or without rhGM-CSF)

········BMT only

-4 0 8 12 16 20 20

.................

-4 0 8 12 16 20 24 Days from autografting (day Ol

FIGURE 40-16. Effect of rhGM-CSF-mobilized hematopoietic progenitor cells on granulocyte (upper panel) and platelet (lower panel) recovery after high-dose melphalan (200 mglm2) treatment and bone marrow transplant (BMT). Compared with the 15 controls receiving BMT only, the 14 breast cancer patients who received circulating progenitor autotransfusion (CPA1) experienced a much faster median recovery, which was particularly striking for platelets. (Modified from Gianni AM, Bregni M, Siena S. Recombinant human granulocyte-macrophage colony-stimulating factor reduces hematologic toxicity and widens clinical applicability of high-dose cyclophosphamide treatment in breast cancer and non-Hodgkin's lymphoma. J Clin Oncol 1990;8:768-778)

DRUG-INDUCED AMENORRHEA AND TREATMENT OUTCOME

Based on adjuvant trial results indicating the treatment benefit was better for premenopausal than for postmenopausal patients and the relatively high frequency of drug-induced amenorrhea, many clinicians hypothesized that the mechanism of action of adjuvant chemotherapy (which included alkylating agents) was chemical castration.425

·427

•449 However,

the results of trials carried out by the NSABP and Milan groups indicated that women who developed amenorrhea as a result of adjuvant treatment failed to show a longer disease-free interval or increased overall survival than patients who did



not develop amenorrhea.450·451 In the Milan series, premenopausal women_who relapsed were equally likely to respond to subsequent ovarian castration whether or not they had develOped prior amenorrhea. 286

•452 Although some investigators

reported that the effect of drug-induced amenorrhea was documented almost exclusively in ER-positive tumors, adjuvant cytotoxic therapy has been equally efficacious in patients with ER-negative and ER-poSitive tumors. 453

•454

Two cooperative study groups reported the endocrine effects of adjuvant chemotherapy in premenopausal breast cancer patients. Bianco and coworkers evaluated 221 patients who received adjuvant CMF for six or nine cycles with or without tamoxifen.455 They found a significant correlation between drug-induced amenorrhea and subsequent treatment outcome. Drug-induced amenorrhea was age related, with women closer to menopause more likely to develop amenorrhea. The existence of a statistical association does not necessarily mean that drug-induced amenorrhea is causally related to a superior treatment outcome. The reported findings may represent an increasing ovarian sensitivity to alkylating agents in older women approaching physiologic menopause. In 387 patients given adjuvant CMFP for at least six cycles, Goldhirsch and colleagues documented a marginally significant difference in the 4-year relapse-free survival rate, favoring women with drug-Induced amenorrhea (68%) over women without amenorrhea (61 %).456 This effect was almost exclusively observed in patients 40 years of age or older and was unrelated to ER status. The investigators thought it was unlikely that the entire effect of cytotoxic adjuvant therapy was mediated through endocrine manipulation, because the treatment effects of chemotherapy were much larger than the outcome differences observed for women who did or did not have amenorrhea. 456

The preponderance of evidence suggests that the major benefit from adjuvant chemotherapy is related to its direct cytotoxic effect on micrometastatic breast cancer, .but a possible treatment effect on the female endocrine axis cannot be excluded.

ADJUVANT ENDOCRINE THERAPY

The concept that at least a fraction of breast tumors are hormor:te dependent and therefore responsive to various endocrine manipulations is an old one, supported by innumerable laboratory and clinical observations. 424

•457 The discovery of

steroid receptors in the mid-1970s provided the rationale for a more selective application of endocrine tr~atment. In advanced disease, endocrine ablative (e.g. 1 surgical castration, ovarian irradiation) or additive endocrine therapies have a response rate of 30% in unselected patients, of about 50% in ER-positive patients, and of 70% to 80% if ER and progesterone receptors (PGR) are positive. Only 5% to 10% of patients with receptor-negative tumors respond to hormonal manipulations. The response rates are proportional to the level of hormone receptor measured by the dextran-coated charcoal method, and the receptor status in the primary tumor appears to correspond largely to that of metastases.

Since the initial proposal of Schnitzinger in 1889 to use adjuvant oophorectomy in breast cancer patients, there have been more than 20 trials of various types of ovarian ablation. The earliest endocrine studies were limited by the same de-

fects as the studies using adjuvant chemotherapy performed before 1972: use of historic, nonmatched, or nonrandomized controls and insufficient information on nodal status. An analysis of the results was further hampered by the absence of adequate statistical methodology. 424

'457 Several recent trials

with ovarian ablation, including several prospective randomized studies, have shown improvement in recurrence-free and overall survival rates in premenopausal women. 424

·457

-459 The

trial that was instrumental in reviving interest in the use of adjuvant ovarian ablation was the Toronto study-460 In. this trial, patients 45 years or older were randomized, after a radical mastectomy, into one of three groups: radiation-induced menopause, radiation-induced menopause followed by 5 years ofprednisone (7.5 mg daily), or no adjuvant systemic therapy_ Although postmenopausal patients showed no survival gain from either of the adjuvant regimens, patients who were premenopausal at entry and given prednisone and castration had a significantly increased survival rate at 10 years compared with the group given no adjuvant therapy. In this trial, patients younger than 45 years were not included in the prednisone therapy arm, and the use of ovarian ablation alone in this group was not associated with a statistically significant survival advantage compared with the no-treatment group.

Adjuvant chemotherapy is being compared with adjuvant castration in premenopausal patients in Scotland in a trial initiatedin March of 1980. By May 1990, 332 premenopausal patients with node-positive operable breast cancer regardless of ER status were randomized to receive intravenous CMF or to undergo ovarian ablation, each with or without prednisolone for 5 years. The low rate of accrual is a function of the relatively small number of premenopausal patients with breast cancer, the reluctance of surgeons to disregard ER status (known for 80%) in assigning adjuvant therapy, and the reluctance of patients to be entered onto the trial. The initial local treatment was by mastectomy or local excision and sampling or clearance of the axilla, with immediate postoperative radiation therapy if indicated_ Although the preliminary reported findings suggest no difference in the recurrence rate between CMF and ovarian ablation, the trial is still open, and formal analysis is not yet available.458

Another form of adjuvant hormonal treatment is the use of ihe antiestrogen tamoxifen. Two characteristics oftamoxifen encouraged research clinicians to assess its role as an adjuvant therapy. It is known to be effective in postmenopausal receptor-positive breast cancer patients with metastatic disease, especially those with limited disease and a disease-free interval in excess of 2 years, and there are fewer side effects compared wiih chemotherapeutic agents.

In 1977, the Nolvadex Adjuvant Trial Organization designed the first modern adjuvant trial with randomization to tamoxifen or to a no-treatment control arm. In a series of 1131 patients with node-positive or node-negative breast cancer, the 5- and IO-year results provided evidence that overall survival was moderately, but significantly, improved with the use of adjuvant tamoxifen.461 Based on the preliminary results of this trial, additional trials testing the value of adjuvant tamoxifen were initiated among major research centers. The results of large randomized studies, activated in Europe during the late 1970s, confirmed the benefit of adjuvant tamoxifen, especially in patients presenting with strongly ER-positive tu-

r '


mors. 462- 464 Based on these results, tamoxifen has become

widely used in clinical practice during the past decade, primarily in postmenopausal patients with ER-positive tumors and with one to three positive nodes. The recommended dose is 20 mg per day for a minimum of three consecutive years. 465

•466 However, based on the available results, many

oncologists believe that treatment with adjuvant tamoxifen should be given for 5 or more years, if not indefinitely, in patients remaining disease free. The benefits of prolonged administration of tamoxifen need to be balanced against potential long-term toxicity, and this is being addressed in current trials.

CHEMOTHERAPY WITH ENDOCRINE THERAPY

With the results achieved using adjuvant chemotherapy and adjuvant tamoxifen, it is logical to ask whether the combined use of cytotoxic and endocrine modalities results iri an additive or synergistic effect. The biologic rationale for combining these two treatment modalities is tumor cell heterogeneity. It is likely that breast tumors consist of various populations of cells with different sensitivities to cytotoxic and hormonal agents, and these two modalities may be effective in killing different tumor populations. During the past decade, a few research groups have tested adjuvant chemoendocrine treatment, particularly in postmenopausal patients. The only large study that yielded a positive result, favoring chemoendocrine treatment (PF plus tamoxifen) over chemotherapy (PF) alone, is that of the NSABP. The investigators observed that the benefit was almost entirely restricted to the subset of patients older than 50 years of age with more than four involved axillary nodes and with positive estrogen and progesterone receptors. 467 Trials from the Eastern Cooperative Oncology Group468 and the Mayo Clinic469 failed to show a significant advantage for the addition of tamoxifen to combination chemotherapy. There is not yet clear evidence from individual trials that the addition of tamoxifen to effective chemotherapy (CMF or CMF-like regimens) has improved treatment outcome over chemotherapy alone. However, the addition of chemotherapy to tamoxifen has consistently resulted in superior treatment outcome compared with adjuvant tamoxifen alone (given for 3 to 5 years) in postmenopausal patients with ER-positive tumors. 470

-472

The pharmacologic reasons for these disparate clinical findings are unknown. In vitro studies demonstrated that endocrine therapies may decrease the cytotoxic effect of chemotherapeutic drugs by unfavorably altering tumor cell kinetics, and this effect may occur clinically if chemotherapy and hormonal treatments are delivered concurrently.473 It is important to compare sequential and concurrent administration of the two modalities in properly designed clinical trials. It may be determined that relatively short intensive chemotherapy should be given first to kill rapidly proliferating tumor cells and that tamoxifen should be given after chemotherapy to kill or retard the growth of slowly proliferating cells. Although the potential benefit of giving tamoxifen after cessation of chemotherapy was suggested by studies of the NSABP465

and ECOG,466 this important issue remains unsettled. Additional studies are required to develop effective combined chemotherapy and endocrine therapy regimens.


ADJUVANT THERAPY IN NODE-NEGATIVE PATIENTS

For many decades, it was believed that histologically nodenegative breast cancer represented an almost invariably curable disease. This was a logical consequence of the Halstedian hypothesis of cancer spread. About 20 years ago, however, the evaluation of large surgical series indicated that the overall 10-year relapse-free survival rate after radical mastectomy in this patient subgroup was only about 70% to 75%. 136·138 As in patients with node-positive tumors (regardless of tumor size and menopausal status), about half of all recurrences manifest within the first 3 years after locoregional treatment. 474 Research efforts using new prognostic variables indicate that, in given subsets of node-negative patients, the distant recurrence rate is as high as 50%. These clinical findings are in agreement with the modern hypothesis that stresses the importance of early hematogenous dissemination of cancer cells.

During the late 1980s, the reports of a few randomized trials using node-negative patients with ER-negative tumors supported the strategy of using adjuvant systemic therapy in high-risk subsets of node-negative patients. A Clinical Alert released in May 1988 by the National Cancer Institute, in which the results of these trials were described before their publication, informed physicians and women about the need to reconsider the overall prognosis of node-negative breast cancer and encouraged their participation in clinical trialsGil an attempt to refine patient subsets and treatment outcome.

The results of combination chemotherapy in node-negative and receptor-negative tumors are summarized in Table 40-24.215·304·475·476 Although the three series have some dissimilarities, the use of adjuvant treatment yielded comparable results at follow-up of 5 or more years. Combination chemotherapy reduced the annual odds of recurrence by at least 30% in patients with node-negative breast cancer. In the Milan trial, the 8-year total survival was significantly improved using adjuvant chemotherapy, with no difference in benefit between premenopausal and postmenopauSal patients. 304

•475 Based on

these results, CMF(P) for six monthly cycles is recommended for node-negative patients who are selected to receive adjuvant chemotherapy. As for node-positive patients, chemotherapy should be started within 4 weeks after surgery and administered at full dose with minimal dose reductions for myelosuppression.

Should all node-negative patients be treated with adjuvant chemotherapy? Currently, the 'answer is no. Although there is no established method to integrate the various morphologic and biologic indicators of prognosis, it is possible to identify patients with a low ( <15%) risk ofrecurrence. 477 This includes patients with tumors characterized by tumor size less than 1 cm in diameter, grade I malignancy, with positive steroid receptors, particularly if ER and PGR are highly positive, or with a low proliferative rate measured by flow cytometric DNA analysis (S phase). Most of these patients can be spared chemotherapy. ·The converse is true for patients having tumors larger than 3 cm in the largest diameter, grade III tumors, negative steroid receptors (particularly if both types are negative), or high S-phase components. These patients have a risk of recurrence that is greater than 30%. If two or more of these unfavorable prognostic factors are present (e.g., large' tumor size and high S-phase component), adjuvant chemo-



TABLE 40-24. Chemotherapy in Node-Negative Breast Cancer: Results of Modern Adjuvant Trials in Receptor-Negative Tumors

Study Factors Milan'°'·"' NSABP 215 Intergroup478

Selection Criteria Menopausal status Pre and post Tumor size Tl-T3a Estrogen receptor status Negative

Regimen Intravenous CMF Number of patients 90 Median follow-up Sy

Relapse-Free Survival(%) Control group 39 Chemotherapy 80 p value 0.0002

Pre and post Any ·Negative

M-F 737 5y

66 78 0.0007

Pre and post Any Negative: all T Positive: T > 3 cm CMFP 425 5y

61 83 <0.0001

C, cylophosphamide; M, methotrexate; F, fluorouracil; P, prednisone.

therapy is strongly advised. Additional research into the identification of prognostic factors should simplify this process.

The current indications for adjuvant endocrine therapy in node-negative tumors are not well defined. Although relapsefree survival has been significantly improved by the use of prolonged tamoxifen in a large series of patients with ERpositive tumors, the advantage in terms of total survival remains somewhat questionable. 216 However, the use of tamoxifen has been associated with favorable effects on bone and blood lipids and is relatively nontoxic.

The 1990 Consensus Development Conference suggested that the oncologist have a thorough discussion with the patient in a:rriving at a final decision.206 The clinician should first estiinate the risk ofrecurrence for that individual patient after locoregional therapy alone and then should explain the potential benefits and risks of endocrine and cytotoxic treatments.

INDICATORS FOR SELECTING ADJUVANT TREATMENT IN NODE-NEGATIVE PATIENTS

Laboratory studies and clinical experience have demonstrated that breast cancer is a highly heterogenous disease in its pathologic and clinical behavior. It is not surprising that the clinical course of this malignancy-the risk of recurrence and the response to systemic therapy-often varies. The major determinants of prognosis are tumor cell burden and drug resistance. During the 1970s, the criteria defining high-risk patients were the presence and number of histologically involved axillary nodes, tumor size, and tumor grade. These firmly established prognostic features were later followed by steroid receptor values and assessments of tumor cell proliferative activity (e.g., thymidine labeling index, DNA ploidy, and S-phase fraction by DNA flow cytometry). Because even these indicators are not absolute, clinicians need additional information to better define tumors associated with poor or good prognoses.

A variety of biochemical and immunocytochemical variables have been the subjects of numerous studies (Table 40-25).

There is abundant literature on the putative utility of these individual prognostic indicators, which does not enable the clinician. to compare one factor with the others and to ,.§_elect the most useful ones among them. The sharpness of p 'values in any given series is not enough to provide complete information. Provisional data from many research groups conflict about the additional predictive potential of these newer prognostic factors compared with conventional prognostic factors. The constellation of newly proposed variables, such as ERBB2 oncogene amplification, epidermal growth factor receptor, protease cathepsin D, stress-response proteins, monoclonal antibodies for detecting peritumoral lymphatic and blood vessel invasion, high-molecular-weight mucin-like antigens, a variety of serum tumor markers, and timing of breast cancer surgery during the menstrual cycle, have been claimed to influence treatment outcome and to help in the treatment de-

TABLE 40-25. Risk Factors for Relapse in Node-Negative Breast Cancer

Established

Primary tumor size Histologic tumor grade Steroid receptor status Tumor cell proliferative activity

Requiring Additional Investigation DNA ploidy Oncogene amplification (e.g., ERBB2) Protease cathepsin D Epidermal growth factor receptor Stress-response proteins Haptoglobin-related protein Laminine receptor expression Serum tumor markers (e.g., CA 15-3) Detection of occult bone marrow micrometastases by monoclonal

antibodies


cision, but they have left the average clinician confused. What is required to resolve this issue is a sufficiently large series of consecutive patients who have been uniformly staged, managed, and evaluated using the established and the newer prognostic factors. 477

•478 It is less critical to determine the under

lying biologic principles or laboratory evidence supporting these various factors than to determine their utility in making treatment decisions. The search for new biologic indicators of prognosis remains important for clinicians, but the novel markers should be "handled with care" in the selection of high-risk patients.

THE INTERNATIONAL OVERVIEW

In 1984, an international overview (i.e., metaanalysis) of the effect on mortality of the various adjuvant systemic treatments (e.g., chemotherapy, endocrine therapy, immunotherapy) was organized by the Early Breast Cancer Trialists' Collaborative Group with the cooperation of several individual investigators who contributed their data. Performing a metaanalysis requires the use of proper statistical procedures, which includes obtaining treatment results in all trials, published and not published, and evaluating the effects from each trial individually before combining them. The major advantage of metaanalysis is the ability to demonstrate a moderate, but clinically relevant, reduction in mortality in a disease that is common. Such a reduction in mortality may not be demonstrated in individual randomized studies because of insufficient patient numbers, even in trials accruing 300 to 400 patients.

The first metaanalysis was conducted in 1987 for a total of about 30,000 patients and evaluated the 5-year results of adjuvant treatment. When all drug treatments (e.g., single agents and various combinations) were considered, the analysis demonstrated a significant overall reduction in the annual odds of death for patients treated with chemotherapy, compared with no-treatment controls.449 When only CMF-type regimens were analyzed, the reduction for all patients was 23%, with a 37% reduction for patients younger than 50 years old and a 9% reduction for older patients. The use of tamoxifen was associated with an overall reduction in the odds of death of


16%, and among women older than 50 years, the reduction was 20%.

The updated metaanalysis at 10 years was conducted in 1991 for a total of75,000 women enrolled in 133 randomized trials involving 31,000 ( 41 %) recurrences and 24,000 (32%) deaths. Thirty thousand women were included in tamoxifen trials, 11,000 in' polychemotherapy trials, 15,000 in other chemotherapy comparisons, 3000 in ovarian ablation trials,

. and 6000 in immunotherapy trials.479 Table 40-26 summarizes the essential data and demonstrates the following points:

!. The long-term (IO-year) results essentially confirmed the intermediate (5-year) results.

2. The cumulative difference in total survival was larger at 10 than at 5 years. Tamoxifen produced a highly significant mortality reduction during the first 5 years and an additional mortality reduction during the next 5 years. For polychemotherapy, the absolute mortality difference at 5 years was about doubled by I 0 years.

3. Longer courses of chemotherapy (e.g., 12 months) were no more effective than shorter courses (e.g., 6 months).

4. At the IO-year analysis, polychemotherapy was documented to reduce the annual odds of mortality, even in women 50 years of age or older (relative reduction = 14±5%). c:

5. For the ER-poor ( <10 fmol/mg cytosol total protein) subset, the IO-year mortalit)> reduction after adjuvant tamoxifen was only 11 %, compared with 21 % for the ER-positive subset.

6. Among the 17 46 women younger than 50 years, ovarian ablation was associated with a highly significant reduction in the annual odds of recurring (26%) and of dying (25%).

7. The proportional annual odds reductions were similar for node-positive and node-negative patients, but the absolute improvement in· 10-year survival was about twice as great for the former as for the latter group.

8. The use of adjuvant immunotherapy was not associated with a significant improvement in recurrence-free survival or in mortality.

TABLE 40-26. Main Results of the Updated International Overview

Age <50 Years• Age ;,,50 Years•

Types of Systemic Adjuvant Recurrence Death Recurrence Death Therapy Compared (%±SD) (%±SD) (%±SD) (%±SD)

Polychemotherapy alone vs no treatment 37± 5 27± 6 22±4 14 ± 5

Chemotherapy plus tamoxifen vs chemotherapy 7± 4 3± 5 28 ± 3 20±4

Tamoxifen for about 2 years vs no treatment 27 ± 7 (17 ± !O]t 30±2 19 ± 3

Chemotherapy plus tamoxifen vs tamoxifen [32 ± 16]t (-6 ± 23]t 26±5 10 ± 7

• Data are expressed as a relative reduction in the annual odds of recurrence or death from any cause at 10 years. t Brackets denote statistically unstable results with SD :?:. 9. :(:Only 383 patients (Early Breast Cancer Trialists' Collaborative Group, 1992).407



The updated results of the International Overview should have a major impact on clinical practice, because they confirm the results from the individual trials and indicate that the benefit is greater at 10 years than at 5 years. The validation of the benefit from adjuvant systemic treatments reemphasizes the importance of the biologic concepts that are the basis of this treatment. The highest recurrence rates in treated patients were documented to be within the first 3 to 4 years, implying that a considerable fraction of patients have resistant tumor cells and that they do not benefit from the prolonged administration of the same polychemotherapy. Despite the considerable amount of information contained in the Overview, the results do not necessarily reflect the evolving state of the art, nor should they be expected to provide a detailed guide for physicians in the choice of treatment outside the context of clinical trials. For example, the most recent metaanalysis did not evaluate the use of full- and intensive-dose chemotherapy, sequential drug regimens, or the role of anthracyclines. The updated International Overview provides evidence for the long-term benefit of adjuvant systemic therapy and a measure of what a practicing physician can reasonably expect from "standard" adjuvant therapy. This is particularly true for the use of tamoxifen, because most patients were treated with 20 mg/day for a minimum of 2 years.

Polychemotherapy includes a heterogenous group of treatments delivered by means of various drugs, doses, and intensities. It is possible that the results of the metaanalysis may underestimate the benefit seen using full-dose standard therapy. This also applies to patients treated with CMF, the most widely tested drug combination, although this drug regimen was "the only one which separately demonstrated survival advantage in the present overview.' '479

TOXICITY OF ADJUV ANT THERAPY

Treatment-related morbidity is still a major concern in the use of adjuvant chemotherapy. Table 40-27 summarizes the early toxic effects and their average frequency for the four most commonly used polydrug regimens. Marked alopecia (requiring the patient to wear a wig) occurs in fewer than 10% of patients after CMF, but it is more frequently observed in women receiving CMFVP and is even more common if using an doxorubicin-containing regimen, such as FAC or AC. After CMFVF administration, paresthesias can be documented in about half of women. About a third of patients undergoing treatment complain of anticipatory vomiting, which obviously reflects the patient's psychological distress. Weight gain during adjuvant chemotherapy has been consistently documented in at least half of women, with an average gain of 3 to 4 kg. This has occurred irrespective of menopaus:il status or. whether drug-induced amenorrhea occurs in preinenopausal patients. Although none of 3000 women treated at the National Cancer Institute of Italy with adjuvant CMF alone or combined with doxorubicin required supportive therapy for life-threatening toxic effects, other trial reports indicate that toxic deaths may occur in fewer than 0.5% of patients.

Thromboembolic complications after adjuvant chemotherapy were recently described by several research groups, but the actual incidence of these complications remains unknown. 468·480.4st They are probably uncommon. In the Milan 15-year experience with CMF chemotherapy, these episodes

TABLE 40-27. Acute and Late Toxicity From Commonly Used Adjuvant Chemotherapy Combinations

Toxicity

Acute Vomiting Oral mucositis Marked alopecia Leukopenia* Thrornbocytopeniat Conjunctivitis Cystitis

Delayed Amenorrhea Cohgestive heart failure

Late

Acute leukemia

"' <2500/mm3.

t <75,000/mm3.

CMF CMFVP (%) (%)

>90 >90 <ID 25 slD 40 slD <12 <ID <15 s30 s15 s15

70 70 0 0

<I <I

FACorAC (%)

>90

>90 s25 <5

<5

80 1-2

<I

(Data are average~ derived from the published literature)

were almost never documented. It is possible that the type of drugs used and the duration of chemotherapy are partially responsible for these events. In a survey of ECOG studies, thromboembolic complications were more than doubled when prednisone (1.5%) and prednisone plus tamoxifen (3.5%) were added to CMF (0.5%). 468

The major delayed toxicity of adjuvant chemotherapy is irreversible amenorrhea, which is secondary to the administration of alkylating agents, such as cyclophosphamide. The incidence of amenorrhea is age related ( <40 years: 40%, >40 years: 95%), and it is reversible in about 40% of women younger than 40 years of age. The available results suggest that the current adjuvant treatments have not been followed by a high incidence of chronic organ failure. After FAC or AC chemotherapy, the incidence of congestive heart failure is only I% to 2% if the cumulative dose of doxorubicin does not exceed 300 to 350 mg/m2 • The Dana-Farber Cancer Institute reported that the frequency of congestive heart failure increased from 1.6% to 6% when the duration of AC chemotherapy was prolonged from 15 to 30 weeks. 429 The increase in the frequency of laboratory-proven and symptomatic myocardial damage in patients treated with conservative surgery and irradiation to the left breast combined with doxorubicincontaining regimens partially depended on the technique of irradiation used.

The primary toxic effects associated with high-dose chemotherapy and autologous bone marrow support are myelosuppression-associated infections, thrombocytopenia, and therapy-related organ system toxicity. 446 Use of hematopoietic growth factors (e.g., rhGM-CSF, rhG-CSF, rh!L-3) . has shortened the period of absolute granulocytopenia to a few days. With rhGM-CSF, the number of platelet transfusions required by the patient has been drastically reduced (see Fig. 40-16). Growth factor-supported high-dose sequential adju-

J 102 of 122 Celltrion, Inc., Exhibit 1006

vant chemotherapy is becoming more tolerable anc! practically devoid of iatrogenic mortality. 446

-448

The major late toxic effect of adjuvant chemotherapy is the risk of secondary neoplasms, mainly acute nonlymphocytic leukemia (Table 40-28). There may be a slightly increased risk of myeloproliferative disease in patients treated with melphalan-containing regimens (1.3%), but induced ~cute leukemias after adjuvant CMF have been reported only occasionally. 482•483 There is no apparent increase in the incidence of solid tumors, although longer follow-up is reqi.iired to assess this possibility. 483•

484

The major toxicities of adjuvant tamoxifen are menopausallike symptoms, atrophic vaginitis, occasional uterine bleeding, anc! thrombophlebitis. The estrogenic effect of tamoxifen may decrease circulating levels of antithrombin III, predisposing to thromboembolic disorders.457•

485

There are insufficient data on the long-term consequences of prolonged tamoxifen use, but there is evidence that tamoxifen does not have deleterious effect on bone density after prolonged administration, and this drug produces a qeneficial blood lipid profile in women. It has been reported that in athymic ntice transplanted with both hormone-dependent breast and endometrial tumors and treated with tamoxifen, the estrogen-stimulated growth of the breast tumor is controlled, but the growth of ·the endometrial tumor is not. 486

Physicians should be aware of the possibility that an occult endometrial carcinoma may not be controlled during prolonged adjuvant tamoxifen therapy for breast cancer.487

·488

The cumulative frequency of infiltrating endometrial cancer in women receiving tamoxifen is 0.5%, compared with 0.1 % in the control group. 489 When the Stockholm trial (in which higher doses of tamoxifen were used) is excluded, the differences in the frequency of endometrial cancer ~sing tamoxifen doses of 20 mg daily are not dramatic. This twofold increase in ris~ of endometrial cancer is similar in magnitude to

Systemic Adjuvant Therapy· 1313

that associated with postmenopausal estrogen-replacement therapy.

PRACTICAL GUIDELINES FOR ADJUV ANT THERAPY

Despite 20 years of progress in the development of adjuvant therapy, many issues remain unsettled. Physicians are still encouraged to enter patients into prospective studies. Treatment guidelines Provide recommendations for patients who are treated outside the context of clinical trials. The Consensus Development Conferences held in Bethesda in 1985 and 1990 provided guidelines for the use of adjuvant chemotherapy and tamoxifen. 206

•490 During the past few years, additional infor

mation has become available from recent studies, particularly from the Overview report and studies of node-negative breast cancer.

Node-Positive Tumors

The available data indicate that patients with node-positive tumors benefit from adjuvant treatment. Regardless of hormone receptor status, an established combination chemoth~rapy (e.g., full-dose CMF for six monthly cycles) is now the standard of care for a premenopausal patient. For a postmeriopausal patient, the hormone receptor status det((rmines the approach. In patients with receptor-positive tun\ors, tamoxifep. is given for a minimum of 3 years, and in patients with receptor-negative tumors, chemotherapy (e.g., CMF) is given. If there are other unfavorable prognostic indicators (e.g., undifferentiated tumors, aneuploidy, high tumor cell proliferative activity), it is reasonable to use chemotherapy in addition to tamoxifen in receptor-positive patients. In patients presenting with more than three positive nodes (especially thqse having '2o!O nodes) full-dose doxorubicin, given every 3 weeks for four cycles, followed by intravenous CMF for four

TABLE 40-28. Frequency of ~econd Primary Malignancies After Adjuvant Systemic Treatment

Median Total No. Follow-up Second Postsurgical

Investigations of Patients (y) Malignancies Treatment

Milan2sa,4a3 845 14 Acute leukemia Control CMF

Solid tumors* Control CMF

NSABP482 8013 10 Acute leukemia or Control Myeloproliferative Radiation therapy dise~se L-PAM ±other drugs

Danish487 1710 9 Endometrial ca. Radiation therapy Radiation therapy plus

TAM 30 mg/day X I y Stockholm488 1846 4.5 Endometrial ca. Control

TAM 40 mg/day x 2 y TAM 40 mg/day X 5 y

* Contralateral breast cancer and cutaneous basal cell carcinomas excluded; L-PAM, L-phenylalanine mustard (melphalan); TAM, tamoxifen.

Actuarial flisk (%)

0 0 8.1 4.7 0.1 1.4 1.3

0.27 I.OJ

<0.5 1.0 5.5



monthly cycles should be considered, regardless of menopausal and receptor status. In this high-risk subset, tarnoxifen can be administered after chemotherapy is completed if the tumor is receptor positive.

Node-Negative Tumors

The available data do not clearly indicate which patients in this subset should be routinely treated with adjuvant therapy. There is consistent evidence that, regardless of menopausal status, receptor-negative tumors benefit from adjuvant chemotherapy (e.g., full-dose CMF for six monthly cycles). Other high-risk subsets, including patients with tumors larger than 3 cin, undifferentiated tumors, and tumors with a high proliferative activity, should be treated with adjuvant therapy. Conversely, patients with favorable tumors (e.g., size <1 cm, well differentiated, low proliferative activity) do not routinely require adjuvant therapy, particularly if all three favorable indicators are present. In women who have tumors with positive hormone rec~ptors, there is evidence that adjuvant tamoxifen prolongs relapse-free and overall survival, although the fraction of patients who benefit from this systemic treatment remains to be defined.

PRIMARY CHEMOTHERAPY

Primary (i.e., neoadjuvant) chemotherapy is a somewhat novel treatment approach in women presenting with resectable tumors. This strategy is based in part on the experience achieved in the treatment of stage Ill breast cancer. Two clinical developments have helped to bring about the use of chemotherapy as initial treatment in earlier disease stages: the similarity of treatment outcome between breast-conserving procedures and radical mastectomy in small tumors and the demonstrated efficacy and safety of adjuvant postoperative chemotherapy. 191 '491

In an attempt to reduce the frequency of mutilating surgery for tumors larger than 3 cm, the Milan Cancer Institute in 1988 initiated a prospective study using neoadjuvant chemotherapy in patients with operable breast cancers larger than 3 cm. ' 0

'-478

·492 It was hoped that the use of chemotherapy be

fore surgery would improve long-term results compared with the classic strategy of surgery followed by chemotherapy. The updated results of the Milan Cancer Institute for 227 evaluable

patients presentµig with operable breast cancer 3 cm or larger are summarized in Table 40-29. The degree of tumor response was :inversely proportional to the initial tumor size, and pathologic complete remission was documented in 8 ( 4%) of 220 patients subjected to surgery. Conservative surgery (i.e., quadrantectomy or tumorectomy with ample free margins plus full axillary dissection) was performed in a total of 201 (91 %) of 220 patients.

There was additional important information from this prospective study. Tumor response was similar for the four drug combinations (CMF, FAC, FEC with epirubicin substituting for doxorubicin, FNC with mitoxantrone substituting for anthracycline), and there was no difference in response rates for three and four cycles of CMF or FAC. Three cycles of single-agent doxorubicin (75 mg/m2 every 3 weeks) yielded an objective tumor response (79%) comparable to that of combination chemotherapy (78%). Age, menopausal status, tumor proliferative activity, and DNA ploidy did not influence the degree of tumor reduction, but the frequency of response was greater in receptor-negative tumors. With a median follow-up of 18 months after surgery, local recurrences were detected in fewer than 2% of the 201 patients treated with conservative surgery and postoperative breast irradiation. In this protocol, high-energy radiation therapy was started 4 to 6 weeks after surgery, the breast was irradiated wi.tlJ two opposing tangential fields, and treatment consisted o~ five fractions each week to a total dose of 6000 cGy in 6 weeks.There is insufficient follow-up to evaluate relapse-free and total survival rates in this trial.

A French study initiated by J acquillat and others attempted to avoid surgery altogether by the use of primary chemotherapy followed by irradiation and further systemic therapy. 287

For 192 patients with stage I through lllA breast cancers, they reported a 5-year relapse-free survival of 100% for stage I, 82% for stage IIA, 61 % for stage IIB, and 46% for stage lllA. The overall rate of breast preservation at 5 years was 94%. They concluded that most women with breast cancer should be given the option of breast-pre::;erving treatm·ent. A more recent French study included 272 patients with operable breast cancers measuring greater than 3 cm, who were prospectively randomized to primary chemotherapy followed by locoregional treatment or to mastectomy with postoperative adjuvant chemotherapy given to high-risk patients (i.e., nodepositive and node-negative ER-negative tumors). 493 Conservative treatment was feasible for 63o/o of women.treated with

TABLE 40-29. Operable Breast Cancer Candidates for Mastectomy Treated With Primary Chemotherapy (Milan)

Initial T Patients Subjected Tat Surgery (cm)* Conservative (cm) to Surgery <3 0 pCRt Surgery

3.0-4.0 110 102 II 5 107 4.1-5.0 73 58 10 3 67

>5 37 23 5 27

•Initial clinical size compared with size at surgery after three or four cycles of primary chemotherapy. t pCR, pathologic complete remiss!on. ·

'•"T


primary chemotherapy. At a median follow-up of 34 months, overall survival, but not relapse-free survival, favored the group of patients treated with primary chemotherapy. This interesting study requires longer follow-up to draw meaningful conclusions, particularly about the adequacy of locoregional treatment and patterns of disease recurrence.

Other research groups are conducting randomized clinical trials comparing primary chemotherapy and postoperative chemotherapy. For example, the NSABP is evaluating the use of four cycles of doxorubicin and cyclophosphamide (AC regimen) given before or after local regional therapy. Additional studies are required to determine the optimal integration of primary and adjuvant chemotherapy. What can firmly be stated today is that the delivery of short-term full-dose primary chemotherapy in large but resectable breast cancers can result in downstaging of the primary tumor to less than 3 cm in diameter in most patients, ·allowing breast-conserving treatment. These results challenge tlie classic indication for primary mastectomy and offer the possibility of an effective and safe alternative for women concerned about preservation of body integrity. Currently in the United States, the use of primary chemotherapy in operable breast cancer should be restricted to patients entered onto clinical trials.

FOLLOW-UP EXAMINATIONS AFTER PRIMARY THERAPY

The follow-up of patients after primary therapy should ideally be carried out in a simple, regular, and uncostly manner. An important objective is the early detection of persistent or recurrent disease in the ipsilateral breast and new cancers in the contralateral breast or in other high-risk organs, such as the uterus, large intestine, and lung. For patients treated using breast-conserving treatment, periodic mammography of both breasts is recommended. The optimal periodicity of performing mammography of the treated breast has not been established, but annual mammography would be considered a minimum. Some clinicians favor performing mammography of the. treated breast every 6 months for the first 5 years, especially in patients whose cancer was detected only on mammography.

The benefit of early detection of distant metastases has not been established, and the periodicity of screening tests for metastases (e.g. 1 bone scans, chest radiographs, serum markers, liver function tests) should be individualized, based on the needs of the patient. One reasonable program is to obtain a chest radiograph and radionuclide bone scan every 6 months during the first 3 years.and then annually. The routine use of circulating tumor markers deserves a word of caution. 494 CEA and CA 15-3 are elevated at presentation in approximately 20% of patients with stage I or II breast cancer. CA 15-3 appears to be a more sensitive marker than CEA. If the marker(s) persist(s) at the same level after surgery, this should not be considered tantamount to signaling micrometastatic disease. Only if the abnormal values progressively increase over time is there a high likelihood of metastatic disease. Isolated abnormal laboratory findings usually create considerable anxiety for the patient and the physician, but they should not be used as a basis to alter therapy.

Treatment of Overt Metastatic Breast Cancer

TREATMENT OF OVERT METASTATIC BREAST CANCER

1315

The treatment of clinically overt metastatic breast cancer is currently in a static phase. Despite innumerable trials using the available cytotoxic drugs, hormonal agents, biologic response modifiers, and various combinations of these three modalities, the survival of patients with metastatic disease has not consistently and substantially improved over that of the previous decade. Overall, the frequency of durable complete remissions has not dramatically changed, even if highdose chemotherapy is combined with autologous bone marrow reinfusion. With the exception of taxol, there are few promising new drugs or innovative strategies on the immediate horizon. The clinical situation is somewhat complicated by the fact that many patients with widespread disease have relapsed after prior adjuvant chemotherapy, and their tumors may be resistant to most effective drugs. Current medical treatment in this large patient subset is palliative, and recurrent breast cancer remains incurable. However, the judicious application of available treatment can control the disease for several months or years in some Patients.

In treating patients with metastatic disease with hormonal or chemotherapy, it is important for the clinician torJocument carefully the patient's response to treatment. The rules for assessing a response have been established by the UICC and are shown in Table 40-30.

ENDOCRINE THERAPY

Manipulation of the endocrine system is the oldest form of systemic therapy for breast cancer. Overall, approximately one third of unselected patients with metastatic disease respond to endocrine therapies (Table 40-31). The discovery of steroid receptors contributed considerably to the selection of patients for endocrine therapy. In advanced disease, the objective response rate is 50% to 60% in patients with ERpositive tumors and 65% to 75% if the tumor is also PGR positive. 495 Only 5% to 10% of patients with receptor-negative tumors respond to hormonal manipulation.

Receptor levels measured in the primary tumor, in the ab- . sence of intervening hormonal therapy, are remarkably similar to levels measured in recurrent disease. In a multivariate analysis of ER, PGR, age, site, and number of metastases, the quantitation of the ER level was shown to be the most important predictor of hormone response to tamoxifen, with response rates increasing from approximately 20% for women with ER values of 3 to 10 fmol/mg of protein to approximately 75% response for women with tumors having ER values greater than 30 fmol/mg of protein. Patients with PR levels greater than 10 fmol/mg of protein have a significantly higher response rate (55%) than those with values less than 10 fmol/ mg of protein (25%). In general, neither the level of ER nor the presence of PR correlates with the duration of response. 496

Ablative Procedures

Oophorectomy commonly induces an almost immediate subjective response in menstruating wonien .whose tumors are ER-positive. The median duration of response is longer than 8 months, and 20% to 30% of patients achieve complete re-



TABLE 40-30. Assessment of Response to Medical Therapy in Clinically Advanced Breast Cancer: Recommendations of UICC, 1979

A. Recording of lesions This should be done by anatomic site. Dimensions should be stated when applicable and the method of evaluation stated (e.g., direct measurement, photOgraph, radiograph). Pleural effusion, ascites, hepatomegaly, pulmonary shadows, etc., may be the result of nonmalignant processes. When possible, histologic proof of involvement should be obtained if these abnormalities are to be used for evaluation of response or stratification in a clinical trial 1. Soft tissue. Breast: ipsilateral or contralateral.

Skin: intracutaneous or subcutaneous or both. Lymphedema.

2. Bone. Sites recorded (state whether lytic and/or blastic). 3. Visceral. Lung (nodular or diffuse), pleura (nodular and/or malignant effusion), ascites, abdominal or pelvic masses, central nervous

system. B. Menopausal Status

There are three physiologic categories of menopausal status which may be classified in various ways. The following is suggested: 1. Premenopausal-a menstrual period has occurred within the previous year. 2. Early postmenopausal-last period: 1-5 years. 3. Late postmenopausal-last period: >5 years. Women who have had a hysterectomy with one or both ovaries left in place may be considered premenopausal if <50 years of age and postmenopausal if >55 years of age: those 50-55 years are classified as early postmenopausal. Vaginal cytology or hormone studies may clarify the true menopausal status. Younger women who have had an artificial menopause should be considered separately.

C. Disease-Free Interval This is the time from treatment of the primary tumor by surgery or radiation therapy or adjuvant systemic therapy to the time of the first recurrence. 1. No free interval (e.g., stage IV or Ml) ) 2. <2 years ( 3. >2 years

D. Definition of Response 1. Measurable lesions. Ideally, all lesions should be measured at each assessment. When multiple lesions are present, this may not be

possible and, under such circumstances, a representative number of eight or more lesions may be selected for measurement. For bidimensional lesions, regression is defined as whet) all lesions disappear, or the sum of the products of the diameters of each individual lesion, or those selected for study, decreases by 50% or more, with no lesion increasing in size. In each case, no new lesions should appear. Progression is defined as when (i) new lesions appear, (ii) there is a 25o/o or more increase in the sum of the products of the diameters of each lesion measured. If an increase of less than 25% makes additional treatment necessary, this is also regarded as progression.

2. For unidimensional lesions, in the case of regression, the same rules apply as for bidimensional lesions, except that regression is taken as a decrease of 50% or more in one measurement. In situations such as infiltration of the breast, liver involvement, and mediastinal enlargement, objective regression is a 50% or greater decrease in that measurement which is regarded as being in excess of that usual for the site under consideration.

3. Evaluable, but nonmeasurable lesions (e.g., osseous metastases, pulmonary infiltration, pleural effusion, skin infiltration). Serial evidence of appreciable change documented by radiography or photography must be obtained and be available for subsequent review. The assessment must always be objective. Pathologic fractures or collapse of bones are not necessarily evidence of progressive disease. Neither the development nor healing of skin ulcers should be taken as sole evidence of change.

E. Categories of Response Objective regression applies to 1 and 2. 1. Complete response: disappearance of all known disease. In case of lytic bone metastases, these must be shown radiologically to have

calcified. 2. Partial response: ~50% decrease in measurable lesions and objective improvement in evaluable, but nonmeasurable lesions. No new

lesions. It is not necessary for every lesion to have regressed to qualify for partial response, but no lesion should have progressed. 3. No change: lesions unchanged (i.e., <50% decrease or <25% increase in the size of measurable lesions).

Note: if nonmeasurable but evaluable lesions represent the bulk of disease and these clearly do not respond, even though measurable lesions have improved, this is considered as "no change," not "objective regression."

4. Progressive disease: Mixed-some lesions regress while others progress or new lesions appear. Failure-progression of some or all lesions or appearance of new lesions. No lesions regress.

F. Duration of Response In a patient who has an objective regression, this is to be dated from the start of therapy until new lesions appear or any one existing · lesion increases by 25% or more above its smallest size recorded. It is essential to categorize a patient as having a regression at a stated time. It is also essential that all baseline studies should have been repeated at this time.

G. Survival Survival is dated from time of commencement of treatment to death.

(Adapted from Hayward JL, Carbone PP, Heuson JC. Assessment of response to therapy in advanced breast cancer. A project of the Programme on Clinical Oncology of the International Union against Cancer, Geneva, Switzerland. Eur J Cancer 1977;13:89-94)


Treatment of Overt Metastatic Breast Cancer 1317

TABLE 40-31. Endocrine Therapies in Clinically Metastatic Breast Cancer: Response Rates in Unselected Patients

No. of Response Range Therapy Patients Rate(%) (%) Effective in

Tamoxifen 1269 32 16-52 Pre- and postmenopause Oophorectomy 3380 33 21-41 Premenopause or within 1 year

from last menstrual period Progesiins 3479 31 9-67 Pre- and postmenopause Aminoglutethimide 1153 32 16-43 Postmenopause LHRH analogs . 293 40 32-45 Premenopause Estrogens 1683 26 15-38 Postmenopause Androgens 2250 21 10-38 Pre- and postmenopause Adrenalectomy 3739. 32 23-46 Pre- and postmenopause Hypophysectomy 1174 36 22-58 Pre- and postmenopause

(Harris JR, Hellman S, Henderson IC, Kinne DW. Breast diseases. 2nd edition. Philadelphia: JB Lippincott, 1991)

missions that can last for y~ars for some women. Although the frequency and duration of objective response is about the same after radiation castration (2000 cGy). the benefit from radiation therapy becomes evident only after the second to the tenth week. In the absence of steroid receptor values, ovariari ablation or other endocrine manipulations (see Table 40-31) are indicated in premenopausal women with slowgrowing metastatic disease, a long disease-free interval, and age over 35 years. The response rate to oophorectomY. iii premenopausal women younger than 35 is lower than the response rate in older premenopausal women. In one study, an objective response to castration occurred in about 200/o of amenorrheic women who relapsed after CMF adjuvant chemotherapy. 452 This observation supports the concept that the benefit from chemotherapy is not necessarily related to ovarian suppression and that drug-induced amenorrhea is not tantamount to castration. There are no data indicating that maintenance therapy with tamoxifen can improve the treatment outcome of patients who respond to ovarian ablation.

Other endocrine ablative procedures, such as adrenalectomy and hypophysectomy, are considered obsolete forms of treatment for patients with metastatic breast cancer because of the availability of newer endocrine treatments, such as antiestrogens, progestins, aminoglutethimide, and luteinizing hormone-releasing hormone analogs.

Antiestrogens

During the past 20 years, the nonsteroidal antiestrogen tamoxifen has replaced diethylstilbestrol and androgens and has become the most widely used hormonal therapy for patients with metastatic breast cancer. Tamoxifen appears to• exert its main antiproliferative effect by competing with estrogen for binding to ER proteins. Presumably, the drug-ER complex inhibits gene transcription and protein synthesis of factors important to tumor growth. The degree of tumor-cell inhibition correlates with the affinity of tamoxifen and its active metabolites for the ER. After binding to the ER. tamoxifen antagonizes many of the cellular events affected by estro-

gen. 457•497

-499 Other cellular mechanisms of tamoxifen may

augment tumor inhibition. The precise subcellular actions for tumor inhibition by tamoxifen remain uncertain. In model systems .• this compound slows estrogen-induced growth and leads to a cell-cycle blockade. Tumor cells are prevented from entering late phases of the cell cycle and accumulate in the early G1 phase. In vitro and in vivo data suggest that the predominant effect of tamoxifen iS'cytostatic.500 At low concentrations, the inhibitory effects of tamoxifen, mediated through ER, can be completely reversed by the addition of estradiol, but at. very high concentration, the inhibitory effects are not reversible and are probably not mediated through the ER. Although tamoxifen behaves primarily as an estrogen antagonist, it may act as a partial agonist for some organs; in fact, the drug shows weak estrogenic properties. 499

Acquired tamoxifen resistance represents a major cause of treatment failure in all stages of breast cancer. It has been reported that tamoxifen resistance in MCF-7 human breast cancer cell lines is associated with changes in tamoxifen metabolism and the accumulation and retention of the drug by the breast cancer cells. 501

Tamoxifen has complex endocrine effects that depend on treatment duration and dose, menopausal status, and the target organ. In postmenopausal women, the normally elevated gonadotropins, follicle-stimulating hormone (FSH) and Juteinizing hormone (LH), decrease with tamoxifen therapy (although levels remain in the normal range), but serum estradiol and progesterone levels are unaffected. In premenopausal women, estradiol and progesterone levels show a striking elevation of up to three times normal, and FSH and LH levels remain unchanged or only slightly increased from pretherapy levels. The antiestrogenic properties oftamoxifen usually are not sufficient to suppress ovarian function. Recent data from the University of Wisconsin have demonstrated that long-term adjuvant tamoxifen for stage I or II breast cancer continues to result in an increase in steroidogenesis in premenopausal women. 502 Because ovulation continues, there is a risk for pregnancy. and patients should be advised to use contraception. Approximately half of patients on Jong-term tamoxifen



therapy continue to have regular ovulatory menstrual cycles. In women developing oligomenorrhea or amenorrhea while on treatment, menses return to normal in more than 90o/o.

The standard therapeutic dose is 20 mg/ day oraily and treatment is continued until there is documented evidence of disease progression. The use of doses higher than 20 mg/day is associated with a greater increase in menstrual irregularities but not an increase in the objective response rate. Tamoxifen is currently recommended for postmenopausal patients with metastatic disease and ER-positive tumors. The frequency of

·response is related to the level of the ER protein. Tamoxifen is an active agent for premenopausal patients with metastatic disease, although its value is less well documented. The response rates appear to be comparable to those reported for ovarian ablation, although there are no firm data indicating that they are equivalent.499

·503

•504 This drug can be offered as

a reasonable alternative-to women who wish to avoid surgical or radiation castration. It is not yet clear whether the response to tamoxifen can predict response to subsequent ovarian ablation. 499

The toxicity from tamoxifen is minimal and transient. 505

The frequency of nausea and vomiting. weight gain, vaginal bleeding, skin rash, edema, leukopenia, thrombophlebitis, and abnormal liver function tests is less than 3o/o. Transient ocular disturbances of the lens, retina, and optic nerve are rare (<I%). 506 Thrombocytopenia, flare, and hypercalcemia may occur in about 5% of patients and hot flashes in 7% to 10%. The main side effects from prolonged tamoxifen therapy were described in the section on adjuvant endocrine therapy.

Other antiestrogens have been studied to identify compounds with less estrogenic effect or no cross-resistance to tamoxifen. Trioxifene, droloxifene, zindoxifene, toremifene, and !CI 164384 have been studied, but they have not been proven to be more active or less toxic than tamoxifen. The results of randomized multicenter trials comparing toremifene and tamoxifen are not yet available, and toremifene therefore remains experimental. The antiestrogens clomiphene and nafoxidine have no role in the current treatment of breast cancer.

Progestins

The semisynthetic progestins medroxyprogesterone acetate and megestrol acetate are the two most active compounds of this class of hormones available for the treatment of metastatic breast cancer. 507 The other effective compounds, norethistrone and hydroxyprogesterone caproate, are rarely used in the current treatment of breast cancer. The mechanism of action of the progestins is not well understood. Several potential mechanisms have been suggested, and it is probable that the predominant effects vary with drug dosage. 508 Studies with the MCF-7 hormone-sensitive cell line suggest that at low concentrations megestrol acetate causes direct cytotoxicity mediated by the hormone receptors. Other investigators have shown that there is dose-dependent inhibition of the hypothalamic-pituitary-adrenal axis that results in suppression of adrenal steroid production. A third potential mechanism for this class of hormones is through their effects on the autocrine growth factors and their receptors. The presence and levels of ER predict response to the progestins, but there is

no evidence that patients whose tumors contain. high PGR levels are more likely to respond than those with low levels.

Medroxyprogesterone acetate has been evaluated, mostly in Europe, in a variety of doses and schedules. 509

- 511 At the end of several trials, it was concluded that very high doses (>1000 mg/day) were not required to influence the response rate, duration of response, or survival. The most frequently used dose schedule is 1000 mg/day given orally or intramuscularly for the first month, followed by 500 mg/day once or twice each week. The drug is almost exclusively administered in postmenopausal patients, for whom the overall response rate is 33% (range, 10-67%). The objective response is higher in women who have not received prior endocrine therapy ( 40-45%) compared with those who have been treated with tamoxifen (5-10%). In patients presenting with considerable liver involvement, the tumor response is practically nil.

Megestrol acetate has been widely evaluated, primarily in the United States, and has shown an overall response rate of 28% (range, 14-56%).612- 514 The therapeutic dosage commonly used is 160 mg/day in divided oral doses. It remains to be confirmed through the ongoing randomized studies conducteCI by Cancer and Leukemia Group B that higher doses (800-1600 mg/day) yield superior objective response and survival compared with lower doses. Randomized trials have confirmed that, for patients without prior endocrine therapy, megest·rol acetate is therapeutically equivalent to tamoxifen and that the response rate is low (<I 0%) if megestrol acetate is used after two or three other hormon~treatments.515

Prolonged high doses (;,,1000 mg/day) of medroxyprogesterone acetate are associated with an increased incidence of side effects, including gluteal abscesses after intramuscular administration (15%), facies lunaris (10%), increased sweating (6%), fine tremors and leg cramps (20%), weight gain (>50%), fluid retention (,;JO%), hypertension (,;10%), worsening of diabetes mellitus, hypertrichosis ( ,;50%), skin rash, and hypercalcemia. Unlike the experience with medroxyprogesterone acetate, toxicity at the highest dose of megestrol acetate is milder and transient and includes fluid retention, hypertension, hyperglycemia, and mild congestive heart failure. 514 Because of superior oral bioavailability and ease of administration, megestrol ·acetate is becoming. the progestin of choice in the United States. The considerable increase in appetite and body weight associated with the use of medroxyprogesterone acetate and megestrol acetate at high doses is being used in the management of c'achexia.

Aminoglutethimide and Aromatase Inhibitors

Aminoglutethimide was developed to effect a medical adrenalectomy and function as an antiestrogen. The major source of estrogen in postmenopausai' women is from the adrenal gland, which produces androstenedione. Androstenedione is converted by an aromatase reaction in peripheral tissues to estrone and estradiol. Aminoglutethimide blocks several steroid hydroxylations and cleavage enzymes. It is capable of inhibiting the conversion of cholesterol to o-5-pregnenolone, blocking all adrenal steroid synthesis. The drug can inhibit the aromatase reaction responsible for conversion of androstenedione to estrone in peripheral tissue.

Adrenal suppression with aminoglutethimide results in

--108 of 122 Celltrion, Inc., Exhibit 1006

negative feedback to the pituitary, increased levels of ACTH, and further stimulation of the adrenal gland to overcome the block. Patients treated with aminoglutethimide are usually given glucocorticoid replacement to prevent addisonian symptoms. 516 When aminoglutethimide is given orally at the dose of 1000 mg/day, the recommended glucocorticoid is hydrocortisone (20 mg twice daily or in a more physiologic schedule of 10 mg in the morning, 10 mg in the afternoon, and I 0 mg before bed time). Clinical trials demonstrated that low doses of aminoglutethimide (250-500 mg/day) can suppress circulating estrogens to the same level as full doses without compromising treatment efficacy.517- 519 The standard dose in clinical practice is 250 mg twice daily combined with 40 mg/day of hydrocortisone, although the deletion of glucocorticoid replacement has been shown to be safe and effective.

In patients with previously untreated metastatic breast cancer, aminoglutethimide has replaced surgical adrenalectomy and has been shown to be as effective as any other form of endocrine therapy in postmenopausal patients. 520 The overall response rate is about 35% (range, 25-53%). When used as second-line endocrine treatment, the overall response rate is approximately 25%. Compared with tamoxifen, it appears that aminoglutethimide is more likely to induce objective response in sites of bone metastases.

Most of the side effects induced by aminoglutethimide can be documented during the first 6 weeks of treatment and include lethargy (35-45%), skin rash (s30%), orthostatic dizziness (20%), ataxia (.10%), and mild hypertension (5%). In fewer than I% of patients, thrombocytopenia or leukopenia may be observed, and in a few cases, fatal pancytopenia has occurred. To avoid or diminish lethargy and pruritic skin rash, it is advisable to start treatment for a few days with hydrocortisone only. If pruritus develops, aminoglutethimide should not be discontinued, because this symptom can be easily alleviated by doubling t)le dose of hydrocortisone for about 1 week.

The development of new aromatase inhibitors and adrenal suppressants is an active area of drug development, with several promising agents, including 4-hydroxyandrostenedione and CGS J 6949A, on the horizon. 521 - 524 Both drugs are effective in postmenopausal patients, with response rates of approximately 30%. Neither has been shown to be superior to aminoglutethimide, but several clinical studies are ongoing, and these drugs remain experimental.

Trilostane is not an aromatase inhibitor, but it does inhibit the 3/l-hydroxysteroid dehy<lrogenase system and therefore the conversion of pregnenolone to progesterone, of 17 a

hydroxypregnenolone to 17 a-hydroxyprogesterone, and dehydroepiandrosterone to androstenedione. Trilostane is effective in about 25% of postmenopausal patients with metastatic breast cancer, and its administration must be supplemented with glucocorticoids. 525 Side effects include diarrhea (;,,30%), anorexia, lethargy, dizziness, and headache. This compound is experimental.

Inhibitors of Pituitary Function

Based on the observation that interruption of pituitary function by hypophysectomy has a beneficial effect in patients with


metastatic breast cancer, medical inhibitors of pituitary function are being evaluated. Few objective responses have been. documented after the use of prolactin inhibitors, such as levodopa or bromocriptine. Danazol, a syntl;ietic steroid whose primary effect is to inhibit pituitary function, has been shown in small trials to have some efficacy in postmenopausal women with metastatic breast cancer. 525 All of these drugs appear inferior to the well-established endocrine manipulations, such as castration, tamoxifen, progestins, and aminoglutethimide.

Luteinizing Hormone-Releasing Hormone Analogs

Luteinizing hormone-releasing hormone (LHRH) agonists or antagonists are promising new areas of research and treatment for breast cancer. The agonist drugs, including buserelin, leuprolide (Leupron), and goserelin (Zoladex) decrease FSH and LH excretion, decrease prolactin excretion, and decrease the levels of sex hormones. The mechanism of action usually ascribed to this group of compounds is a medical oophorectomy. 526 However, it is possible that these agents have a direct effect on breast cancer.

Several phase II studies have provided convincing evidence that LHRH agonists are effective in the treatment of premenopausal women with metastatic disease and have shown a total response rate of 40% (range, 32-50%). Some postmenopausal women may respond to these agents. 527

-529 Gos

erelin is easy to administer through a depot preparation injected monthly as a single subcutaneous injection of 3.6 mg. Toxicity is minimal and consists m~ly of hot flashes. Cessation of menses persists only as long as the patient continues taking the drug. Prospective randomized trials ongoing in Europe and North America are comparing LHRH agonists (with or without tamoxifen) with oophorectomy in premenopausal women who have metastatic diseas.e and with adjuvant chemotherapy.

COMBINED ENDOCRINE THERAPY

Theoretically, combining two or three endocrine therapies may provide therapeutic advantages, because different endocrine therapies can inhibit tumor growth through different mechanisms. However, it is possible that combinations of hormones may produce an antagonistic interaction. To investigate this, several trials were activated during the past decade.530 The most frequently tested combinations have been tamoxifen with fluoxymesterone, glucocorticosteroids, medroxyprogesterone acetate, or goserelin, and the results are intriguing. The response rate after polyhormonal treatment often is sµperior to the use of a· single endocrine therapy. However, this superiority in the response rate has not been associated with a significant and reproducible improvement in the duration of response or survival. 531 This is probably due to the fact that patients who respond to a single endocrine therapy subsequently respond to other hormonal treatments, and women treated with combinations of hormones lose this sequential response. Patients treated with a combination of hormonal therapies experience the combined toxicities as well. Combinations of hormones are not recommended outside of a clinical trial.



GUIDELINES FOR ENDOCRINE THERAPY

Considering that the probability of responding to an endocrine therapy in an unselected patient is only about 30%, hormonal therapy is only indicated for patieqts whose tumors are receptor-positive. The ER protein is a good predictor ofresponse to all forms of endocrine therapy and helps to exclude women who are not likely to benefit from hormonal treatment. The response rate to hormonal treatr:µent is related to the dominant site of disease and is highest in women with disease in soft tissue, lower in those with bone metastases~ and still lower in those with limited nodular pulmonary metastases. Physicians should remember that t4e :predictive value of a response to one endocrine therapy for a subsequent response to the second appears to be independent of the type of endocrine therapy used and that the level of ER protein is a more powerful predictor of response to endocrine therapy than prior response to therapy. The median duration ofresponse to a single course of endocrine therapy is between 1 ijild 2 years, and this appears to be independent of the type of hormonal therapy. For patients with negative-receptors, the response rate is consistently less than 10%, regardlesiof clinical presentation. Ifreceptors are not available and a biopsy is not feasible, medical oncologists should follow the established clinical guidelines outlined in Table 40-32.

Premenopausal Patients

For women older than 35 years of age and regularly menstruating, surgical castration is generally recommended as the first endocrine treatment. A good (not necessarily equivalent) alternative is the use of tamoxifen. GoE;~relin may prove to be an even better treatment than oophorectomy or tamoxifen, but it is still being evaluated.

If a patient who responded to oophorectomy relapses, tamqxifen is the treatment of choice. Second-line endocrine ~re~tment is associated with about a 50o/o response rate in patients who are ER-positive and who respond to a first course of endocrine therapy. Fewer than 25% of patients with receptor-positive tumors who f~il to respond to a first course of endocrine therapy respond to a subsequent endocrine therapy. !n patients who respond to tamoxifen, third-line treatment is a progestational agent (medroxyprogesterone acetate or me-

TA~LE 40-32. General Criteria to Select Patients for Endocrine Manipulations or Chemotherapy

Endocrine Therapy Chemotherapy

Slow-grow~ng disease particµlarly in soft tissue* or skeleton

Free interval > 2 rears Age > 35 years Objective response to prior

endocrine manipulations

* Inflammatory carcinoma excluded.

Rapidly growing disease, massive liver involvement, lung or skin involvement with lymphangitic metastases

Free interval < 2 years Any age group Negative response to

first endocrine manipulations

gestrol acetate) or aminoglutethimide. The latter drug is preferred in patients with disease predominantly in bone. If a patient wit4 a receptor-positive tumor progresSflS after surgical castration, combination chemotherapy is indicated.

Postmenopausal Patients

The treatment of choice is tamoxifen for pos.tmenopausal women, unless the patient has rapidly growing viscera! (e.g., liver, lymphangitic lung) involvement. In case of a favorable response, sec'ond- or third-line endocrine treatment. is ~dicated, and the sequence is usually the same as outlined for premenopausal patients. It is rare for a patient to respond to a fourth endocrine manipulation, and after the third hormonal treatment, patients are usually treated with combination chemotherapy. ·

CHEMOTHERAPY

General Principles

In the treatment of clinically disseminated breast cancer, the aim of conventional chemotherapy is to palliate symptoms and improve the quality of life. Because·the selection of patients for chemotherapy is largely a process of exclusion, clinicians should carefully evaluate the general indications outlined in Table 40-32 before deciding to start chemotherapy.532

Because ti)e adlijinistration of chemotherapy always involves the principle of dose intensity, physici;J'S should carefully evaluate patients to rule out certain organ dysfunctions (e.g., high blood urea nitrogen for methotrexate, high serum bilirubin or cardiac insufficiency for doxorubicin, poor bone marrow reserve due to prior extensive irradiation of the spine and pelvis) that may preclude delivering full doses of drugs. 440

Age is not a substantial impediment to the use of chemotherapy, unless increased age is associated with one or more organ dysfunctions.

During treatment, measurable lesions should be objectively evaluated to avoid prolongation of ineffective but toxic treatment. The criteria of response for endocrine therapy and che-motherapy are shown in Table 40-30. ·

For most patients, chemotherapy is delivered using a polydrug regimen, because combinations of drugs are superior to the use of a single agent. The one exception is the use of single-agent doxorubictn. which is as effective as several standard combination regimens.533·534 As in the adjuvant setting, treatment should be applied according to the principles of dose intensity whenever safety allows. If the patient has moderate leukopenia ( <3000 peripheral leukocytes/mm3), it is usually preferable to delay the subsequent treatment for.a few days rather than reduce the drug dose by 50%.

In considering the indications for systemic therapy, medical oncologists should not neglect the contribution oflocal treatment in certain clinical situations. The most common events are listed in Table 40-33. For example, a patient with a single chest wall lesion appearing 4 or 5 years after mastectomy can easily and effectively be treated with excision and irradiation, if restaging excludes distant metastases. A similar strategy can be applied in a patient with a long disease-free interval (>2 years) and a small supraclavicular recurrence or an isolated bony metastasis. The decision to add endocrine therapy,


TABLE 40-33. Clinical Situations in Which Surgery or Radiation Therapy are used in Addition to or Instead of Systemic Therapy

Limited recurrence (e.g., chest wall, supraclavicular adenopathy, a single vertebra) after a free interval> 2 years

Impending fracture in a long bone Persistent localized bone pain Spinal cord compression Brain or choroidal metastases Pleural or pericardia! effusion

such as tamoxifen, to local treatment depends on the patient's hormone receptor status, but the use of chemotherapy usually can be delayed because survival is not improved by earlier use. Patients with a single site of recurrence and a short disease-free interval should preferentially be treated with systemic therapy, uSing the isolated lesion as a means of measuring response.

Combination Regimens

The most common drug regimens currently used in the treatment of advanced breast cancer are shown in Table 40-22. · Doxorubicin-containing regimens yield a 10% to 20% higher response rate than polydrug regimens without doxorubicin. However, it has not been demonstrated that the use of doxorubicin or its analog epirubicin can improve overall.survival in this setting. The same applies to CMFP when compared with CMF, for which the magnitude of the advantage provided by prednisone is minimal. The role of vincristine remains undefined, because there is no difference iri the response rate between CAFVP and CAF. 532'535'536

Table 40-34 presents the average therapeutic results that can be expected after adequate delivery of an effective combination regimen. The reported differences reflect differences in patient selection, response criteria, freq~ency of assessment, and total tumor volume. The response rate is higher among patients who present with disease predominantly in soft tissues than those with visceral and osseous metastases. The initial signs of response are usually measurable within the first 2 months after starting treatment, but in the case of .osteolytic metastases, it may take a few months to detect partial bone recalcification. Patients who do not show signs of tumor shrinkage within the first 6 months of chemotherapy usually do not show an objective response using additional treatment with the same drug regimen. Because it is difficult to determine the optimal second-line treatment for a patient who no longer responds to doxorubicin, we usually prefer CMF as initial chemotherapy, and use doxorubicin, alone or in combination, when the disease progresses. An additional reason for choosing this sequence is that CMF and doxorubicin are not cross-resistant and are roughly equivalent in efficacy. 537

The optimal duration of chemotherapy in patients with metastatic breast cancer is controversial. The results of recent trials indicate that the continuous use of chemotherapy results in a twofold to threefold longer median time to progression than a .six-course induction regimen followed by observation.538·539 However, total survival is essentially the same in the two treatment groups. The physician should carefully as-

Treatment of Overt Metastatic Breast Cancer

TABLE 40-34. Response After Combination Chemotherapy in Patients Previously Untreated With Cytotoxic Drugs

1321

Total with PR or CR (%) Total with CR (%) Time to initial response (median) Duration of response (median) Survival of responders (median)

45-800/o of all patients 5-25% of all patients 4-Bwk 5-13 mo

15-33 mo

PR, partial response; CR, complete response.

sess whether the patient is more concerned about the discontinuation of treatment without a ·complete remission or about the continuing symptoms of treatment, including nausea, vomiting, mucositis, and hair loss. In the few patients who achieve a complete remissiori, chemotherapy is discontinued and resumed after detection of tumor recurrence. Occasionally, patients who hav" complete responses may live in unmaintained remissions for as long as 5 years after the completion of their chemotherapy. 535

Other important strategies in the treatment of advanced breast cancer using drug combinations are the use of alternating noncross-resistant regimens, the sequencing of endocrine therapy and chemotherapy, and the use of combinations of chemotherapy arid endocrine therapy. The common rationale for all of these strategies is that bre~t tumors are characterized by multiple clones, and the patient may benefit from the use of multiple different treatments. It is possible that some clones are sensitive to chemotherapy and resistant to endocrine manipulations and that other clones manifest the opposite behavior. Several trials carried out over the past 15 years have shown some improvement in the response rate and the frequency of complete remission, but they have not demonstrated a consistent increase in the duration of response or in survival.2·540-542 These strategies have not been established as standard practice in the treatment of disseminated breast cancer.

Salvage Chemotherapy

Salvage chemotherapy is increasingly used for women who develop clinically disseminated breast cancer after adjuvant systemic therapy. Patients who fail after adjuvant tamoxifen can be treated with one of the standard drug combinations (see Table 40-22). There is no evidence that the response rate, tfle duration of response, or total survival are inferior compared with previously untreated patients with similar disease extent. If the disease-free interval from the completion of adjuvant chemotherapy is longer than 12 months, retreatment with the same drug regimen can yield a complete and partial remission rate of about 50o/o, with a median duration of response of 22 months. In women who had recurrences within the first 12 months after starting adjuvant CMF, retreatment with CMF was totally ineffective, and full-dose doxorubicin (75 mg/m2 every 3 weeks) yielded a 38% response rate, with a median response of 17 months.452 As in malignant lymphomas, the disease-free interval can differentiate patients who require noncross-resistant therapy and those who can be retreated with the same regimen.



Patients with metastatic breast cancer who are refractory to first-line chemotherapy generally do not respond to secondand third-line drug regimens. The most common situation is for a patient with metastatic disease to develop progressive disease after treatment with CMF. As in women who develop recurrences after adjuvant CMF chemotherapy, doxorubicin (or epirubicin) represents the treatment of choice, and it can induce a second response in 25% to 40% of patients. The use of full-dose doxorubicin alone appears to yield results superior to its administration with other drugs. If there is an objective response, treatment can be continued, but not to a cumulative dose exceeding 450 to 500 mg/m2 to avoid the risk of congestive heart failure. In patients with a poor.perf~rmance status or bone marrow reserve. massive liver involvement, or lymphangitic pulmonary metastases, weekly administration of 15 to 20 mg/m2 of doxorubicin is associated with fewer cardiac and noncardiac complications without a reduction in its effectiveness compared with a dose of 50 to 75 mg/m2

every 3 weeks.543

There is no third-line treatment of choice for patients who have become refractory to doxorubicin. Among the several regimens tested, a combination of mitomycin C and vinblastine is probably the most effective, yielding a response rate between 25% and 350/o.544- 547 Various dose regimens of this combination have been attempted to avoid excessive bone marrow toxicity. The regimen using mitomycin (10 mg/m2

)

on day 1 and vinblastine (5 mg/m2) on days 1and15, recycled

every 4 weeks, yielded a response rate of about 30% with less hematologic toxicity than regimens using mitomycin at the dose of 20 mg/m2 every 6 weeks.546 Single-agent treatments with nitrosourea derivatives (BCNU or CCNU), vinorelbine, platinum derivatives, or mitoxantrone are associated with lower response rates. 2

NEW DRUGS AND TREATMENTS

Taxol

Taxol is the single new drug that holds promise in the treatffient of breast cancer. Taxol is a complex molecule derived from the needles and bark of the Pacific yew tree, Taxus brevifolia. It functions as a mitotic spindle poison by interfering with the formation of the structural apparatus used by dividing cells to partition chromosomes between daughter cells. Although other plant-derived drugs (e.g., colchicine, podophyllotoxin, vinblastine, vincristine) interfere with cell division by inhibiting assembly of the microtubules that make up the spindle, taxol instead promotes assembly ofmicrotubules and

, stabilizes those already formed. This prototype ofa novel class of antimicrotubule agents that induces excessive polymerization of tubulin has demonstrated significant activity in ad- · vanced refractory ovarian epithelial neoplasms and objective responses in various malignancies including breast cancer. 548 ·

The administration of taxol is associated with several nonprohibitive side effects (e.g., neutropenia, neurotoxicity, myalgias, alopecia, vomiting, diarrhea, bradycardia, asymptomatic ventricular tachycardia). To avoid hypersensitivity reactions (probably due to the cremophor vehicle), the time span of intravenous infusion was prolonged, and routine premedication was implemented. Using a 6-hour intravenous infusion, the recommended phase II starting dose is 225 mg/

m2 every 3 weeks. 549 Sequences of taxol and cisplatin have yielded only mlld to modest neurotoxicity in 27% of patients. 548 Taxol appears to merit broad investigation at the phase II level. Because of the scarcity of taxoL research is being directed toward identifying active taxol analogs synthesized from the abundant but inactive taxol-like natural chemicals, taxanes.

Biochemical Modulation

Biochemical modulation refers to strategies that favorably alter the interaction of conventional therapeutic agents with their target end points in malignant and nonmalignant cells. The development of clinically useful modulatory strategies requires a solid knowledge of the mechanism of action of chemotherapeutic agents and the means by which malignant cells become insensitive to their cytotoxic effects. Several investigations illustrated the value of modulating the cytotoxicity of fluorouracil with leucovorin and demonstrated that the addition of leucovorin in the treatment of patients With advanced colorectal carcinoma is associated with a significant survival advantage compared with patients given fluorouracil alone.

Therapeutic trials using a combination of fluorouracil and leucovorin in patients with refractory breast cancer and as initial therapy in patients with previously untreated advanced disease have shown encouraging results in the past few years.550" 51 Phase II trials from Vanderbilt and Baylor Universities using a combination of mitoxantrone with leucovorinmodulated fluorouracil in previously treated advanced breast carcinoma suggest that this may be an effective second-line chemotherapy reIDmen. 55~·553

The anthracenedione mitcixantrone is considered a substitute for doxorubicin because it is less toxic. Randomized trials comparingmitoxantrone (14 mg/m2

) and doirnrubicin as second-line treatment of metastatic breast cancer demonstrated comparable activity but less alopecia, vomiting, mucositis, and cardiotoxicity in the mitoxantrone-treated group. 554 The comtiination of mitoxantrone, methotrexate, and mitomycin (MMM regimen) was as effective as CMF.555 The results from the Vanderbilt and Baylor groups, using a 21-day cycle of mitoxantrone on day 1 and 3 days offluorouracil and leucovorin, showed impressive 65% and 42% response rates, respectively, with a median duration of 6 months. It is likely that the reported findings in these previously treated populations may underestimate its effect in previously untreated patients. The combination of mitoxantrone, fluorouracil, and high-dose leucovorin appears to be active and well tolerated, and comparison with other standard combinations as first- or secondline treatment for breast cancer is indicated.

Bone Marrow Transplantation

Dose-intensity analyses provided the impetus to study a variety of high-dose regimens using autologous bone ·marrow support. 440•

441 In patients with stage IV disease, this approach .has produced high complete remission rates (35-50%), but they are insufficient to conclude that this form of therapy is superior to conventjonal drug treatments. 446·

556- 558 The duration of re

sponse in these series has not yet been convincingly demonstrated to be prolonged compared with more conventional treatments. The results have been achieved in highly selected

_.-112 of 122 Celltrion, Inc., Exhibit 1006

series of previously untreated patients, and the median followup time is still short. For clinically advanced disease, it is possible that the major obstacle to curative high-dose chemotherapy is the high frequency of drug-resistant tumor cells. If this is the case, high-dose treatment is unlikely to work in these patients, and future trials should instead concentrate on the adjuvant treatment of high-risk women (;,,10 positive axillary lymph nodes).

New techniques, including the administration of hematopoietic growth factors, have considerably decreased the toxicity and mortality associated with dose-intensive regimens · and reduced the hospital costs of the procedure. 446_..48 Additional follow-up of the current studies and randomized studies using high-dose chemotherapy with autologous bone marrow transplantation are required to assess the impact of this approach.

Biologic Response Modifiers

Lymphokines, cytokines, and monoclonal antibodies, all products of biotechnology, were devised to enable the immune system to inhibit the growth of human malignancy. 559 Several of the biologic response modifiers (e.g., interferon-a, lymphokine-activated killer cells, recombinant human tumor necrosis factor) have been investigated in phase I trials as treatment of metastatic breast cancer. The results achieved so far are disappointing or controversial. The interferons may be used as modulatory agents, because they represent a new class of biochemical modulators that appear capable of circumventing mechanisms by which malignant cells became resistant to the fluoropyrimidines.

TREATMENT OF METASTATIC COMPLICATIONS

Metastases to the Central Nervous System

Approximately 25% of patients with metastatic breast cancer eventually develop a nervous system problem: brain metastases, epidural spinal cord compression syndrome, carcinomatous meningeal infiltration, choroidal metastases, brachia! plexus syndrome, or paraneoplastic syndrome. Brain metastases are often multiple, and they usually become symptomatic late in the course of patients with disseminated disease. After diagnosis by MRI or CT scans, standard therapy is prompt administration of dexamethasone (16-24 mg/day) followed by whole-brain irradiation. About two thirds of patients improve, and some patients remain free of symptoms for many months or years. Resection before radiation therapy is indicated if the diagnosis is in doubt or if there is a single accessible metastasis. Small-field or gamma knife irradiation has been used in conjunction with standard external-beam irradiation in patients with limited metastases. The impact of systemic treatment on brain metastases has not been fully evaluated. Single-agent (e.g., lomustine, fluorouracil, high-dose methotrexate) chemotherapy, combinations of drugs, or tamoxifen may produce objective responses in 25% to 50% of patients.

Carcinomatous meningeal infiltration is a disease complication that is associated with a poor prognosis. Fewer than 10% of patients survive 12 months. The treatment involves intrathecal administration of methotrexate, thiotepa, and cytosine arabinoside in single-drug or combination regimens.


Treatment is optimally administered through an Ommaya reservoir and is usually started with methotrexate (IO mg/ m') given twice weekly until the cerebrospinal fluid cytology becomes negative. The frequency of methotrexate administration can be gradually decreased, first to a weekly course and eventually to a single administration every 2 months. Radiation therapy is generally not recommended because treatment would involve more than 40% of the bone marrow, and it cannot be delivered concomitantly with intrathecal chemotherapy. In patients with symptomatic cranial nerve palsies, whole-brain irradiation is usually given. Choroidal metastases are rare (,;;5% of all patients) and may be bilateral. The treatment of choice is radiation therapy, which can improve visual acuity in most patients.

Epidural spinal cord compression syndrome is one of the most serious disease complications. After localizing the exact level of compression by MRI or CT scans, dexamethasone and radiation therapy should be promptly instituted. In the presence of rapid development of paralysis and sphincter dysfunction, laminectomy should be considered. However, the results of surgery followed by radiation therapy have not been demonstrated to be superior to the results achieved by irradiation alone. Brachia! plexopathy can be caused by tumor infiltration or radiation damage, if prior irradiation has been given to that area. Early diagnosis of brachia! plexopathy secondary to tumor and prompt administration of radiation therapy (and systemic treatment if the patient has not been previously receiving treatment) can alleviate neurologic symptoms in most patients. If symptoms are longstanding, the likelihood of alleviating them is considerably decreased.

Bone Metastases and Hypercalcemia

Pathologic bone fractures can be devastating, and it is preferable to attempt to prevent them. A femoral fracture may result in marked limitation of mobility, and a vertebral body compression fracture may result in severe pain or profound neurologic dysfunction due to spinal cord injury. In-patients with known metastatic disease, attention must be paid to areas of bony involvement and suspicion maintained for areas where pathologic fractures can occur. If recognized promptly, fracture can be prevented in most areas of bony destruction by the early use of radiation therapy or by orthopedic stabilization. The choice of treatment depends on the site and extent of involvement and the degree of pain; in patients with painful lesions with significant (>33%) destruction of the cortex in the peritrochanteric region, the treatment is usually surgery.

Hypercalcemia is the most common life-threatening metabolic abnormality in patients with breast cancer and osseous metastases. This complication is associated with significant morbidity and mortality, especially in patients in the latter stages of their disease (see Chap. 60). Besides the use of systemic anticancer therapy, the bisphosphonates and gallium nitrate are new agents for the treatment of hypercalcemia that work as potent inhibitors of bone resorption.560 To directly compare therapeutic effectiveness, a randomized, doubleblind, multicenter study of gallium nitrate compared with etidronate was recently conducted using patients with cancerrelated hypercalcemia. 561 Both drugs were given daily for 5 consecutive days. Gallium nitrate was given by continuous intravenous infusion (200 mg/m2/day), and etidronate was



given as a 4-hour intravenous infusion (7.5 mg/kg). In 71 randomized patients, normocalcemia was achieved in 82% of patients given gallium nitrate and in 43% of those treated with etidronate. The conclusion of the study is that gallium nitrate is highly effective and superior to etidronate for acute control of moderate to severe cancer-related hypercalcemia.

REFERENCES l. American Cancer Society. Cancer statistics. CA 1992;42:30-31. 2. Harris JR, Hellman S, Henderson IC, Kinne DW. Breast diseases. 2nd ed. Philadelphia:

JB Lippincott, 1991. 3. Lippman ME, Dickson RB. Regulatory mechanisms in breast cancer. Boston: Kluwer

Academic Publishers, 1990. 4. McGuire WL, Johnson BE, Seeger RC. Oncogenes in clinical cancer: A panel discussion.

Breast Cancer Res Treat 1987; 10:217-227. 5. Lee EY-H. 'fumorsuppressor genes: A new era for molecular genetic studies of cancer.

Breast Cancer Res Treat 1991;19:3-13. 6. Liotta LA. Gene products which play a role in cancer invasion and metastasis. Breast

Cancer Res Treat 1988;11:113-124. 7. O'Malley BW. Steroid hounone receptors as transactivators of gene expression. Breast

Cancer Res Treat 1991;18:67-71. 8. Clarke R, Dickson RB, Brunne R. The process of malignant progression In human

breast cancer, Ann Oncol 1990; 1:401-407. 9. Slamon DJ, Clark GM, Wong SG. Human breast cancer: Correla1ion of relapse and

survival with amplification of1he HER-2/neu oncogene. Science 1987;235:177-182. IO. Tandon AD, Clark GM, Chamness GC. HER-2/neu oncogene protein and prognosis

in breast cancer. J Clin Oncol 1989;7:1120-1128. 11. Palk S, Hazan R, Fisher ES. Pathologic findings from the National Surgical Adjuvant

Breast and Bowel Project: Prognostic significance of erbB-2 protein overexpresslon in primary breast cancer. J Cl!n Oncol 1990;8: 103-112.

12. Lippman ME, Dickson RB, Bate S. Autocrine and paracrine growth regulation of human breast cancer. Breast Cancer Res Treat 1986;7:59-70.

13. Lippman ME, Dickson RB, Gelmann EP. Growth regulation of human breast carcinoma occurs through regulated growth factor secretion. J Cell Biochem 1987;35:1-16.

14. Osborne CK, Arteaga CL. Autocrlne and paracrine growth regulation In breast cancer: Clinical Implications, Breast Cancer Res Treat 1990;15:3-l 1.

15. Yee D, Favoni.RE, Lippman ME. Identification of insulin-like growth factor binding proteins in breast cancer cells. Breast Cancer Res Treat 1991;18:3-10.

16. Andersson I, Aspegren K. Janson L. Mammographic screening and mortality from breast cancer: The Malmo mammographic screening trial. Br Med J 1988;297:943-948.

17. Chu K, Smart C, Tarone R. Analysis of breast cancer mortality and stage distribution by age for the Health Insurance Plan clinical trial. JNCI 1988;80:1125-1132.

18. UK Tria1 of Early Detection of Breast Cancer Group. First results on inortality reduction In the UK trial of early detectiqn of breast cancer. Lancet 1988;2:411-416.

19. Roberts M. Alexander FE, Anderson T, et al. Edinburgh trial of screening for breast cancer: Mortality at seven years. Lancet 1990;335(8684):241-246.

20. Rutqvlst L, Miller A, Andersson I, et al. Reduced breast-cancer mortality with mammography screening-an assessment of currently available data. Int J Cancer 1990;5: 76-84.

21. Tabar L, Fagerberg C, South M, Day N, Duffy S. Update of the Swedish two-country trial of mammographic screening for breast cancer. Radial Clin North Am 1992;30: 187-210.

22. Wald N, Cuckle H, Frost C. Breast cancer screening: The current position. Br Med J 1991;302:845-846.

23. Fisher E, Sass R, Fisher B. Biologic considerations regarding the one and two-step procedures In the management of patients with Invasive carcinoma of the breast. Surg Gynecol Obstet 1985:161:245-249.

24. MacMahon B, Cole P. Brown J. Etiology of human breast cancer: A review. JNCI 1973;50:21-42.

25. Henderson B, Ross R, Ludd H. Do regular ovulatory cycles increase breast cancer risk? Cancer 1985;56:1206-1208.

26. MacMahon B, Trichopoulos D, Brown J. Age at menarche, probability of ovulation and breast cancer risk. Int J Cancer 1982:29:12-16.

27. Trichopoulos D, Yen S, Brown J. The effect of westernization on urine estrogens, frequency of ovulation, and breast cancer risk. Cancer 1984;53:187-192.

28. Bernstein L, Ross R, Lobo R. The effects of moderate physlcaJ activity on menstrual cycle patterns In adolescence: Implications for breast cancer prevention. Br J Cancer 1987;55:681-685.

29. Henderson B, Ross R. Bernstein L. Estrogens as a cause of human cancer: The Richard and Hinda Rosenthal Foundation Award Lecture. Cancer Res 1988;48:246-253.

30. Thomas D, Lllienfeld A. Geographic, reproductive and socloblological factors. In: Stoll BA, ed. Risk factors in breast cancer. Chicago: Heinemann Medical Books, 1976: 25-53.

31. Trichopoulos D. MacMahon B, Cole P. Menopause and breast cancer risk. JNCI 1972;48:605-613.

32. MacMahon B, Cole P. Lin T. Age at first birth and breast cancer risk. Bull WHO 1970;43:209-221.

33. Trichopoulos P, Hsieh C, MacMahon B. Age at any birth and breast cancer risk. Int J Cancer 1983;31:701-704.

34. Brinton L, Hoover R, Fraumenl J. Reproductive factors In aetiology of breast cancer. Br J Cancer 1983;47:757-762.

35. Pike M, Henderson B, Casagrande J. Oral contraceptive use and early abortion as risk factors for breast cancer ln young women. Br J Cancer 1981;43:720-776.

36. Hadjlmichael 0, Boyle C, Meigs J. Abortion before first live birth and risk of breast cancer. Br J Cancer 1986;53:281-284.

37. Moolgavkar S. Day N, Stevens R. Two stage moCl.el for carcinogenesis: Epidemiology of breast cancer in females. JNCI 1980;65:559-569.

38. Anderson T, Ferguson D, Raab G. Cell turnover In the "resting" human breast: Jnfiuence of parity, contraceptive pill, age and laterallty. Br J Cancer 1982;46:376-382.

39. Henderson I. What can a woman do about her risk of dying of breast cancer? Curr Prob! Cancer 1990;14:166-230.

40. Miller D, Rosenberg L, Kaufman D. Breast cancer before age 45 and oral contraceptive use: New findings. Am J Epidemlol 1989;129:269-280.

41. Bemsteln L, Pike M, Krai1o M, Henderson B. Update of the Los Angeles study of oral contraceptives and breast cancer. In: Mann R, ed. Oral contraceptives and breast cancer. London: Parthenon Publishing Group, 1990:169.

42. Meirik 0, Lund E, Adami H. Oral contraceptive use and breast cancer in young women. Lancet 1986;2:650-653.

43. McPhefliOn· K, Vessey M, Neil A. Early oral contraceptive use and breast cancer. Results of another case-control study. Br J Cancer 1987;56:653-660.

44. The Centers for Disease Control Cancer and Steroid Hormone Study, Long-teun oral contraceptlye use and the risk of breast cancer. JAMA 1983;259:1591-1595.

45. Paganini-Hill A, Ross R, Gerkins V. Menopausal estrogen therapy and hip fractures. Ann Intern Med 1981;95:28-31.

46. Carroll K. Experimental evidence of dietary factors and honnonedependent cancers. Cancer Res 1975;35:3374-3383.

47. Buell P. Changing incidence of breast cancer in Japanese-American women. JNCI 1973;51 :1479-1483.

48. Kinlen L. Meat and fat consumption and cancer mortality. A study of strict religious orders In Britain. Lancet 1982;1:946-949.

49. Phillips R, Garfinkel L, KuzmaJ. Mortality among California Seventh Day Adventists for selected cancer sites.JN Cl 1980;65: 1097-1107, .

50. Willett W, Stampfer M, Coldltz G, Dietary fat and risk of breast cancer. N Enil J Med 1987;316:22-28.

51. Jones G, Schatzkln A, Green S. Dietary fat and breast cancer In the National Health and Nutrition Examination Survey I epldemiologic follow-up study. JNCI 1987;79: 465-471.

52. Zollinger T, Phillips R, Kuzma J. Breast cancer survival rates among Seventh-Day Adventists and non-Seventh Day Adventists. Am J Epldemlol 1984:119:503-509.

53. Longnecker P, Berlin J, Orz M. A metaanalysis of alcohol consumption In relation to breast cancer risk. JAMA 1988;206:652-656.

54. Garftnl<el L, Boffetta P, Stellman S. Alcohol and breast cancCc.A cohort study. Prev Med 1988;17:686-693. . / I....

55. Harvey E, Schairer C, Brinton L. Alcohol consumption and breast cancer. JNCI 1987;78:657-661.

56. Veer P, Kok F, Hermus R, Sturmans F. Alcohol dose, frequency and age at first exposure in relation to the risk of breast cancer. Int J Ep!demiol 1989;18:511-517.

57. Ottman R, King M, Pike M, Henderson B. Practical guide for estimating risk for famlllal breast cancer. Lancet 1983;2:556-558.

58. Anderson D. Genetic study of breast cancer: Identification of a high risk.group. Cancer 1974;34:1090-1097.

59. Sattln R, Rubin G, Webster L. Family history and the risk of breast cancer. JAMA 1985:253:1908-1913.

60. Baak J, Van Dop H, Kurver P, Hermans J. The value of morphometry to classic prognosticators in breast cancer. Cancer 1985;56:374-382.

61. Anderson D, Badzioch M. BUaterality in famUial breast cancer patients. Cancer 1985;56: 2092-2098.

62. Lynch H, Guingis M, Broclkey F. Genetic heterogenicity and familial carcinoma of the breast. Surg Gynecol Obstet 1976;142:693-699.

63. Tokunaga M, Land C, Yamamoto T, et al. Incidence of female breast cancer among atomic bomb survivors, Hiroshima and Nagasaki, 1950-1980. Radlat Res 1987;112: 243-272.

64. Miiler A. Howe G, She~an G. Mortality from breast cancer after irradiation during fluoroscopic examinations In patients being treated for tuberculosis. N Engl J Med 1989;321:1285-1289.

65. Shore R, Hildreth N. Breast cancer among women given x-ray therapy for acute postpartum mastltls. JNCI 1986;77:689-696.

66. Hildreth N, Shore L, Dvoretsky P. The risk of breast cancer after Irradiation of the thymus in infancy. N Engl J Med 1989;321:146-151.

67. Dupont W, Page D. Risk factors for breast cancer In women with proliferative breast disease. N Engl J Med 1985;312:146-151.

68. Cancer Committee of the College of American Pathologists. ls "fl.brocystlc disease" of the breast precancerous? Arch Pathol Lab Med 1986; 110:171-173.

69. Dupont W, Page b. Relative risk of breast cancer varies with time since diagno'sis of atypical hyperplasia. Hum Pathol 1989;20:723-725.

70. Dupont W, Page D, Rogers L, Parl F. Influence of exogenous estrogens, proliferative breast disease, and other variables on breast cancer risk. Cancer 1989;63:948-951.

71. GaU M, Brinton L. Byar D, et al. Projecting Individualized probabilities of developing


breast cancer for white females who are being examined annually. JNCI 1989;81: 1879-1886.

72. Schechter M. Breast cancer risk factors: Can we select women for prophylactic mastectomy? Can J Surg 1985:28:242-244.

73. Ariyan S. Prophylactic mastectomy for precancerous and high risk lesions of the breast. CanJ Surg 1985;28:262-267.

74. Snyderman R. Cancer. Prophylactic mastectomy: Pros and cons. Cancer 1984;53: 803-808.

75. Eldar S, Mequid M, Beatty J. Cancer of the breast after prophylactic subcutaneous mastectomy. Am J Surg 1984;148:692-693.

76. Lynch M, Watson P, Conway T. Breast cancer family history as a risk factor for early onset breast cancer. Br Cancer Res Treat 1988;11:263-267.

77. Woods J, Meland N. Conservative management of full thickness nipple areolar necrosis after subcutaneous mastectomy. Plast Reconstr Surg 1989:84:258-254.

78. Pennisi V. Total mammary adenectomy with histologic evaluation and Immediate reconstruction. Plast Reconstr Surg 1981;68:510-518.

79. GoodnightJ, QuagllanaJ, Morton D. Failure of subcutaneous mastectomy to prevent the development of breast cancer. J Surg Oncol 1984;26:198-201.

80. Jackson C, Palmquist M, Swanson J. The effectiveness of prophylactic subcutaneous mastectomy in Sprague Dawley rats induced with 7 ,12-dimethylbenzanthracene. Plast Reconstr Surg 1984:73:249-260.

81. Temple W, Lindsay R, Magi E, Urbanski S. Technical considerations for prophylactic mastectomy in patients at high rlsk for breast cancer. Am J Surg 1991; 161 :413-415.

82. Moon R, Thompson H, Becci P. N-(4-hydroxyphenyl) retlnamide, a new retinoid for prevention of breast cancer In the rat. Cancer Res 1979;39:1229-1346.

83. Silverman J, Katayama S, Radok R. Effect of short-term administration of N-(4-hydroxyphenyl)-all trans-retinamide on chemically induced mammary tumors. Nu tr Cancer 1983;4:186-191.

84. McCormick D, Menta R, Thompson C. Enhanced Inhibition of mammary carclno· genesis by combined treatment with N-( 4-hydroxyphenyl) retlnamlde and ovarlectomy. Cancer Res 1982;42:508-512.

85. Preece P, Baum M, Mansel R. Importance ofmastalgla in operable breast cancer. Br MedJ 1982;284:1299.

86. Divitt J, Barr J. The clinical recognition of cystic carcinoma of the breast. Surg Gynecol Obstet 1984;159:130-132.

87. Bell D. l:lajdu S, Urban'J. The role of aspiration cytology in the diagnosis and management of mammary lesions in office practice. Cancer 1983:51 :1182-1189.

88. Strawbridge H, Bassett A, Foldes I. Role of cytology in management of lesions of the breast. Surg Gynecol Obstet 1981; 152: 1-7.

89. Urban J. Egeli R. Non-lactational nipple discharge. Cancer 1978;28:3. 90. McLaughlin C, Coe J. A study of nipple discharge In the non-lactating breast. Ann

Surg 1963;157:810. 91. Seltzer M, Perloff L, Kelley R, Fitts W. The significance of age in patients with nipple

discharge. Surg Gynecol Obstet 1970;131:519. 92. Murad T, Contessa G, Mouriesse H. Nipple discharge from the breast. Ann Surg

1982;195:259. 93. Di Petro S, Coopmansde Yoldi G, Bergorzl. S. Nipple discharge as a sign ofpreneoplastic

lesion and occult carcinoma of the breast: Clinical and galactographic study in 103 consecutive patients. Tumor! 1979;65:317-324,

94. Morrow M. Nipple discharge in breast diseases. In: Harris J, Hellman S, Henderson IC Kinne D, eds. Breast diseases. 2nd ed. Philadelphia: JB Lippincott, 1991:73-76.

95. Kline T, Joshi L, Neal H. Fine needle aspiration of the breast: Diagnoses and pitfalls. Cancer 1979:44:1458-1464.

96. Barrows G, Anderson T, LambJ. Fine needle aspiration ofbreast cancer. Relationship of clinical factors to cytology results in 689 primary malignancies. Cancer 1986;58: 1493-1498.

97. Dixon J, Anderson T, Lamb J. Fine needle asplration cytology, In relationship to clinical examination and mammography in the diagnosis of a solid breast mass. Br J Surg 1984;71:593-596.

98. Bicker T, Schandorf H, Naujoks H. Long-term follow-up In patients with mammary gland changes found unsusplclous by aspiration cytology. Cancer Detect Prev 1988; 11: 319-322. .

99. Goodson W, Mailman R, Miller T. Three year follow-up ofbenign fine-needle aspiration biopsies of the breast. Am J Surg 1987;154:58-61.

100. Fentiman I, Millis R. Hayward J. Value of needle biopsy in outpatient diagnosis of breast cancer. Arch Surg 1980;115:652-653.

101. MinkowltzS, Moskowitz R, KhaflfR, Alderete M. Tru-cut needle biopsy of the breast. An analysis of its specificity and sensitivity. Cancer 1986;57:320-323.

102. Shabot M, Goldbert I. Schick P. Aspiration cytology is superior to Tru·cut needle biopsy in establishing the diagnosis of clinically suspicious breast masses. Ann Surg 1982;196:122-126.

103. Bertario L, Reduzzl D, Piromalll D. Outpatient biopsy of breast cancer. Influence on survival. Ann Surg 1985;201:64-67.

104. Walker G, Foster R, McKegney C, McKegney F. Breast biopsy. A comparison of outpatient and inpatient experience. Arch Surg l 978;113:942-946.

105. Doberneck R. A study of cost effectiveness. Ann Surg 1980:192:152-156. 106. Morrow M. Management ofnonpalpable breast lesions. PPO Updates 1990;4:1-1 l. 107. Schwartz G, Feig S, Patchefsy A. Significance and stagingofnonpalpable carcinomas

of the breast. Surg Gynecol Obstet 1988;166:6. 108. Ciatto S, Cataliottl L, Dlstante V. Nonpalpable lesions detected with mammography.

Review of 512 consecutive cases. Radiology l 987;165:99. 109. Choucair R, Holcomb M, Matthews R. Hughes T. Biopsy ofnonpalpable breast lesions.

Am J Surg 1988;156:453.

References 1325

110. Tabar L, Gad A. Screening for breast cancer. The Swedish trial. Radiology 1981;138: 219-222.

111. Moskowitz M. The predictive value of certain mammographic signs in screening for breast cancer. Cancer 1983;51: 1007-1011.

112. Sickles E. Further experience with micro focal spot magnification mammography in the assessment of clustered microcalcifications. Radlology 1980;137:9.

113. Sickles E. Micro focal spot magnification mammography using xeroradlographic and screen film recording systems. Radiology 1979;131 :599.

114. Powell R, McSweeney M, Wilson C. X-ray calcifications as the only basis for breast biopsy. Ann Surg 1983;197:555-559.

115. GallagherW, CardenosaG, RubensJ. Minimal volume excision ofnonpalpable breast lesions. AJR 1989;153:957-961.

116. Tinnemans J, Wobbes T, Holland R. Mammographic and hlstopathologic correlation of nonpalpable lesions of the breast and the reliability of frozen section diagnosis. Surg Gynecol Obstet 1987;165:523.

117. Azavedo E, Auer G, Svane G. Stereotacdc fine needle biopsy in 2594 mammographically detected nonpalpable lesions. Lancet 1989:1:1033.

118. Dowlatshai K, Gent H, Schmidt R. Nonpalpable breast tumors: Diagnosis with ste· reotaxic localization and fine-needle aspiration, Radiology 1989;1770:427-433.

119. Gent H, Springer E. Dowlatshahl K. Stereotaxic ne'edle localization and cytological diagnosis of occult breast lesions. Ann Surg 1986;204:580.

120. Schmidt R. Morrow M, Bibbo M, Cox S. Mammographic screenlng: Potential cost benefits of stereotactlc aspiration cytology. Adminlstr Radiol 1990;9:35.

121. McDlvitt R, Stewart F, Berg J, Tumors of the breast. In: Atlas of tumor pathology. Bethesda, MD: Armed Forces Institute of Pathology, 1967.

122. World Health Organization. Hlstologic typingofbreast tumors. Tumori 1982;68:181. 123. Wellings S, Jensen H. On the origin and progression of ductal carcinoma in the

human breast. JNCI 1973;50:1111-1118. 124. Fisher E, Kenny J, Sass R, et al. Medullary cancer of the breast revisited. Breast

Cancer Res Treat 1990:16:215-229. 125, Harris M, Howell A. Chrissohou M. Swindell R, Hudson M, Sellwood R. A comparison

of metastatic pattern of infiltrating lobular and infiltrating ductal carcinoma of the breast. Br J Cancer 1984;50:23-30.

126. SprattJ, Donegan W. Cancer of the breast. Philadelphia: WB Saunders, 1967. 127. Fisher B, Slack N. Ausman R. Location of breast carcinoma and prognosis. Surg

Gynecol Obstet 1969;129:705-716. 128. Haagensen C. Diseases of the breast. Philadelphia: WB Saunders, 1971 :486-491. 129. Holland R, Veling S, Mravunac M, Hendriks J. Histologic multlfocality of Tis, Tl-2

breast carcinomas. lmpllcations for clinical trials of breast-conserving surgery, Cancer 1985;56:979-990.

130. Carter C. Allen C, Henson D. Relation of tumor size, lymph node status, and survival In 24,740 breast cancer cases. Cancer 1989;63:181-187.

131. Pickren J. Significance of occult metastases. Cancer 1961 ;14:1266-1271. 132. Bucalossi P: Veronesi U, Zlngo L, Conti C. Enlarged mastectomy for breast cancer:

Review of 1,213 cases. Am J Roentgenol Radium Ther Nucl Med 1971;111:119-122,

133. Butcher H. Radical mastectomy for mammary carcinoma. Ann Surg 196m0:883-884.

134. Haagensen C. Diseases of the breast. 3rd ed. Phlladelphla: WB Saunders, 1986:872. 135. Schottenfeld D, Nash A, Robbins G. Beattle E. Ten-year results of the treatment of

primary operable breast cancer. Cancer 1976;38: 1001-1007. 136. Valagussa P. Bonadonna G, Veronesi U. Patterns of relapse and survival following

radical mastectomy: Analysis of 716 consecutive patients. Cancer 1978;41:1170-1178.

137. Haagensen C. Treatment of curable carcinoma of the breast. Int J Radiat Oncol Biol Phys ·1977;2:975-980.

138. Fisher B, Slack N, Katrych D, Wolmar N. Ten-year follow-up results of patients with carclnoma of the breast In a co-operative clinical trial evaluating surgical adjuvant chemotherapy. Surg Gynecol Obstet 1975;140:528-534.

139. Spratt J, Donegan W. Carcinoma of the breast. Philadelphia: WB Saunders, 1971: 13& .

140. Payne W, Taylor W, Knonsarl S. Surgical treatment of breast cancer: Trends and factors affecting survival. Arch Surg 1970;101:105-113,

141. Ferguson D, Meier P, Karrlson T, et al. Staging of breast cancer and survival rates: An assessment based on 50 years of experience with radical mastectomy. JAMA 1970;248:1337-1341.

142. Nemoto T, Vana J, Bedwanl R. Management and survival of female breast cancer. Cancer 1980;45:2917-2924.

143. Smith J, Gamez-Araujo J, Gallager H, White E, McBride C. Carcinoma of the breast. Cancer 1977;39:527-532.

144. Veronesi U, Rilke F, Lulnl R. Distribution of axillary node metastases by level of invasion. An analysis of 539 cases. Cancer 1987;59:682-687.

145. Rosen P. Lesser M, Kinne D. Discontinuous or "skip" metastases in breast carcinoma: Analysis of 1228 axillary dissections. Ann Surg 1983;187:276-283.

146. Fisher B, Slack N, Bross ID. Cancer of the breast: Size of neoplasm and prognosis. Cancer 1969;24:1071-1080.

147. Adair F. BergJ.Joubert L, Robbins G. Long-term follow-up ofbreastcancerpatlents: The 30 year report. Cancer 1974;33:1145-1150.

148. Handley R. Carcinoma of the breast. Ann R Coll Surg Engl 1975;57:59-66. 149. Veronesi U, Casclnelll N. Greco M. Prognosis of breast cancer patients after mas

tectomy and dissection of internal mammary nodes. Ann Surg 1985;202:702-707. 150. Halsted W. The results of radical operations for the cure of carcinOma of the breast.

Ann Surg 1907;46:1-19.



151. Warren S, Witman E. Studies on tumor metastases: The distribution of metastases In cancer of the breast. Surg Gynecol Obstet 1937;57:1018.

152. Saphillo 0, Parker M. Metastases from primary carcinoma of the breast with special reference to spleen, adrenal glands and ovaries. Arch Surg 1941;42:1003.

153. Khansur T, Halek A, Patel B. Evaluation of bone scan as a screening work·UP in primary and local·regional recurrence of breast cancer. Am J Clin Oncol 1987;10: 167-170.

154. Baker R. Preoperative assessment of the patient with breast cancer. Surg Clln 1984;64: 1039-1050.

155. Butzelaar R, Van Dongen J, Van Der Schoot J. Evaluation of routine preoperative bOne scintigraphy in patients with breast cancer. Eur J Cancer Clio Oncol 1977;13: 19.

156. Gerber F. Goddreau J, Kirchner P: Efficacy of preoperative and postoperative bone scanning In the management of breast carcinoma. N Engl J Med 1977;297:300.

157. Green D, Jeremy R, Towson J. The role of fluorine 18 scanning In the detection of skeletal metastases. Aust N Z J Surg 1973:43:251.

158. Lee Y. Bone scanning In patients with early breast carcinoma: Should it be a routine staging procedure? Cancer 1981;47:486.

159. Nomura Y, Kondo H, Yamagata J. Evaluation of liver and bone scanning In patients with early breast cancer, based on results obtained from more advanced cancer patients. Eur J Cancer 1978;14:1129.

160. Welner S, Sachs S. An assessment of routine liver scanning In patients with breast cancer. Arch Surg 1978;113:126.

161. Sears H, Gerber F, Sturtz D, Liver scan and carcinoma of the breast. Surg Gynecol Obstet 1975;140:409.

162. Chaudry M, Millis R, Hoskins E. Bilateral primary breast cancer: A prospective study of disease incidence. Br J Surg 1984;71:711-714.

163. Lee Y. Carcinoembryonlc antigen as a monitor of recurrent breast cancer. J Surg Oncol 1982;20:109.

164. Colomer R, Ruiba1 A, Navarro M. Circulating CA 15-3 levels in breast cancer. Our present expertence. Int J Biol Markers 1986;1:89.

165. Gion M, Mione R, Dlttadi R. Evaluation of CA 15-3 serum levels In breast cancer patients. J Nucl Med Allied Sci 1986;30:29.

166. Hayes D, ZurawskiV, Kufe D. Compartson of circulating CA 15-3 and carcinoembryonic antigen levels in patients with breast cancer. J Clln Oncol 1986;4:1542.

167. Huvos A, Hutter V, Berg J. Signlflcance of axillary macro metastases and mlcrome· tastases In mammary cancer. Ann Surg 1971;1:44-46.

168. Halsted W. The results of operations cure of cancer of the breast performed at Johns Hopkins Hospital. Johns Hopkins Hosp Bull 1984;4:497-555.

169. Haagensen C, Bodlan C. A personal experience with Ha1sted's radical mastectomy. Ann Surg 1984:199:143-150.

170. Donegan W, Perez·MesaC, Watson F. A blostatistical study oflocally recurrent breast carcinoma. Surg Gynecol Obstet 1966;122:529-540.

171. Brinkley D, Haybittle J. The curability of breast cancer. Lancet 1975:2:95-97, 172. Fentiman I, Cuz.ickJ, Millis R. Which patients are cured of breast cancer? Br Med J

1984;289:1108-111 L . 173. Rosen P, Groshen S, Salgo P, Kinne D, Hellman S. A long.term follow·up study of

survival in stage I and stage II breast carcinoma. J Clin Oncol 1989;7:355-366. 174. Say C, Donegan W. A biostatistical evaluation of complications from mastectomy.

Surg Gynecol Obstet 1974;138:370-376. 175. Forrest A, Roberts M, Cant E. Simple mastectomy and pectoral node biopsy: The·

Cardiff-St. Mary's Trial. World J Surg 1977;1:320-323.

176. Turner-Warwick R. Lymphatics of the breast. Br J Surg 1958;46:574-582. 177. Urban J. Management of operable breast cancer. The surgeon's view. Cancer 1978;42:

2066-2077, 178. Meir P. Ferguson D, Harrison T. A controlled trial of extended radical mastectomy.

Cancer 1985;55:880-891. 179. Veronesi U, Valagussa P, Inefficacy of internal mammary node dissection In breast

cancer surgery. Cancer 1981;47: 170-175. 180. Morrow M, Foster R. Staging a breast cancer. A new rationale for internal mammary

node biopsy. Arch Surg. 1981;116:748-751. 181. Wilson R. Donegan W, Mettlin C. The 1982 national survey of carcinoma of the

breast in the United States by the American College of Surgeons. Surg Gynecol Obstet 1984;159:309-318.

182. Mann B, SametJ, Hunt W. Changing treatment of breast cancer In New Mexico from 1969 through 1985. JAMA 1988;259:3413-3417.

183. Altken D, Minton J, Complications associated with mastectomy. Surg Clin North Am 1983;63:1331-1352.

184. Robinson G, Van HeerdenJ, Payne WEA. The primary surgical treatment of carcinoma of the breast. A changing trend toward modified radical mastectomy. Mayo Clin Proc 1976;51:433-442.

185. Baker R, Montague A, Childs J. A comparison of modified radical mastectomy to radical mastectomy Jn the treatment of operable breast cancer. Ann Surg 1979;189: 553-559.

186. Turner L, Swindell R, Bell W. Radical versus modified radical mastectomy for breast cancer. Ann R_ Coll Surg Engl 1981;63:239-243.

187. Maddox W, Carpenter J, Laws H. A randomized prospective tr:la1 of radical mastectomy versus modified radical mastectomy in 3 i I breast cancer patients. Ann Surg 1983; 198: 207-212.

188. Fisher B. Redmond C, Fisher E. Ten year results of a randomized clinical trial com· paring radical mastectomy and total mastectomy with or without radiation. N Engl J Med 1985;312:674-681.

189. Kinne D, De Cosse J, Modified radical mastectomy for carcinoma of the breast. Ann Surg 1982;48:543-556.

190. Lagios M, Westdahl P, Rose M. The concept and implications of multicentrtcity In breast carcinoma. Pathol Annu 1981;16:83-102.

191. Fisher B, Redmond C, Poisson R, et al. Eight year results of a randomized clinical trial comparing total mastectomy and lumpectomy with or without Irradiation in the treatment of breast cancer. N Engl J Med 1989;320:822-828.

192. Lagios M, Richards V, Rose M, Yee E. Segmental mastectomy without radiotherapy. Short·tenn follow·up. cancer 1983;52:2173-2179 ..

193. Clark R, Wilkinson R, Mahoney L. Breast cancer: A 21-year experience with conservative surgery and radiation. Int J Radiat Oncol Biol Phys 1982;8:967-975.

194. Montgomery A, GreeningW, Levene A. Clinical study of recurrence rate and survival time of patients with carcinoma of the breast treated by biopsy excision without any other therapy, JR Soc Med 1978;71:339-342.

195. Tagart R. Partial mastectomy forbreast cancer. Br Med J 1978;2:1268. 196. Freeman C, Belliveau N. Kim T. Limited surgery with or without radiotherapy for

early breast carcinoma. J Can Assoc Radiol 1981;32:125-128. 197. Veronesi W, Salvadori B, Luini A. Conserva~ive treatment of early breast cancer.

Long·term results of 1232 cases treated with quadrantectomy, axillary dissection, and radiotherapy. Ann Surg 1990:211:250-259.

198. Recht A, Silen W, Schnltt S, et al. Time course of local recurrence following con· servative surgery and radiotherapy for early stage breast cancer. Int J Radiat Oneal Biol Phys 1988;15:255-261.

199. Veronesi U, Saccozzi R, Del Vecchio M. Comparing radical mastectomy With quadrantectomy, axillary dissection, and radiotherapy in patients with small cancers of the breast. N Engl J Med 1981;305:6-11.

200. Veronesi U, Banfi A, Salvador! B. Breast conservation is the treatment of choice In small breast cancer: Long·term results of a randomized trial. Eur J Cancer 1990;26: 668-670,

201. Sarrazin D, Le M, ArrlagadaR, et al. Ten-year results of a randomized trial comparing a conservative treatment of mastectomy in early breast cancer. Radiother Oncol 1989;14:177-184.

202. Straus K, Lichter A, Lippmann M, et al. Results of the National Cancer Institute early breast cancer trial. JNCI Monographs 1992; 11:27-32.

203. Habibollahi F, Fentiman I, Chandry M. Conservation treatment of operable breast cancer. Proc Am Soc Clin Oncol 1987;6:59.

204. van Dongen JA, Bartellnk H, Fentlman IS. et al. Randomized clinical trial to assess the value of breast-conserving therapy ln stage I and II breast cancer, EORTC 10801 trial. JNCI Monographs 1992;11:15-18.

205. Blichert-Toft M, Rose C, Andersen JA, et al. Danish randomized trial comparing breast conservation therapy With mastectomy: Six years of life table analysis. JNCI Monographs 1992;11:19-26.

206. National Institute for Health Consensus Conference. Treatment of early stage breast cancer. JAMA 1991;265:391-395.

207. Margo Iese R, Poisson R, Shibata H. The technique of segmental mastectomy (lump· ectomy) and axillary dissection: A syllabus from the Natiori:tl..Surgical Adjuvant Breast Project workshops. Surgery 1987;102:828-834.

208, Veronesi U, Volterranl F, Luini A. Quadrantectomy versus lumpectomy for small size breast cancer. Eur J Cancer 1990;26:671-673.

209. Harris J, Recht A. Conservative surgery and radiotherapy. In: Harrls J, Hellman S, Henderson IC, Kinne D, eds. Breast diseases. 2nd ed. Philadelphia: JB Lippincott, 1991 :399-404.

210. Vicini F, Eberlein T. Connolly J, et al. The optimal extent of resection for patients with stages I or II breast cancer treated with conservative surgery and radiotherapy. Ann Surg 1991;214:200-205.

211. Holland R. Connolly JL, Gelman R. et al, The presence of an extensive lntraductal component following a limited excision correlates with prominant residual disease In the remainder of the breast, J Clin Oncol 1990;8(1):113-118.

212. Fourquet A. Campana F, Zafrani B. Prognostic factors of breast recurrence in the conservative management of early breast cancer: A 25·year follow·up. Int J Radiat Oncol Biol Phys 1989:17:719-725.

213. Delouche G, Bache lot F. Premont M, KurtzJ. Conservation treatment of eariy breast cancer: Long·term results and complications. Int J Radlat Oncol Biol Phys 1987;13: 29-34.

214. Kurtz J, Jacquemier J. Amalric R, et al. Why are local recurrences after breast· conserving therapy more frequent In younger patients? J Clin Oncol 1990;8(4):591-598.

215. Fisher B, Redmond C, Dimitrov N, et al. A randomized clinical trial evaluating se· quential methotrexaie and fluorouracil in the treatment of patients with node negative breast cancer who have estrogen receptor negative tumors. N Engl J Med 1989;320: 473-478.

216. Fisher B, Costantino J, Redmond C, et al. A randomized clinical trial evaluating tamoxifen in the treatment of patients with node negative breast cancer who have estrogen receptor positive tumors. N Engl J Med 1989;320:479-484.

217. FrazierT, Wong R. Rose D. Implications of accurate pathologic margins in the treat· ment of primary breast cancer, Arch Surg 1989;124:37-38.

218. Carter D. Margins of "lumpectomy" for breast cancer. Hum Pathol 1986:17:330-332.

219. Danforth D, Findlay P, McDonald H. Complete axillary lymph node dissection for stage 1-11 carcinoma of the breast. J Clin Oncol 1986:4:655-662.

220. Chevinsky A. Ferrara J, James A. Prospective evaluation of clinical and pathologic detection of axillary metastases in patients with carcinoma of the breast. Surgery 1990:108:612-618.


221. Schwartz G, Rosenberg A, DanoffB. Lum pee to my and level I axillary dissection prior to irradiation for operable breast cancer. Ann Surg 1984;200:554-560.

222. Siegel B. Mayze! K, Love S. Level I and II axillary dissection in the treatment of early stage breast cancer. Arch Surg 1990;125:1144-1147.

223. Clark D, Martinez A, Cox R. Breast edema following staging axillary node dissection in patients with breast carcinoma treated by radical radiotherapy. Cancer 1982;49: 2295-2299. •

224. Brismar B, Ljungdahl I. Postoperative lymphoedema after treatment of breast cancer. Acta Chir Scand 1983;149:687-689.

225. Feigenberg Z, Zer M, Dintsman M. Comparison of postoperative complications following radical and modified radical mastectomy. World J Surg 1977;1:207-211.

226. Budd D, Cochran R, Stutz D. Surgical morbidity after mastectomy operations. Am J Surg 1978:135:218-220.

227. Larson D, Weinstein M, Goldberg I, et al. Edema of the arm as a function of the extent of axillary surgery in patients with stage 1-11 carcinoma of the breast treated with primary radiotherapy. Int J Radiat Oncol Biol Phys 1986;12:1575-1582.

228. Pezner R, Patterson M, Hill L. Arm edema in patients treated conservatively for breast cancer: Relationship to palient age and axillary node dissection technique. Int J Radiat Oncol Biol Phys 1986;12:2079-2083.

229. Chin L, Cheng C, Siddon R, Rice R, Mijnheer B, Harris J. Three dimensional photon dose distributions with and without lung corrections for tangential breast intact treatments. Int J Radlat Oncol Biol Phys 1989;17:1327-1335.

230. Bornstein B, Cheng C, Rhodes LM, et al. Can simulation measurements be used to predict the irradiated lung volume in the tangential fields In patients treated for breast cancer? lntJ Radiat Oncol Biol Phys 1.990;18:181-187.

231. Pierce SM, Recht A, Lingos TI, et al. Long-term radiation complications following conservative surgery and radiation therapy in patients with early stage breast cancer. Int J Radlat Oncol Blol Phys 1992;23:915-923.

232. Sadowsky N, Semine A, Harris H. Breast imaging: A critical aspect of breast conserving treatment. Cancer 1989;65:2113-2118.

233. Healey E, OstC.en R, Schnltt S, et al. Can the clinical and mammographlc findings at presentation predict the presence of an extensive intraductal component in early stage breast cancer? Int J Radiat Oneal Blol Phys. 1989;17:1217-1221.

234. FrazierT, Noone R. An objective analysis oflmmedlate simultaneous reconstruction In the treatment of primary carcinoma of the breast. Cancer 1985:55:1202-1205.

235. Kroll S, Ames F, Singletary S, Schusterman M. The oncologic risks of skin preservation at mastectomy combined with Immediate reconstruction of the breast. Surg Gyn Obstet 1991;172:17-20.

236. Johnson C, Van Heerden J, Donohue J. Oncological aspects of immediate breast reconstruction following mast_ectomy for malignancy. Arch Surg 1989;124:819-824.

237. Webster D, Mansel R, Hughes L. Immediate reconstruction of the breast after mastectomy. ls it safe? Cancer 1984;53:1416-1419.

238. Stevens L, McGrath M, Durss R. The psychological Impact of lnimediate breast reconstruction for women with early breast cancer. Plast Reconst Surg 19!34;73:619-626.

239. Vinton A, Traverso W, Zehring R. Immediate breast reconstruction following mastectomy is as safe as mastectomy alone. Arch Surg 1990;125:1303-1308.

240. Eberlein T, Cresop L, Smith B. Prospective evaluation of immediate reconstruction following mastectomy. Paper Session, Clinical Congress of the American College of Surgeons, 1991.

241. Kessler DA. The basis of the FDA's decision on breast implants. N Engl J Med 1992;326: 713-715.

242. Dea pen D, Pike M, Casagrande J, Brody G. The relationship between breast cancer and augrnentation'mammoplasty: An epidemiologic study. Plast Reconstr Surg 1986;77: 361-367. .

243. Glasser J, Lee N, Wingo P. Does breast augmentation Increase the risk of breast cancer? [Abstract) Epiderriic Intelligence Service Conference. Atlanta: Centers for Disease Control, 1989.

244. May D, Stroup N. The in.cldence'of sarcomas of the breast among women in the United States, 1973-1986. Plast Reconstr Surg 1991;87:193-194.

245. Weissman M, Vecchione D, Albert L, Moore M. Connective tissue disease following breast augmentation: A preliminary test of the human adjuvant disease hypothesis. Plast Reconstr Surg 1988;82:626-630.

246. Van Nunen S, Gatenby P, Basten"· A. Post-mammoplasty connective tissue disease. Arthritis Rheum 1982:25:694. ·

247. Kumagai Y, Shiokawa Y, Medsger·T, Rodnan G. Clinical spectrum of connective tissue disease after cosmetic surgery. Arthritis Rheum 1984;27:1.

248. Byron M, Venning V, Mowat A. Post-mammoplasty human adjuvant disease. Br J Rheumatol 1984;23:227.

249. Krizek T. Breast reconstnlction after mastectomy. In: Harris J, Hellman S, Henderson IC, Kinne D. eds. Breast diseases. 2nd· ed. PhUadelphia: JB Lippincott, 1991:491-495.

250. Hartrampf C. Breast reconstruction with a transverse abdominal island flap. A retrospective evaluation of 335 patients. Perspect Plast Surg 1987;1 :123.

251. BunkisJ, Walton R, Mathes S. Experience with the transverse lowerrectus abdominis operation for breast reconstruction. Plast Reconstr Surg 1983;72:819-827.

252. Shaw W. Microv3.scular free flap breast reconstruction. Clin Plast Surg 1984;1 l :333-341.

253. Allison A, Howorth M. Carcinoma In a nipple pi-eserved by heterotopic auto implan· talion. N EnglJ Med 1978;298:1132.

254. Parry R, Cochran T, Wolfort F. When is there nipple Involvement in carcinoma of the breast? Plast Reconstr Surg 1977;59:535-537.

References 1327

255. Smith J, Payne W, Camey J. Involvement of the nipple and areola in carcinoma of the breast. Surg Gynecol Obstet 1976;143:546-548.

256. Andersen J, Pallesen R. Spread to the nipple and areola in carcinoma of the breast. Ann Surg 1979;189:367-372.

257. Lagios M, Gates E.. Westdahl P. A guide to the frequency of nipple involvement in breast cancer .. A study of 149 consecutive mastectomies using a serial sub gross and correlated radiographic technique. Am J Surg 1979;138:135-142.

258. McCarty K, Barton T, Georgiade N. Selection of patients for heterotoplc Implantation of the areola and nipple. Surg Gynecol Obstet 1980;150:545-547.

259. Becker H. The use of intradermal tattoo to enhance the final result of nipple-areola reconstruction. Plast Recontr Surg 1986:77:673-675.

260. Cuzick J, Stewart H, Peto R. et al. Overview of randomized trials of postoperative adjuvant radiotherapy in breast cancer. Cancer Treat Rep 1987;71:15-29.

261. Schaake-Konlng C, Van der Linden E, Hart G. Adjuvant chemo· and hormonal therapy in locally advanced breast cancer: A randomized clinical study. Int J Radlat Oncol Blol Phys 1985;11:1759-1763.

262. Overgaard M, ChristensenJ, Rose C, et al. Importance of loco-regional tumour control in high-risk breat cancer patients given adjuvant systemic treatment with or without radiotherapy. Presented at the Fifth EORTC Breast Cancer Working Conference, 1991.

263. Griem K, Henderson I, Gelman R, et al. The 5-year results of a randomized trial of adjuvant radiation therapy after chemotherapy in breast cancer patients treated with mastectomy. J Clin Oncol 1987;5:1546.

264. Baron P, Moore M, Kinne D. Occult primary cancer presenting with axillary metastases. Arch Surg 1990;125:210-214.

265. Vezzonl P, Balestrazti A, Bignaml P. Axillary lymph node metastases from occult carcinoma of the breast. Tumori 1979;65:87-91.

266. Ellerbroek N, Holmes F, Singletary E. Treatment of patients with Isolated axillary nodal metastases from an occult primary carcinoma consistent with breast origin. Cancer 1990;66:1461-1467.

267. Patel J, Nemoto T, Rozner D, Dao T. PickrenJ. Axillary lymph node metastasis from an occult breast cancer. Cancer 1981 ;47:2923-2927.

268. Kemeny M, Rivera D, Terz J, Benfield J. Occult primary adenocarcinoma with axillary metastases. Am J Surg 1986;152:43-47.

269. Iglehart J, Ferguson B. Shingleton W. An ultrastructural analysis of breast carcinoma presenting as Isolated axillary adenopathy. Ann Surg 1982;196:8-13.

270. Bhatia S, Saclarides T, Witt T. Hormone receptor studies in axillary metastases from occult breast cancer. Cancer 1987;59:1170-1172.

271. Rosen P, Kimmel M. Occult breast carcinoma presenting with axillary lumph node metastases: A follow-up study of 48 patients. Hum Pathol 1990;21:518-523.

272. Ashikari R, Rosen P, Urban J, Senoo T. Breast cancer presenting as an.axillary mass. Ann Surg. 1976;183:415-417.

273. Westbrook K. Gallager H. Breast carcinoma presenting as an axillary mass. Am J Surg 1971;122:607-611.

274. Ozzello L, Sanp'itak P. Epithellal-stromal junction of intraductal carcinoma of the breast. Cancer 1970;26:1186-1198.

275. Fitts W, Steiner G, Enterline H. Prognosis of occult carcinoma of the breast. Am J Surg 1963;106:460-463.

276. Vilcoq J, Calle R, Ferme F, Veith F. Conservative treatment of axillary adenopathy due to probable subclinical breast cancer. Arch Surg 1982; 117:1136-1138,

277. Campana F, Forquet A, Ashby M. Presentation of axillary lymphadenopathy without detectable breast primary (TONlB breast cancer): Experience at lnstitut Curie. Radio! Oncol 1989;15:321-325.

278. Donegan W. Cancer of the breast. 3rd ed. Philadelphia: WB Saunders, 1988:389. 279. Swain MS. Selection of therapy for stage Ill breast cancer. SurgClin North Arn 1990;70:

1061-1080. 280. Haagensen C, Stout A. Carcinoma of the breast. Criteria of operability. Ann Surg

1943;118:859-870. 281. FletcherG, Montague E. Carcinoma oft he breast; criteria of operability. AmJ Roent

genol Radium Ther Nucl Med 1965;93:573-584. 282. Sheldon T, Hayes D, Cady B. et al. Primary radiation for locally advanced breast

cancer. Cancer 1987;60:1219-1225. 283. Parker L, Boyages J, Eberlein T, et al. Inflammatory carcinoma of the breast. In:

Harris J, Hellman S, Henderson IC, Kinne D, eds. Breast diseases. 2nd ed-:-rufiladelphla: JB Lippincott, 1991:775-782.

284. BedwlnekJ, Rao D, Perez C. Stage Ill and localized stage IV breast cancer: Irradiation alone vs. irradiation plus surgery. Int J Radial Oncol Biol Phys 1982;8:31-36.

285. Spanos W, Montague E, Fletcher G. Late complicatlonsofradiation only for advanced breast cancer. Int J Rad.lat Oncol Biol Phys 1980;6:1473-1476.

286. Bonadonna G. Conceptual and practical advances In the management of breast cancer. Karnofsky Memorial Lecture. J Clin Oncol 1989;7:1380-1397.

287. Jacqulllat C, Well M, Baillet F. Results of neoadjuvant chemotherapy and radiation therapy in the breast-conserving treatment of250 patients with all stages of infiltrative breast cancer. Cancer 1990;66:119-129.

288. Ragaz K. Emerging modalities for adjuvant therapy of breast cancer: Neoadjuvant chemotherapy. NCI Monogr 1986;1:145-152.

289. Honobagyi GN. Comprehensive managementoflocally advanced breast cancer. Cancer 1990;66:1387-1391.

290. Delena M, Varin! M, Zucali R. Multinodal treatment for locally advanced breast cancer. Cancer Clin Trials 1981 ;4:229-236.

291. Perloff M, Lesnick G, Korzun A. Combination chemotherapy with mastectomy or radiotherapy for stage Ill breast carcinoma: A Cancer and Leukemia Group B study, J Clin Oncol 1988;6:261-269.



292. Terz J, Romeroo C, Kay S. Preoperative radiotherapy for stage Ill carcinoma of the breast. Surg Gynecol Obstet 1978; 14 7:497-502.

293. Zucali R, Uslenghl C, Kenda R, Bonadonna G. Natural history and survival of inoperable breast cancer treated with radiotherapy and radiotherapy followed by radical mas· tectomy. Cancer 1976;37:1422-1431.

294. Balawajder I, Antich P, Boland J. An analysis of the role of radlotherapy alone and in combination with chemotherapy and surgery in the management of advanced breast carcinoma. Cancer 1983;51:574-580.

295. Gohn P. Heinonen E, Klefstrom P. Adjuvant postoperative radiotherapy, chemotherapy, and immunotherapy in stage III breast cancer. Cancer 1984;54:670-674.

296. Rubens R, Bartelink H, Englesman E. Locally advanced breast cancer: The contribution of cytotoxic and endocrine treatment of radiotherapy. An EORTC Breast Cancer Cooperative Group trial (10792). Eur J Cancer Clin Oncol 1989;25:667-678.

297. Israel L, Breau J, Morere J. Two years of high dose cyclophosphamide and 5-fluorouracU followed by surgery after 3 months for acute Inflammatory breast carcinomas. Cancer 1986:57:24-28.

298. Awain S, Sorace R, Bagley C. Neoadjuvant chemotherapy in the combined modality approach of locally advanced nonmetastatic breast cancer. Cancer Res 1987;47:3889-3894.

299. Morrow M, Bravennan A, Thelmo W. Multimodal therapy for locally advanced breast cancer. Arch Surg 1986;121:1291-1296.

300. Feldman L, Hortobagyi G, Buzdar A. Pathological assessment of response to induction chemotherapy in breast cancer. Cancer Res 1986;46:2578-2581.

301. Broadwater J, Edwards M. Kuglen C. Mastectomy following preoperative chemotherapy. Ann Surg 1991;213:126-129.

302. Hortobagyi GN, Ames FC, Buzdar AU. Management ofst?ge III primary breast cancer with primary chemotherapy, surgery and.radiation therapy. Cancer 1988;62:2507-2516.

303. Valagussa P, Zambetti M, Bonadonna G. Prognostic factors in locally advanced non inflammatory breast cancer. Long-tenn results following primary chemotherapy. Breast Cancer Res Treat 1990;15:137-147.

304. Bonadonna G. Evolving concepts in the systemic adjuvant treatmentofbreast cancer. Cancer Res 1992;52(8):2127-2137.

305. Buzzonl R, Bonadonna G, Valagussa P. Adjuvant chemotherapy with doxorubicln plus cyclophospharnide, methotrexate, and fluorouracil in the treatment ofresectable breast cancer with more than three positive axillary nodes. J Clin Oncol 1991;9:2134-2140.

306. De Lena M, Zucall R, Vlganottl G. Combined chemotherapy and radiotherapy approach ln locallf advanced breast cancer. Cancer Che mother Pharmacol 1978;1 :53-59.

307. Fastenberg NA, Buzdar AU, Montague ED. Management of inflammatory carcinoma · . of the breast. A combined modality approach. Am J Clin Oncol 1985;8: 134-141.

308. Perez CA, Fields NC. Role of radiation therapy for locally advanced and Inflammatory carcinoma of the breast. Oncology 1987;1:81-93.

309. Fowble BF, Glover D. Rosato ER. Combined modality treatment of inflammatory breast cancer. Int} Radial Oncol Biol Phys 1986;12:11-12.

310. Rouesse J. Sarrazin D, Mourlesse H. Primary chemotherapy In the treatment of inflammatory breast carclnoma: A study of230 cases from the Institut Gustave-Roussy. J Clln Oncol 1986;4:1765-1771.

311. Roue"sse J, Friedman S, Mouriesse H. Therapeutic strategies In inflammatory breast carcinoma based on prognostic factors. Breast Cancer Res Treat 1990;16:15-22.

312. Fields JN, Kuske RR, Perez CA. Prognostic factors In inflammatory breast cancer. Univariate and multivariate analysis. Cancer 1989;63:1225-1232.

313. Nance F, De Loach D, Welsh R, Becker W. Paget's disease of the breast. Ann Surg 1970;171:684-874.

314. Ashikari R, Park K. Huvos AU J. Paget's disease of the breast. Cancer 1970;26:680-685.

315. PaoneJ, Baker R. Pathogenesis and treatment of Paget'sdisease of the breast. Cancer 1981;48:825-829.

316. Paget J. Disease of the mammary areola preceding cancer of the mammary gland. St Bart Hosp Rep 1874;10:79-89.

317. Osteen R. Paget's disease of the nipple. In: HarrisJ, Hellman S, Henderson IC, Kinne D, eds. Breast diseases. 2nd ed. Philadelphia: JB Lippincott, 1991:797-803.

318. Maier W, Rosemond G, Harasym E. Paget's disease In the female breast. Surg Gynecol Obstet 1969;128:1253-1263.

319. Inglis K. Paget's disease of the nipple. London: Oxford University Press, 1936:10. 320. Muir R. Pathogene'sis of Paget's disease of the nipple and associated lesions. Br J

Surg 1935;22:728-737. 321. Orr J, Parish D. The nature of the nipple changes In Paget'sdisease. J Pacho! Bacterial

1962;84:201-208. 322. Lagios M, Westdahl P, Rose M. Concannon S. Paget's disease of the nipple. Alternative

management in cases without or with minimal extent of underlying breast carcinoma. Cancer 1984;54:545-551.

323. Fourquet A. Campana F, Vielh P. Paget's disease of the nlpple without detectable breast tumor: Conservative management with radiation therapy. Int J Radiat Oncol Biol Phys 1987;13:1463-1465.

324. Boring C, Squires T, TongT. Cancer Statistics, 1991. CA 1991;41:19-51. 325. Kozak F, Hall J, Band P. Familial breast cancer in males. A case report and review

of the literature. Cancer 1986;58:2736-2739. 326. Jackson A, Muldal S, Ockey C, O'Connor P. Carcinoma of the male breast in association

with Klinefelter syndrome. Br Med J 1965;1:223-225. 327. El-Gazarerli M, Abdul-Aziz A. On bilharzlasls and male breast cancer in Egypt: A

preliminary report and review of the literature. Br J Cancer 1963;17:566-571. 328. Eldar S. Nash E, Abrahamson J. Radiation carcinogenesis in the male breast. Eur J

Surg Oncol 1989:15:274-287.

329. Ramantanis G, Besbeas S, Garas J. Breast cancer In the male: A report of 138 cases. World J Surg 1980;4:621-624.

330. Axelsson J, Andersson A. Male breast cancer. World J Surg 1983;7:281-287. 331. Heller K, Rosen P, Schottenfeld D. Male breast cancer: A clinlcopathologic study of

97 cases. Ann Surg 1978;188:60-65. 332. Treves N. Holleb A. Cancer of the male breast. A report of 146 cases. Cancer 1955;8:

1239-1250. 333. Langlands A, Maclean N, Ken G. Cancer of the male breast: Report of a series of 88

cases. Clin Radiol 1976;27:21-25. 334. Donegan W, Perez-Mesa C. Carclnoma of the male breast. A 30-year review of 28

cases. Arch Surg.1973;106:273-279. 335. Gupta S, Khanna N, Khanna S. Paget'sdisease of the male breast. A cllnlcopathologic

study and a collective review. J Surg Oncol 1983;22 151-156. 336. Treves N. Inflammatory carcinoma of the breast in the male patient. Surgery 1953;34:

810-820. 337. Friedman M, Hoffman P, Dandolos E. Estrogen receptors in male breast cancer:

Clinical and pathologic correlations. Cancer 1981;47:134-137. 338. Everson R, Lippman M, Thompson E. Clinical correlations of steroid receptors and

male breast cancer. Cancer Res 1980;40:991-997. 339. Kinne D, Hakes T. Male breast cancer. In: Harris J, Hellman S, Henderson JC, Kinne

D, eds. Breast diseases. 2nd ed. Philadelphia: JB Lippincott, 1991 :782-790. 340. Bezwoda W, Hesdorffer C, Dansey R. Breast cancer In men, Clinical fe~tures, honnone

receptor status, and response to therapy. Cancer 1987;60:1337-1340. 341. Krayblll W, Kaufman R, Kinne D. Treatment of advapced male breast cancer. Cancer

1981;47:2185-2189. 342. Yap H, Toshima C, Blumenschein G. Chemotherapy for advanced male breast cancer.

JAMA 1980;243:1739-1741. 343. ·Pitel H. Buzdar A, Hortobagyi G. Role of adjuvant chemotherapy in male breast

cancer. Cancer 1989;64:1583-1585. 344. Bagley C; Wesley M, Young R. Adjuvant chemotherapy in males wlth cancer of the

breast. Am J Clin Oncol 1987;10:55-60. 345. White T. Prognosis of breast cancer for pregnant and nursing women; analysis of

1413 cases. Surg Gynecol Obstet 1955;100:661. 346. Applewhite R, Smith L, Divicenti F. Carcinoma of the breast with pregnancy and

lactation. Am Surg 1973;39:161-104. 347. Treves N, Holleb A. A report of 549 cases of breast cancer in women 35 years of age

or younger. Surg Gynecol Obstet 1958; 107:271-283. 348. TorresJ, Mlckal A. Carcinoma of the breast in pregnancy. CUn Obstet Gynecol 1975;18:

219-225 . 349. Byrd B. Bayer D, RobertsonJ. Treatment of breast tumors associated with pregnancy

and lactation. Ann Surg 1962;155:940-947. 350. Donegan W. Breast cancer and pregnancy. Obstet Gynecol 1977;60:244-252. 351. King R, Welch J, Martin J, Coulam C. Carcinoma of the breast associated with preg

nancy. Surg Gynecol Obstet 1985;160:228-32. 352. Nugent P, O'Connell T. Breast cancer and pregnancy. Arch Surg 1985;120:1221-

1224. 353. Holleb A, Farrow J. The relation of carcinoma of the breast and pregnancy in 283

patients. Surg Gynecol Obst et 1962; 115:65-71. 354. PetrekJ. Breast cancer and pregnancy. In: HarrisJ, Hellman S, Henderson JC, Kinne

D, eds. Breast diseases. 2nd ed. Philadelphia: JB Lippincott, 1991:809-816. 355. Ribeiro G, Palmer M. Breast carcinoma associated with pregnancy: A clinician's di·

lemma. Br Med J 1977;2:1524-1527. 356. Bunker M, Peters V. Breast cancer associated with pregnancy or lactation. Am J

Obstet & Gy.necol 1963;85:312-319. 357. Schaplra D, Chudley A. Successful pregnancy following continuous treatment with

combination chemotherapy before conception and throughout pregnancy. Cancer 1984;54:800-803.

358. Sweet D, Kinzie J. Consequences of radiotherapy and antineoplastlc therapy for the fetus. J Reprod Med 1976;17:241-246.

359. Fisher B. Ten-year results of a randomized trial comparing radical mastectomy and total mastectomy with or without radiation:{:! Engl J Med 1985;312(11):674-681.

360. Zimmerman K, Montague E, Fletcher G.l Frequency, anatomical distribution and managemeilt of !peal recurrences after definitive therapy for breast cancer. Cancer 1966;19:67-74.

361. Gilliland M, Barton R, Copeland E. The implications of local reCl!rrence of breast cancer as the first site of therapeutic failure. Ann Surg 1983;197:284-287.

362. Marshall K. Redfern A, Cady B. Local recurrences of carcinoma of the breast. Surg Gynecol Obstet 1974;139:406-408.

363. Fentiman I, Matthews P. Davison 0. Survival following local skin recurrence after mastectomy. Br J Surg 1985;72:14-16.

364. Recht A, Hayes D. Local recurrence following mastectomy. In: Harris J, Hellman S. Henderson IC, Kinne D, eds. Breast diseases. 2nd ed. Philadelphia: JB Lippincott. 1991:527-540.

365, Aberizk W, Silver B, Henderson I, et al. The use of radiotherapy for treatment of isolated loco regional recurrence ofbreast cancer after mastectomy. Cancer 1986;58: 1214-1218.

366. Beck T, Hart N, Woodard D, Smlth C. Local or regionally recurrent carcinoma of the breast: Results of therapy ln 121 patients. J Clin Oncol 1983;1:400-405.

367. Bedwinek J, Fineberg B, Lee J. Analysis of failures following local treatment of lsOlated local-regional recurrence of breast cancer. Int J Radial Oncol Biol Phys 1981;7:581-585.

368. McKenna R, McMurtrey M. Larson D, Mountain C. A perspective on chest wall resection In patfents with breast cancer. Ann Thorac Surg 1984;38:482-487.


369. McCormick P."Bains M, Burt M. Local recurrent mammary carcinoma failing mt.iltimodality therapy: A solution. Arch Surg 1989:124:158-161.

370. Chen K, Montague E, Oswald M. Results of irradiation in the treatment oflocoregional breast cancer recurrence. Cancer 1985;56: 1269-1273.

371. Fisher E, Sass R, Fisher B. Pathologic findings from the National Surgical Adjuvant Breast Project JI. Relation of local recurrence to multlcentricity. Cancer 1986;57: 1717-1724.

372. Fowble B, Solin L, Schultz D. Breast recurrence following conservative surgery and radiation: Patterns of failure, prognosis, and pathologic findings from mastectomy specimens with implications for treatment. Int J Radiat Oncol Biol Phys 1990;.19: 833-842.

373. Recht A, Silver B, Schnltt S, et al. Breast relapse following primary radiation therapy for early breast cancer. I. Classification, frequency and salvage. Int J Radiat Oncol Biol Phys 1985;11:1271-1276.

374. Kurtz J, Amalric R, Brandone M. Local recurrence after breast conserving surgery and radiotherapy. Frequency, time course and prognosis. Cancer 1989;63:1912-1917.

375. Solin L, Fowble B, Schultz D. The detection of local recurrence after definitive irradiation for early stage carcinoma of the breast. Cancer 1990;65:2497-2502.

376. Stamper P, Recht A, BerenbergA, et al. Mammo graphic detection of recurrent cancer in the irradiated breast. AIR 1987;148:39-43.

377. Haffty B, Goldberg N, Fischer D. Conservative surgery and radiation therapy in breast carcinoma: Local recurrence and prognostic lmplkations. Int J Radiat Oncol Biol Phys 1989;17:727-732.

378. Kurtz J, Amalric R, Brandone H. Results of wide excision for mammary recurrence after breast conserving therapy. Cancer 1988;61;1969-1972.

379. St om per P, Connolly J, Meyer J, Hanis J. Clinically occult ductal carcinoma in situ detected with mammography: Analysis of 100 cases with radiologic-pathologic correlations. Radiology 1989;172:235-241.

380. Rosa! J. Borderline epithelial lesions of the breast. Am J Stirg Pathol 1991;15:209-221.

381. Andersen J, Nielsen M. Christensen L. New aspects of the natural history of in situ and invasive carcinoma In the female breast: Results from autopsy Investigations. Verh Dtsch Ges Pathol 1985;69:88-95.

382. Wetzels R, Holland R, van Haelst U, Lane E, Leigh I, Ramaekers F. Detection of basement membrane components and basal cell keratin 14 in noninvasive and invasive carcinomas of the breast. Am J Pathol l 989;134:571-579.

383. Meyer J. Cell kinetics ofhls~ologic variants of in situ breast carcinoma. Breast Cancer Res Treat 1986:7:171-180.

384. van de Vljver M, Peterse J, Moo! W, et al. Neu protein overexpresslon In breast cancer: Association with comedo-type ductal carcinoma in situ and limited prognostic value In stage II breast cancer. N Engl J Med 1988;319:1239-1245.

385. Lagios M, Margolin F, Westdahl P, Rose M. Mammographlcally detected duct carcinoma in situ: Frequency of local recurrence following tylectomy and prognostic effect of nuclear grade on local recurrence. Cancer 1989:63:618-624.

386. Silverstein M, Waisman J, Gierson E, Colburn W, Giimagaml P, Lewinsky B. Radiation therapy for lntraductal carcinoma: Is it an equal alternative? Arch Surg 1991;126: 424-427.

387, Page D, Dupont W, Rogers L, Landenberger M. lntraductal carcinoma of the breast: Follow-up after biopsy only. Cancer 1982;49:751-758.

388. Rosen P, Braun D, Kinne D. The clinical significance of pre-invasive breast carcinoma. Cancer 1980;46:919-925.

389. Alpers C, Wellings S. The prevalence of carcinoma in situ In normal and cancerassociated breast. Hum Pathol 1985;16:796-807.

390. Bartow S, Pathak D, Black W, Key C, Teaf S. Prevalence of benign, atypical, and malignant breast lesiOns in populations at different risk for breast cancer. Cancer 1987;60:2751-2760.

391. Aasmundstad T, Haugen O. Recurrence-free survival in patients with lntraductal carcinomas grouped according to combinations of DNA ploldy and histopathological classification. Eur J Cancer 1990;26:956-959.

392. Kinne D, Petrek J, Osborne M, Fracchia A, De Palo A, Rosen PP. Breast carcinoma in situ. Arch Surg 1989: 124:33-36.

393. Carpenter R, Boulter PS, Cooke T, Gibbs MN. Managehient of screen detected ductal carcinoma in situ of the female breast. Br J Surg 1989:76:564-567.

394. Amesson L, Sm eds S, Fagerberg G. Follow-up of two treatment modalities for ductal cancer In situ of the breast. Br J Surg 1989;76:672-675.

395. Lagios M. Ductal carcinoma in situ: Pathology and treatment. Surg Clln North Am 1990;70:853-871.

396. Fisher E, Leeming R, Anderson S, Redmond C, Fisher B. Conservative management oflntraductal carcinoma (DCIS) of the breast. J Surg Oncol 1991;47:139-147.

397. Graham M, Lakhani S, Gazet J. Breast conserving surgery In the management of hi situ breast carcinoma. Eur J Surg Oncol 1991;17:258-264.

398. Gallagher W, Koerner F. Wood W. Treatment of intraductal carcinoma with limited surgery: Long-term follow-up. J Clln Oncol 1989;7:376-380.

399. McCormick B. Rosen P, Kinne D, Cox L, Yahaiom J. Ductal carcinoma In situ of the breast: Does conservation surgery and radiotherapy provide acceptable local control? Int J Radiat Oncol Biol Phys {Abstract] 1990;19:132.

400. KurtzJ, Jacquemier J, TorhorstJ, et al. Conservation therapy for breast cancers other than infiltrating ductal carcinoma. Cancer 1989:63:1630-1635.

401. Solin L, Fowble B. Schultz D, et al. Definitive irradiation for intraductal carcinoma of the breast. Int J Radlat Oncol Biol Phys 1990;19:843-850.

402. Bornstein B, Recht A, Connolly J, Hanis J. Results of treating ductal carcinoma in situ of the breast with conservative surgery and radiatioil therapy. Cancer 1991;67: 7-13.

References 1329

403. Stotter AT, McNeese M, Oswald MJ, Ames FC, Romsdahl MM. The role of limited surgery with irradiation in .primary treatment of ductal in situ breast cancer. Int J Radlat Oncol Biol Phys 1990;18:283-287.

~04. Fourquet A, Zafranl B, Campana F, Durand J, Vi1coq J. Breast conserving treatment of ductal carcinoma in situ. Semln Radial Oncol 1992;2: 116-124.

405. Van Dongen J, Holland R, Peterse J, et al. Ductal carcinoma in situ of the breast: Second EORTC consensus meeting. Eur J Cancer 1992;28(2-3):626-629.

405a. Roland Holland, personal communication. 406. Foote F, Stewart F. Lobular carcinoma in situ: A rare form of mammary cancer. Am

J Pathol 1941;17:491-496. 407. Schwartz G, Feig S, Rosenberg A. Staging and treatment of clinically occult breast

cancer. Cancer 1984;53:1379. 408. Wheeler J, Enterline H, Roseman J. Lobular carcinoma in situ of the breast. Long

term follow-up. Cancer 1974;34:554-563. 409. Frykberg E, Santiago F, Betsill W, O'Brien P. Lobular carcinoma in situ of the breast.

Surg Gynecol Obstet 1987;164:285-301. 410. Haagensen C, Lane N, Lattes R, Bodian C. Lobular neoplasla of the breast. Cancer

1978;421:737-769. 411. Lemanne D, Simone M, Martino S, Swanson M. Breast carcinoma ln situ: Greater

rise in ductal carcinoma in situ vs lobular carcinoma in situ. Proc Am Soc Clin Oncol 1991;10:45.

412. Pope T, Fechner R. Wilhem M. Lobular carcinoma in situ of the breast: Mammographic features. Radiology i988;!68:63-66.

413. Silverstein M, Gamagami P, Colburn W. Nonpalpable breast lesions: Diagnoses with slightly overpenetrated screen film mammography and hook wire-directed biopsy in 1014 cases. Radiology 1989;171:632-638.

414. McDlvitt R, Hutter R, Foote F, Stewart F. In stw lobular carcinoma. A prospective followup study indicating cumulative patient r1sks. JAMA 1967;201:82-86.

· 415. Rosen P. Lieberman P, Braun D. Lobular carcinoma In situ of the breast. Am J Surg · Pathol 1978;2:225-251.

416. Andersen J. Lobular carcinoma In situ of the breast .. An approach to rational treatment. Cancer 1977;39:2597-2602.

417. Andersen J. Lobular carcinomaln situ. A long-term follow-up In 52 cases. Acta Pathol Microblol Scand 1974;82:519-533.

418. Haagensen C; Bodian C, Haagensen D. Breast carcinoma, risk and detection. Philadelphia: WB Saunders, 1981:238.

419. Fisher B, Fisher ER. The interrelationship of hematogenous and lymphatic tumor cell dissemination. Surg Gynecol Obstet 1966;122:791-798.

420. Martin DS, Fugman RA. A role of chemotherapy as an adjuvant to surgery. Cancer Res 1957;17:1098-1101.

421. Sldpper HE, Schabel FMJ. Tumor stem cell heterogeneity: Implication with respect to the classification of cancers by chemotherapeutic effect. Cancer Treat Rep 1984;68: 43-61.

422. Sch~bel FM, Griswold DP, Corbett TH. Increasing the therapeutic response rates to anticancer drugs by applying the basic principles of pharmacology. Cancer 1984;54: 1160-1167.

423. Bonadonna G, Valagussa P. The contribution of medicine to the primary treatment of breast cancer. Cancer Res 1988:48:2314-2324.

424. Pritchard Kl. Current status of adjuvant endocrine therapy for resectable breast cancer. Semin Oncol 1987;14:23-33.

425. Fisher B, Carbone P, Economou SG. L-Phenylalanlne mustard (L-PAM) in the management of primary breast cancer: A report of early findings. N EnglJ Med 1975:292: 117-122.

426. Bonadonna G, Brusamollno E, Valagussa P. Combination chemotherapy as an arljuvant treatment In operable breast cancer. N Engl J Med 1976;294:405-410.

427. Bonadonna G, Valaguss AP. Rossi A. Ten-year results with CMF-based adjuvant chemotherapy in resectable breast cancer. Breast Cancer Res Treat 1985;5:95-115.

428. Bonadonna G, Valagussa P. Dose-response effect ofadjuvant chemotherapy In breast cancer. N Engl J Med 1981;304:10-15.

429. Shapiro CL, Henderson JC, Gelman RS. A randomized trlal of 15 vs. 30 weeks of adjuvant chemotherapy in high risk b~ast cancer patients: Results after a median follow-up of9.l years. Proc Am Soc ¢11n0ncol [AbstractJ 1991;10:44.

430. Fisher B, Brown AM, Dimitrov NV, et al. Two months of doxorublcln-cyclophosphamide with and without Interval reinductlon therapy co.mpared with 6 months of cyclophosphamlde, methotrexate and fluorouracil In positive-node breast cancer patients with tamoxifen nonresponsive tumor: Results from the National Surgical Adjuvant Breast and Bowei Project B-15. J Clin Oncol 1990;8:2483-2496.

431. Molltemi A, Bonadonna G, Valagussa P. Cyclophosphamide, methotrexate, and fluorouracll with and withoutdoxorubicin in the adjuvant treatment ofresectable breast cancer with one to three positive axillary nodes. J Clin Oncol 1991:9:1124-1130.

432. Bonadonna G, Monfardini S, De Lena M. Phase I and preliminary phase II evaluation of Adrlamycln (NSC-123127). Cancer Res 1970;30:2572-2582.

433. Buzdar AU, Hortobagyl GN, Kau SW. Doxorubicin-containing adjuvant therapy for patients with stage II breast cancer: M.D. Anderson Cancer Center experience. In: Salmon SE, ed. Adjuvant tlierapy of cancer VI. Philadelphia: WB Saunders, 1990: 210-215. .

434. Dalton W, Brooks R, Jones S. Breast cancer adjuvant therapy trials at Arizona Cancer Center using Adriamycin and cyclophosphamide. In: Salmon SE, ed. Adjuvant therapy of cancer V. Philadelphia: Grune & Stratton, 1987:263-269.

435. Fisher B, Redmond C. Wicherham DL. Doxorubicin-containlng regimens for the treatment of stage II breast cancer: The National Surgical Adjuvant Breast and Bowel Project experience. J Clln Oncol 1989:7:572-582.



436. The Ludwig Breast Cancer Study Group. Combination chemotherapy for node-positive breast cancer: Inadequacy of a single perioperative cycle. N Engl J Med 1988;319: 677-683.

437. The Ludwig Breast Cancer Study Group. Prolonged disease-free survival after one course of perioperative adjuvant chemotherapy for node-negative breast cancer. N Engl J Med 1989;320:491-496.

438. Sertoli M, Pronzato P, Rubagottl A. A randomized study ofperioperative chemotherapy in primary breast cancer. In: Salmon SE, ed. Adjuvant therapy of cancer VI. Phila· delphla: WB Saunders, 1990:196-203.

439. Frei EL, Canellas GP. Dose: A critical factor in cancer chemotherapy. Am J Med 1980;69:585-593.

440. H:rynluk W, Bush H·. The importance of dose Intensity in chemotherapy of metastatic breast cancer. J Clin Oncol 1984;2:1281-1288.

441. Hryniuk WM, Bonadonna G, Valagussa P. The effect of dose intensity in adjuvant chemotherapy. In: Salmon SE, ed. Adjuvanc therapy of cancer V. Philadelphia: Grune & Stratton, 1987:13-23.

442. Henderson IC, Hayes DF, Gelman R. Dose-response in the treatment of breast cancer: A critical review. J Clin Oncol 1986;6:1501-1515.

443. Tannock IF, Boyd NF. De Boer G. A randomized trial of two dose levels of cyclophosphanllde, methotrexate, and fluorouracll chemotherapy for patients with metastatic breast cancer. J Clin Oncol 1986;6:1377-1387.

444. Engelsman E, Klijn JCM, Rubens RD. "Classical" CMF versus a 3·weekly intravenous CMF schedule In postmenopausal patients with advanced breast cancer. An EORTC Breast Cancer Co-operative Group phase III trail (10808). Eur J Cancer 1991;27: 966-970.

445. Abel off M, Beverdige RA, Donehower RC. Sixteen-week dose-intense chemotherapy in the adjuvant treatment of breast cancer. JNCI 1990;82:570-574.

446. Peters WP. High-close chemotherapy and autologou~ bone marrow support for breast cancer. In: DeVita VT Jr, Hellman S, Rosenberg SA, eds. Important advances in

) oncology, 1991. Philadelphia: JB Lippincott, 1991:135-156, 4,47. Gianni AM, Bregnl M, Siena S. Recombinant human granulocyte-macrophage colony

stimulating factor reduces hematologic toxicity and widens clinical applicability of high-close cyclophospham!de treatment In breast cancer and non-Hodgkin's lymphoma. J Clin Oncol 1990;8:768-778.

448. Gianni AM, Siena S. Bregnl M. Growth factor-supported high-close sequential adjuvant chemotherapy ln breast cancer wlth ::>:10 positive nodes. Proc Am Soc Clin Oncol (Abstrac1] 1992;11 (in press).

449. Early Breast Cancer Triallsts' Collaborative Group. Effects of adjuvant tamoxifen and of cytotoxic 1herapy on mortality in early breast cancer. An overview of 61 ran· domized !rials among 28,896 women. N Engl J Med 1988;319:1681-1692,

450. Fisher B, Sherman B, Roakcette H. L·Phenylalanlne mustard (L-PAM) in the management ofpremenopausal patients with primary breast cancer. Lack of association of disease-free survival with depression of ovarian function. Cancer 1979:44:847-857.

451. Bonadonna G, Valagussa P, De Palo G. The results of adjuvant chemotherapy are predominantly caused by the hormonal changes such therapy induces. Opposed. In: Van Scoy·Mosher MB, ed. Medical oncology. Controversies In cancer treatment.Boston: GK Hall, 1981:100-109, 112-115.

452. Valagussa P, Brambllla C, Zambetti M. Salvage treatments in relapsing resectable breast cancer. Recent Results Cancer Res l 989;115:69-76.

453. Padmanabhan N, Howell A, Rubens RD. Mechanism of action of adjuvant chemo· therapy in early breast cancer. Lancet 1986:2:411-414.

454. Valagussa P, Bonadonna G. Mechanisrri of action of adjuvant chemotherapy in early breast cancer. Letter to the editor. Lancet 1986;2: 1035-1036,

455. Bianco AR, De1 Mastro L, Gallo C. Prognostic role of amenorrhea induced by adjuvant chemotherapy in premenopausal patients wlth early bre'ast cancer. Br J Cancer 1991;63:799-803.

456. Goldhlrsch A, Gelber RD, Castiglione M. The magnitude of endocrine effects of ad· juvant chemotherapy for premenopausal breast cancer patients. Ann Oncol 1990;1: 183-188.

457. Gibson DFC, Jordan VC. Adjuvant antiestrogen therapy for breast cancer. Past, present, and future, Surg Clin North Am 1990;70:1103-1113.

458. Stewart HJ. Adjuvant endocrine therapy for operable breast cancer. Bull Cancer 1991;78:379-384.

459. Goldhirsch A, Valagussa P. Old and new trends in the adjuvant treatment of early breast Cancer. Ann Oncol (EdltorialJ l 991;2:320-322.

460. Meakin JW, Allt WEC, Beale FA. Ovarian irradiation and prednisone following surgery and radiotherapy for carcinoma oft he breast. Breast Cancer Res Treat 1983;3(suppl 1):45-48.

461. Baum M, Ebb S, Brooks M. Biological fall out from trials of adjuvant tamoxifen in early breast cancer. In: Salmon SE, ed. Adjuvant therapy of cancer Vi. Philadelphia: WB Saunders, 1990:269-274.

462. Breast Cancer Trials Committee. Adjuvant tamoxifen In the management of operable breast cancer: The Scottish trial. Lancet 1987;2:171-175.

463. Rose C, Thorpe SM, Anderson KW. Beneficial effect of adjuvant tamoxifen therapy In primary breast cancer patients with high estrogen receptor values. Lancet 1985;1: 16-19.

464. Rutqvist LE, Cedermark B, Glas U. The Stockholm trial on adjuvant tamoxifen in early breast cancer. Correlation between estrogen receptor level and trea1ment effect. Breast Cancer Res Treat 1987;10:255-266.

465. Fisher B, Brown A. Wolmark N. Prolonging tamoxifen therapy for primary breast cancer. Findings from the National Surgical Adjuvant Breast and Bowel Project clinical trial. Ann Intern Med 1987;106:694-754.

466. Tormey DC, Rasmussen P, Jordan VC. Long-term adjuvant trutloxlfen study: Clinical update. Breast Cancer Res Treat 1987;9:157-158.

467. Fisher B, Redmond C, Brown A. Adjuvant chemotherapy with and without 1amoxifen in the treatment of primary breast cancer: 5-year results from the NSABP trial. J Clin Oncol 1986;4:459-471.

468. Tormey DC, Gray R, Taylor SGI. Postoperative chemotherapy and chemohormonal therapy in women with node-positive breast cancer. NCI Monogr 1986;1:75-80.

469. Ingle JN, Everson LK, Wieand HS. Randomized trial of observation versus adjuvant 1herapy w:lth cyclophosphamide, fl.uorouracil, prednlsone with or without tamoxlfen follow:lng mastectomy ln postmenopausal women w:lth node-positive breast cancer. J Clin Oncol 1988;6:1388-1396.

470. Fisher B. Redmond C, Legault-Poisson S. Postoperative chemotherapy and tamoxifen compared with tamoxlfen alone In the treatment of positive-node breast cancer patients aged 50 years and older with tumors responsive to tamoxifen: Results from !he National Surgical Adjuvant Breast and Bowel Project B-16. J Clin Oncol 1990;8:1005-1018.

471. Pearson OH. Hubay CA, Go~don NH. Endocrine versus endocrine plus five-drug che· motherapy in postmenopausal women with stage II estrogen receptor-positive breast cancer. Cancer 1989;64:1819-1823.

472. Goldhirsch A. Gelber RD. Adjuvant chemo-endocrine therapy or endocrine therapy alone for postmenopausal patients: Ludwig studies Ill and IV. Recent Results Cancer Res 1989;115:153-162,

473. Osborne CK, Kitten L, Arteaga CL Antago!Usm of chemotherapy-induced cytotoxiclty for human breast cancer cells by anti estrogens. J Clin On col 1989;7:710-717.

474. Herxlerson LC, Hayes DF, Parker LM. Adjuvant systemic therapy for patients with node-negative tumors. Cancer 1990;65:2132-2147.

475. Zambetti M, Bonadonna G, Valagussa P. CMF for node-negative and estrogen receptornegative breast cancer. JNCI Monogr 1992;11:77-83.

476. Mansour EG, Eudey L, Shatila AH. Adjuvant therapy in node-negative breast cancer. ls it necessary for all patients? An Intergroup study. In: Salmon SE, ed. Adjuvant therapy of cancer VI. Philadelphia: WB Saunders, 1990:174-189.

477. McGuire WL, Clark GM. Prognostic factors and treatmentdeclslons in axillary-nodenegative breast cancer. N Engl J Med 1992;326:1756-1761.

478. Bonadonna G, Valagussa P, Brambilla C. Adjuvant and neoadjuvant treatment of breast cancer with chemotherapy and/or endcicrine therapy. Semln Oncol 1991;18:515-524.

479. Early Breast Cancer Trialists' Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic or Immune therapy: 133 randomised trials involving 31 ,000 recurrences and 24,000 deaths among 75,000 women. Lancet l 992;339:1-15, 71-85.

480. Levine MN, Gent M, Hirsh J. The thrombogenlc effects of anticancer drug therapy ln women with stage II breast cancer. N Engl J Med 1988;318:404-407.

481. Weiss RB, Tormey DC, HollandJF. Venous thrombosis during multlmodal treatment of primary breast carcinoma. Cancer Treat Rep 1981;65:677-679.

482. Fisher B. Rockette H, Fisher ER. Leukemia in breast cancer patients followingadjuvant chemotherapy or postoperative therapy: The NSABP experience. J Clin Oncol 1985;3: 1640-1658.

483. Valagussa P, Tanclnl G, Bonadonna G. Second niallgnancies af1erCMF forresectable breast cancer. J Clin Oncol 1987;5:1138-1142.

484. Arrlagada R, Rutqvlst LE. Adjuvant chemotherapy in early breast cancer and incidence of new primary malignancies. Lancet 1991 ;338:535-8,

485. Jordan VC, 1991. Are we giving tamoxifen for too long? Eur J Cancer 1991;27:1072-1074.

486. Jordan VC. Tamoxifen and endometrlal cancer. Letter to the editor. Lancet 1988;2: 1019.

487. Andersson M, Storm HH, Mouridsen HT. Incidence of new primary cancers after tamoxifen therapy and radiotherapy for early breast cancer. JNCI 1991;83:1013- · 1017.

488. Fornander T, Rutqvlst LE, Cedermark B. Adjuvant tamoxifen In early breast cancer: Occurrence of new primary cancers. Lancet 1989;1 :117-120.

489. Nayfield SG, Karp JE, Ford LG. Potential role of tamoxifen In prevention of breast cancer. JNCI 1991;83:1450-1459.

490. Consensus Conference: Adjuvant chemotherapy for breast cancer. JAMA 1985;254: 3461-3463.

491. Veronesi U, Banfi A, Del Vecchio M. Comparison of Halsted mastectomy with quadrantectomy, axillary dlssectlon, and radiotherapy in early breast cancer. Long-term resultS. Eur J Cancer Clin Oncol 22:1085.

492. Bonadonna G, Veronesi U, Brambilla C. Primary chemotherapy to avoid mastectomy in tumors with diameters of three centimeters or more. JNCI 1990;82:1539-1545.

493. Mauriac L, Dumad M, Avril A. Effects of primary chemotherapy in conservative treatment of breast cancer patients wlth operable tumors larger than 3 cm. Results of a randomized trial in a single centre. Ann Oneal 1991 ;2:347-354.

494. Bates C. Clinical applications of serum tumor markers. Ann Intern Med 1991;115: 623-638.

495. Osborne CK, Yochnowitz MG, Knight WA. The value of estrogen and progesterone receptors in the treatment of breast cancer. Cancer 1980;46:2884-2888.

496. Bezwuoda WR. Esser JD, Kansey R. The value of estrogen and progesterone receptor detennlnation in advanced breast cancer: Estrogen receptor level but not progesterone receptor level correlates w:lth response to 1amoxifen, Cancer 1991 ;68:867-872.

497. J0rdan VC. Blochemlcal pharmacology ofantiestrogen action. Pharmacol Rev 1984;36: 245-276.

498. Furr BJA, Jordan VC. The pharmacology and clinical uses of tamoxlfen. Pharma col Ther 1984;35:127-205.


r I !

499. Sunderland MD, Osborne CK. Tamoxifen in premenopausal patients with metastatic breast cancer: A review. J Clin Oncoi I991;9:1283-1297.

500. OsbOme CK, Boldt DH, Clark GM. Effects oftamoxlfen on human breast cancer cell cycle kinetics: Accumulation of cells In early G1 phase. Cancer Res 1983;43:3583-3585.

501. Osborne CK. Coronado E, Allred DC. Acquired tamoxifen resistance: Correlation with reduced breast tumor lev&ls of tamoxifen and isomerizatlon of trans-4-hydroxytamoxifen. JNCI 1991:83:1477-1482.

502. Jordan VC, Fritz NF, Langan-Fahey S. Alteration of endocrine parameters in premenopausal women with breast cancer during long-term adjuvant therapy with tamoxlfen as single agent. JNCI 1991;83:1488-1491.

503. Ingle JN, Krook JE, Green SJ. Randomized trial of bilateral oophorectomy versus tamoxifen In premenopausal women with metastatic breast cancer. J Clin Oncol 1986;4:178-185.

504. Buchanan RB, Blarney RW, Durrant KR. A randomized comparison oftamoxifen with surgical oophorectomy in premenopausal patients with. advanced breast cancer. J Clin Oncol 1986;4:1326-1330.

505. Mourldsen H, PalshofT, Patterson J. Tamoxifen in advanced breast cancer. Cancer Treat Rev 1978;5:131-141.

506. Ashford AR, Donev HA, Hamilton RW. Reversible ocular toxicity related co camoxifen therapy. Cancer 1988;61:33-35.

507. Haller DG, Glick JH. Progestatlonal agents in advanced breast cancer: An overview. Semln Oncol 1986;13:2-8.

508. Allegra JC, Kiefer SM. Mechanisms of action of progestatlonal agents. Semin Oncol 1985;12:3-5.

509. Pannutl F, Marton! A, Di Marco AR. Prospective randomized clinical trial of two different high dosages of medroxyprogesterone acetate (MPA) in the treatment of metastatic breast cancer. Eur J Cancer 1979;15:593-601.

510. Cavalli F, Goldhirsch A. Jung! F. Randomized trial of low-versus high-dose medroxyprogesterone acetate in the induction treatment of postmenopausal patients with breast cancer. J Clin Oncol 1978;2:414-419.

511. Davila E, Vogel CL, East D. Clinical trial of high dose oral medroxyprogesterone acetate in the treatment of breast cancer and review of the literature. Cancer 1988;6 l: 2161-2167.

512. Gregoey EJ. Cohen SC, Oines DW. Megestrol acetate therapy for advanced breast cancer. J Clin Oncol 1985;3:155-160.

513. Muss HB, Wells HB, Paschold EH, et al. Megestrolacetateversus tamoxifen in advanced breast cancer: 5-year analysis-a phase Ill trial of the Pledmont Oncology Association. J Clin Oncol 1988;6:1098-1106.

514. Parnes HL. Abrams.JS, Tchekmedyian S. A phase I/II study of high-dose megestrol acetate Jn the treatment of metastatic breast cancer. Breast Cancer Res 1991 ;18: 171-177.

515. Muss HB, Well HB, Paschold EH. Megestrol acetate versus tamoxifen In advanced breast cancer: 5-year analysis-a phase Ill trial of the Piedmont Oncology Association. J Clln Oncol 1985;12:40-42.

516. Sancen RJ. Henderson IC. A comprehensive guide to the therapeutic use or amlnoglutethimide. Berlin: Karger, 1982.

517. Murray R, Pitt P. Low-dose aminoglutethlmlde without steroid replacement in the postemenopausal women with advanced breast cancer. Eur J Cancer Clin Oneal 1985;21:19-22.

518. Harris AL, Cantwell BMJ, Sainsbuey JR. Low dose amlnoglutethlmlde (125 mg twice daily) with hydrocortisone for the treatment of advanced postemnopausa1 breast cancer. Breast Cancer Res Treat 1986;7: 1-44.

519. Bruning PF, Bonfrer JMG, Hart AAM. Low dose aminoglutethimide without hydro· cortisone for the treatment of advanced postmenopausal breast cancer. Eur J Cancer Clin Oncol 1989;25:369-376.

520. Sancen RJ, Worgul TJ, Samojlik E. A randomized trial comparing surgical adrenalectomy with amlnoglutethimlde plus hydrocortisone in women with advanced breast cancer. N EnglJ Med 1981;305:545-551.

521. Gross PE, Powles TJ, Dowsett M. Treatment of advanced postmenopausal breast cancer with an aromawe Inhibitor, 4-hydroxyandrostenedione. Phase II report. Cancer Res 1986;46:823-826,

522. Dowsett M, Metha H, Smith IE. Endocrine pharmacology of the aromatase Inhibitor 4·hydroxyandrostenedione ( 4·HA) in orally treated breast cancer. Breast Cancer Res Treat 1989;14:151-155.

523. Lipton A, Harvey HA, Demers LM. A phase I trlal ofCGS I6949A: A new aromatase inhibitor. Cancer 1990:65:1279-1285.

524. Stein RC, Dowsett M, Davenport J. Preliminaey study of the treatment of advanced breast cancer In postmenopausal women with the aromatase Inhibitor CGS 16949A. Cancer Res 1990:50:1381-1384,

525. Ingle JN, KrookJE, Sch aid DJ. Evaluation oftrilostane plus hydroconisone in women with metastatic breast cancer and prior honnonal therapy exposure. Am J Clin-Oncol 1990;13:93-97.

526. Santen RJ, Manni A, Harvey H. Gonadotropln releasing hormone (GnRH) analogs for the treatment of breast cancer and prostatic carcinoma. Breast Cancer Res Treat 1986;7:129-145.

527. Harvey HA, Lipton A, Max DT. Medical castration produced by the GnRH analogue leuprolide to treat metastatic breast cancer. J Clin Oncol 1985;3:1068-1072.

528. Jackson IM, Matthews MJ, Diver JMJ. LH-RH analogues in the treatment of cancer. Cancer Treat Rev 1989;16:161-175.

529. Dixon AR, Jackson L, Nicholson RI. The use ofgoserelin (Zoladex) in premenopausal advanced breast cancer. Br J Cancer 1990;62:32-39.

References 1331

530. Stoll BA. Combination endocrine therapy in breast cancer. Eur J Cancer Clin On col 1985;21:413-416.

531. Powles TJ, Ford HT, Nash AG. Treatment of disseminated breast cancer with tamoxlfen, amlnoglutethimide, hydrocortisone, and danazol, used in combination or sequentially. Lancet 1984:1:1369-1373.

532. Henderson IC, Garber JE, Breitmeyer JB. Comprehensive management of dlsseminated breast cancer. Cancer 1990;66:1439-1448.

533. Jones RB, Holland JF. Bhardway S. A phase I-II study of intensive-dose Adraiamycin for advanced breast cancer. j Clin Oncol 1987;5:172-177.

534. Ahman DL, Schaid D, Bisel HF. The effect on survival of Initiating chemotherapy in advanced breast cancer: Polychemotherapy versus single drug. J Clin Oncol 1987;5: 1928-1932.

535. Garber JE, Henderson IC. The use of chemotherapy in metastatic breast cancer. Hematol Oncol Clin North Am 1989;3:807-821.

536. Tranum BL, McDonald B, Thigpen T. Adrlamycin coinbinations in advanced breast cancer: A Southwest Oncology Group study. Cancer 1982;49:835-839.

537. Brambllla C, De Lena M, Rossi A. Response and survival in advanced breast cancer after two non-cross resistant combinations. Br Med J 1976;1 :801-804.

538. Coates A, Gebski V, Bishop JF. Improving the quality of life during chemotherapy for advanced breast cancer: A comparison oflntennlttent and continuous treatment strategies. N EnglJ Med 1987;317:1490-1495. -

539. Muss HB, Case LD, Richards F. Interrupted versus continuous chemotherapy in patients with metastatic breast cancer. N Engl J Med 1991;325:1342-1348.

540. Henderson JC, Hayes DF, Come S. New agents and new medical treatments for advanced breast cancer. Semln Oncol 1987;14:34-64.

541. Brambilla C, Valagussa P, Bonadonna G. Sequential combination chemotherapy in advanced breast cancer. Cancer Chemother Pharmacol 1978;1:35-39.

542. Lippman ME, Cassidy J, Wesley M. A randomized attempt to increase the efficacy of cytotoxic chemotherapy in metastatic breast cancer by honnonal synchronization. J Clln Oncoi 1984;2:28-36.

543. ScheithauerW, Zlelinskl C, Ludwig. Weekly low dose doxorublcin therapy in metastatic breast cancer resistant to previous hormonal and cytostatic treatment. Breast Cancer Res Treat 1985;6:89-93.

544. Konits P, Aisner J, Van Echo d. Mitomycin C and vinblastine chemotherapy in advanced breast can.cer. Cancer 1981;48:1295-1298.

545. Kardinal C, Perry M, Konuni A. Responses to chemotherapy or chemohormonal therapy in advanced breast cancer patients treated previously with adju"Jant che· motherapy. A subset analysis ofCALGB study 8081. Cancer 1988;61:415-419.

546. Brambilla C, Zambetti M, Ferrari L. Mitomycln and vinblastine in advanced refractot;' breast cancer. Tumor! 1989;75:141-144.

547. Buckner JC, Ingle JN, Everson LK. Results of salvage honnonal therapy and salvage chemotherapy in women failing adjuvant chemotherapy after mastectomy for breast cancer. Breast Cancer Res Treat 1989;13:135-142.

-548. Rowlnsky EK, Cazenae LA, Donehower RC. Taxol: A novel lnvestlgational antineo· plastic agent. JNCI 1990;82:1247-1259.

549. Brown T, Havlin K, Weiss G. A phase I trial of taxol given by a 6·hour intravenous infusion. J Clin Oncol 1991;9:1261-1267.

550. Doroshow JH, Leong L, Margolin K. Refractoey metastatic breast cancer: Salvage therapy with fluorouracll and high-dose continuous Infusion leucovorln calcium. J Clln Oncol 1989:7:439-444.

551. Swain SM, Lippman ME, Egan EF. 5-Fluorouracil and high-dose leucovorln in previously treated patients with metastatic breast cancer. J Clin Oncol 1989;7:890-899.

552. Hainsworth JD, Andrews MB, Johnston DH. Mltoxantrorie, fluorouracil, and high· dose leucovorin: An effective, well·tolerated regimen for metastatic breast cancer. J Clin Oncol 1991;9:1731-1735.

553. Jones SE, Menne! RG, Brookes B. Phase II study of mitoxantrone, leucovorln, and infuslonal fluorouracll for treatment of metastatic breast cancer. J Clin Oncol 1991 ;9: 1736-1739. 1 -

554. Henderson IC, Alegra JC, Woodcock T. Randomized clinical trial comparfiig mltoxantrone with doxorublcin in previously treated patients with metastatic breast cancer. J Clin Oncoi 1989;7:560-571.

555. Jodrell DI, Smith IE, Mansi JL. A randomized comparative trial of mitozantronel methotrexate/mltomycin C (MMM) and cyclophosphamide/methotrexate/5FU (CMF) In the treatment of advanced breast cancer. Br J Cancer 1991;63:794-798.

556. Shpall EJ, Jones RB, Bast RCJ. 4-Hydroxyperoxycyclophosphamidepurging of breast from the mononuclear cell fraction of bone marrow In patients receiving high dose chemotherapy and autologous marrow support: A phase I trial. J Clin Oncol 1991;9: 85-93 .•

557. Kennedy MJ, Beveridge RA, Rowley SD. High-dose chemotherapy with reinfusion of purged autologous bone marrow following dose-Intense Induction as Initial therapy for metastatic breast cancer. JNCI 1991 ;83:920-926.

558. Williams SF, Mick R, Desser R. High-dose consolidation therapy with autologous stem cell rescue In stage IV breast cancer. J Clin Oncol 1989;7:1824-1830.

559. Borden EC, Sondel PM. Lymphokines and cytoklnes as cancer treatment: lmmuno· therapy realized. Cancer 1990;65:800-814.

560. Muggla F. Therapeutic approaches to management of metabolic complications of cancer. Semin Oncol 1991;18:1-31.

561. Warrell RPJ, Murphy WK, Shulman P. A randomized double-blind study of gallium nitrate compared with etidronate for acute control of cancer-related hypercalcemfa. J Clln Oncol 1991;9:1467-1475.

562. Seidman H, Mushinskl M, Gelb SEA. Probabilities of eventually developing or dying of cancer: United States, 1985. CA 1985;35:36-56.



563. Jones J, Ribeiro G. Mortallty patterns over 34 years of breast cancer patients In a clinical trial of post-operative radiotherapy. CUn Radiol 1989;40:24-28.

564. Host H, Brennhoud I, Loeb M. Post-operative radiotherapy In breast cancer-Longterm results from the Oslo'study. Int J Radlat Oncol Biol Phys 1986;12:727-732.

565. Rutqvist L, Cedermark B, Glas U, et al Radiotherapy, chemotherapy, iind tamox1fen as adjuncts to surgery" ln early breast cancer: A summary oftfiree randomized trials. IntJ Radial Oncol Biol Phys 1989:16:629-639.

566. Veiez-GarciaE, Carpenter JT, Moore M, et al. Comparison ofadjtiVantchemotherapy versus loco-regional radiotherapy followed by adjuvant chemotherapy in breast ~cer patients with 4 or more positive axillary nodes: A So1Jtheastem C.ancer ~tudy Group study. Eur J Cancer 1992;28A:1833-1837. '

567. Overgaard M, Christensen J, Johansen H, et al. EvalUation of radiotherapy In highrisk breast cancer patients. Int J Radlat Oncol Biol Phys 1990; 19: 1121-1124.

568. Ahmann D, O'Fallon J, Scanlon PW, et al. A preliminary assessment of factors associated with recurrent disease In a surgical adjuvant clinical trial for patients with

breast cancer with special emphasis on aggressiveness of therapy. Am J Clin Oncol 1982:5:371-381.

569. Muss H, Cooper R, Brockschnrldt J, et al. A randomized trial of adjuvant chemotherapy (CT) without radiation therapy (RT) for stage II breast cancer: 11-year follow-up of Piedmont Oncology Association protocol 74176. Breast Cancer Res Treat 1989;14: 185.

570. McArdle C, Crawfoid D, Dykes EH, et al. Adjuvant radiotherapy and chemotherapy In breast cancer. Br J Surg 1980;73:264-266.

571. Owen H, Dockerty M, Gray H. Occult carcinoma of the breast. Surg Gynecol Obstet 1954;98:302-307.

572. Price P, Sinnett HD, Gusterson B, Walsh G, A'Hem RP, McKlnnaJA. Duct carcinoma In situ: Predictors of local recurrence and progression in patients treated by surgery alone. Br J Cancer 1990;61:869-872. ·

573. Hayward.JL, Carbone PP, Heuson JC. Assessment of response to therapy ln advanced breast cancer. A project of the Programme on Clinical Oncology of the International Union against Cancer, Geneva, Switzerland. Eur J Cancer 1977;13:89-94.


1 of 122 celltrion, inc., exhibit 1006 - microsoft · pdf file·cer principles &...

Documents