Bladder TumorsEDITORS

MARK SOLOWAY, ADRIENNE CARMACK, SAAD KHOURY

1st INTERNATIONAL CONSULTATION ON BLADDER TUMORS - HAWAII OCTOBER 3-7, 2004

CO-SPONSORED BY

SIU (International Society of Urology), ICUD (International Consultation on Urological Diseases),

EORTC Urology (European Organisation for Research and Treatment of Cancer)

Distributor : EDITIONS 2176, rue de la Pompe - 75016 Paris - FRANCEFax: +33 1 45 04 72 89 E-mail : editions21@wanadoo.fr

© Health Publication Ltd 2005All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or byany means, electronic, mechanical, photocopying, recording, or otherwise, without prior permission of the publisher.

Accurate indications, adverse reactions, and dosage schedules for drugs are provided in this book, but it is possible that they maychange. The reader is urged to review the package information data of the manufacturers of the medications mentioned.

The Publishers have made every effort to trace the copyright holders for borrowed material. If they have inadvertently overloo-ked any, they will be pleased to make the necessary arrangements at the first opportunity.

“The opinions expressed in this publication do not necessarily represent the official opinion of WHO”

ISBN 0-9546956-4-X

LAY-OUT: S.Taieb

2

3

INTRODUCTION: INTERNATIONAL CONSULTATION ON BLADDER TUMORS

The title of this text makes a statement. All blad-der tumors are not malignant and thus, the title isbladder tumors and not bladder cancer. This isappropriate. Tumors of the urinary bladder areheterogeneous. Not all neoplastic growths in thebladder are a threat to the host. Papilloma, papil-lary urothelial neoplasms of low malignant poten-tial and even most low-grade, non-invasive papil-lary tumors sometimes termed low grade transitio-nal cell carcinoma do not behave as a malignancywith the potential for invasion or metastasis. Yet,almost 50% of our patients with bladder tumorshave one of these benign neoplasms and much ofour attention is directed at treating them.

Bladder tumors are indeed a major health problem.The prevalence of bladder tumors is high due inpart to their propensity to “recur” once resectedand the many etiologic factors related to theirdevelopment. Some of the causative agents areknown such as cigarette smoking and hair dyeswhile others have yet to be determined. It is like-ly that many of the chemicals in our environmentare concentrated in the urine and capable of cau-sing mutations in the bladder urothelium thus lea-ding to a bladder neoplasm. It has been stated thatthe cost of taking care of patients with bladdertumors is among the highest of all neoplasms.

This text details the final report of the consultationon bladder tumors and provides an extensiveresource for all those interested in this disease. Ithas been an honor and privilege to chair this Inter-national Consultation on Bladder Tumors. I wasconfident that the many experts throughout theworld would agree to share their expertise and par-ticipate in this major effort. I subdivided the sub-ject of bladder tumors into eleven areas. Eachcommittee had the task of collecting informationon their subject area and, after an extensivereview, arrive at a consensus or guidelines.Although there are relatively few randomizedtrials dealing with bladder tumors, we have produ-ced many guidelines whose foundation largelyrests on the experience of those who have studiedthis disease for many years. Guidelines may chan-ge as new techniques, knowledge and treatmentsevolve. Most of these, however, are based on an

understanding of the biology of bladder tumorsand will likely remain for many years.

I particularly want to express my deep apprecia-tion to all of the chairs, vice-chairs and committeemembers who unselfishly provided us with theirtime and energy to accomplish this goal. This wasa labor of love. Each of the chairs had the task ofsynthesizing the often diverse opinions of theircommittee members and reaching a consensuswhen possible.

I am particularly grateful to two individuals thatguided the project from its inception to comple-tion. Saad Khoury formulated the concept of aninternational consultation. I believe this is a wor-thy goal and although some may challenge indivi-dual guidelines, most would agree that a periodicreview of the literature with the discussion onappropriate management is a productive endeavor.Working with Mustafa Elhillali and the SIU, Saadprovided the challenge and the mechanism tocarry out the task. The second individual isAdrienne Carmack. While only a mid-level urolo-gy trainee, Adrienne is an accomplished editor,skilled communicator and most importantly, hassuperb time management skills. She was able toattend three international meetings, communicateregularly with the committees, and helped me editthe manuscripts while still taking part in a deman-ding residency program. Without email evenAdrienne could not have accomplished this task.

Many of you will not have the opportunity to readthe entire text but will use this volume as a resour-ce and select topics of interest. I will take thechairman’s prerogative of highlighting some of theguidelines and recommendations from the 11 com-mittees.

Epidemiology and Diagnosis

Bladder tumors are common. It is the most com-mon malignancy in some countries (Egypt) andranks very high in many others e.g. Turkey, China,United States. Since we know some of the riskfactors, e.g. age, cigarette smoking, hair dyes,there is an opportunity for early detection. Unfor-tunately many of our patients are diagnosed at astage when even our optimal treatments are only

4

palliative. Since we have effective treatments forTa – T1 urothelial tumors, it only seems reaso-nable to devote some resources to identifyingbladder cancer at an early stage. This strategy hasmade an impact on some other tumors, e.g. breast,prostate, cervix. Importantly, the majority of blad-der tumors left untreated will eventually requiretherapy so the problem of overdetection does notseem pertinent.

The committee urges pathologists and urologiststo adopt the 2004 WHO/ISUP classification whichdivides grades into low or high eliminating theambiguous Grade 2 category.

Markers

The development of urine based tumor markersprovides an opportunity for early detection as wellas non-invasive surveillance following the initialdiagnosis of a bladder tumor. Of course, urolo-gists will not abandon the cystoscope which is thefoundation of the diagnosis and initial treatment.It is, however, invasive. Some of the monitoringcan be supplemented or substituted by a tumormarker. This is applicable particularly for thosewith low risk tumors. This chapter provides anextensive review of cytology and the various mar-kers.

Low grade, non-invasive tumors

There is sufficient evidence from prospective ran-domized trials to recommend the instillation ofintravesical chemotherapy within hours of resec-tion of low grade, Ta tumors provided that thebladder wall remains intact. Upper tract monito-ring is not necessary in these patients as the risk ofan upper tract urothelial tumor is very low. Officefulguration is felt to be an acceptable mode of the-rapy for small, low grade “recurrences”.

High grade Ta, carcinoma in situ, and T1 (2committees)

The term bladder cancer is appropriate for thesegrades and stages. Not only is the likelihood of asubsequent tumor following initial resection highbut there is a risk of progression to a higher stage.The clinician must assure removal of all of thetumor, accurate staging, and make an effort tominimize the likelihood of a subsequent tumor.Depending on certain prognostic factors such asmultifocality and stage, one might even consider

proceeding with a radical cystectomy in a subsetof these patients, e.g. multifocal high grade T1with CIS. The concept of a reTUR is a strongrecommendation in this group of patients giventhe critical role of staging in the decision makingprocess. In general, this should be performedwithin a few weeks of the initial endoscopic resec-tion of a high grade Ta or T1 bladder cancer. Onemust exclude muscle invasion since this wouldchange the treatment algorithm. Muscle must bepresent in the resection specimen.

BCG is the treatment of choice following resec-tion of a high-grade Ta, CIS or T1 urothelial carci-noma. It is yet to be established that maintenanceBCG is necessary, although the majority of thecommittee members felt that a minimum of one-year maintenance has been shown to reduce therecurrence rate if not the progression rate. Thethree and six month evaluation is critical. If a highgrade tumor recurs or persists within this timeframe, one should consider proceeding withoutundue delay to radical cystectomy.

The committee felt that following the initial resec-tion of a high grade Ta or T1 tumor, a single doseof post-operative intravesical chemotherapy isappropriate to decrease the chance of tumorimplantation. This does not alter the recommen-dation to use BCG.

Muscle invasive urothelial cancer

Radical cystectomy remains the optimal treatmentfor tumors which invade the muscularis propria ofthe bladder. There are occasional patients whomeet the criteria for bladder preservation, whichincludes a “complete transurethral resection” fol-lowed by combination chemotherapy and radia-tion therapy. Careful and persistent monitoring isnecessary. This approach requires the expertise ofthree disciplines (urology, radiation and medicaloncology). The presence of hydronephrosis isgenerally felt to exclude the role of bladder pre-servation.

The committee commented on two of the pro-blems related to this group of patients. A delayfrom diagnosis to cystectomy greater than 12weeks is an adverse prognostic factor. Manypatients who would ordinarily be candidates forcystectomy, are not having an opportunity to havethis therapy. This needs to be explored further.

5

Urinary diversion

Twenty years ago virtually all patients had anexternal appliance upon removal of the urinarybladder. With the introduction of a continent cuta-neous diversion and subsequently the develop-ment of an orthotopic neobladder. Fifty percentof patients now avoid an external appliance.Ileum is the bowel segment of choice for construc-tion of an orthotopic neobladder. Patients must bemotivated and understand the possibility of noc-turnal incontinence and the need for clean inter-mittent catheterization. The orthotopic neobladderas currently performed in many major medicalcenters has the same morbidity as other forms ofurinary diversion and patient acceptance has beenexcellent.

Radiation therapy

External beam radiation therapy remains a com-mon form of treatment for muscle-invasive blad-der cancer in many geographic areas and in parti-cular in patients older than 75. There are groupsof patients in which the success rate is low and thisincludes patients with ureteral obstruction and cli-nical stages T3-T4. Concurrent cisplatin shouldbe administered along with external beam radia-tion therapy unless there is a contraindication suchas impaired renal function. For the advanced cli-nical stages, multiagent systemic chemotherapyshould precede external beam radiation therapy.

Systemic chemotherapy

Cisplatin-based multiagent systemic chemothera-py for urothelial cancer rarely provides a cure inthose who have metastasis; however, manypatients have significant palliation. The currentmost active combinations consist of either MVACor gemcitabine plus cisplatin. Other combinationshave been explored and many are active.

Adverse risk factors for patients with clinicallylocalized muscle invasive bladder cancer includelymphovascular invasion, hydronephrosis, stageT3-T4 and prostatic stromal invasion. Suchpatients can be considered for neoadjuvant che-motherapy although prospective randomized trialsshow only a modest benefit with this approach.There is insufficient evidence to recommend rou-tine adjuvant chemotherapy for patients with pT2and pT3 bladder cancer.

Urothelial carcinoma of the prostate

It is important for the urologist to monitor theprostatic urethra, particularly in the presence of ahigh grade urothelial carcinoma of the bladder.Treatment for urothelial carcinoma of the prostatedepends on the stage: confined to the urothelium,involvement of the prostatic ducts or stromal inva-sion. The treatment for carcinoma confined to theurothelium is transurethral resection followed byBCG if the cancer is high grade. Once the tumorinvolves the stroma radical cystectomy is necessa-ry, assuming that the patient is a candidate formajor surgery. When the tumor involves only theprostatic ducts, some would proceed with trans-urethral resection plus BCG while others think thiswarrants proceeding to cystectomy.

Non-urothelial carcinoma of the bladder

Squamous cell carcinoma, adenocarcinoma andsmall cell carcinoma are uncommon but importantneoplasms of the urinary bladder. Radical cystec-tomy is the treatment of choice for squamous celland adenocarcinoma. Systemic chemotherapy is acritical component of the treatment for small cellcarcinoma of the bladder.

This has been only a few highlights from theextensive amount of material that follows in thisbook. Whenever possible, the authors haveattempted to provide guidelines and indicate thegrades of recommendation as identified in theOxford Center for Evidence-Based Medicine. Onbehalf of all of those who worked so diligently inthis project, we sincerely hope that the informationprovided will be used to help all of us optimize thecare of our patients with bladder tumors.

Mark S. Soloway, M.D.

Professor and Chairman

Department of Urology

University of Miami – Miller School of Medicine

Miami, Florida

Committee 1 Bladder Cancer: Epidemiology, Staging and Grading, and Diagnosis

Z. KIRKALI (TURKEY), T. CHAN (USA), M. MANOHARAN (USA), F. ALGABA (SPAIN), C. BUSCH (SWEDEN), L. CHENG (USA), L. KIEMENEY (NETHERLANDS), M. KRIEGMAIR (GERMANY), R. MONTIRONI (ITALY), W. MURPHY (USA), I. SESTERHENN (USA), M. TACHIBANA (JAPAN), J. WEIDER (USA)

Committee 2 Cytology And Tumor Markers:Tumor Markers Beyond Cytology

V.B LOKESHWAR (USA), T. HABUCHI (JAPAN), H.B. GROSSMAN (USA), W.M. MURPHY (USA), G.P. HEMSTREET, III (USA), S.H. HAUTMANN (GERMANY), A.V. BONO (ITALY), R.H. GETZENBERG (USA), P. GOEBELL (GERMANY), B.J. SCHMITZ-DRÄGER (GERMANY), M. MARBERGER (AUSTRIA), J. A. SCHALKEN (NETHERLANDS), E. MESSING (USA), Y. FRADET (CANADA), M.J. DROLLER (USA)

Committee 3 Low Grade, Ta (Noninvasive) Urothelial Carcinoma of the Bladder

W. OOSTERLINCK (BELGIUM), E. SOLSONA (SPAIN), H. AKAZA (JAPAN), C. BUSCH (SWEDEN), P.J. GOEBELL (GERMANY), P.-U. MALMSTRÖM (SWEDEN), H. ÖZEN (TURKEY), P. SVED (USA)

Committee 4 High Grade Ta Urothelial Carcinoma and Carcinoma in Situ of the Bladder

R. SYLVESTER (BELGIUM), A.V.D MEIJDEN (NETHERLANDS), J.A. WITJES (NETHERLANDS),G. JAKSE (GERMANY), N. NONOMURA (JAPAN), C. CHENG (SINGAPORE),A. TORRES (MEXICO), R. WATSON (AUSTRALIA), K.H. KURTH (NETHERLANDS)

Committee 5 T1 Urothelial Carcinoma of the Bladder

M.A.S JEWETT (CANADA), A.M. NIEDER (USA), M. BRAUSI (ITALY), D. LAMM (USA), M. O’DONNELL (USA), K. TOMITA (JAPAN), H. WOO (AUSTRALIA)

Committee 6 Muscle-Invasive Urothelial Carcinoma of the Bladder

S. B. MALKOWICZ (USA), H. VAN POPPEL (BELGIUM), G. MICKISCH (GERMANY),V. PANSADORO (ITALY), J. THÜROFF (GERMANY), M. SOLOWAY (USA), S. CHANG (USA), M. BENSON (USA), I. FUKUI (JAPAN)

Committee 7 Urinary DiversionR. E. HAUTMANN (GERMANY), H. ABOL-ENEIN (EGYPT), K. HAFEZ (USA), I. HARA (JAPAN), W. MANSSON (SWEDEN), R. D. MILLS (UK), J. D. MONTIE (USA), A. I. SAGALOWSKY (USA), J. P. STEIN (USA), A. STENZL (GERMANY), U.E. STUDER (SWITZERLAND), B.G. VOLKMER (GERMANY)

6

MEMBERS OF THE COMMITTEES (by Committee)

Committee 8 Urothelial Carcinoma of the Prostate

J. PALOU (SPAIN), J. BANIEL (ISRAEL),L. KLOTZ (CANADA), D. WOOD (USA), M. COOKSON (USA), S. LERNER (USA), S. HORIE (JAPAN), M. SCHOENBERG (USA), J. ANGULO (SPAIN), P. BASSI (ITALY)

Committee 9 Chemotherapy for Bladder Cancer: Treatment Guidelines for Neoadjuvant Chemotherapy, Bladder Preservation, Adjuvant Chemotherapy, andMetastatic Cancer

C. STERNBERG (ITALY), S.M. DONAT (USA), J. BELLMUNT (SPAIN), R.E. MILLIKAN (USA), W. STADLER (USA), P. DE MULDER (NETHERLANDS), A. SHERIF (SWEDEN), H. VON DER MAASE (DENMARK), T. TSUKAMOTO (JAPAN), M. SOLOWAY (USA)

Committee 10 Radiotherapy for Bladder Cancer

M. MILOSEVIC (CANADA), M. GOSPODAROWICZ (CANADA), A. ZIETMAN (USA),F. ABBAS (PAKISTAN), K. HAUSTERMANS (BELGIUM), L. MOONEN (NETHERLANDS), C. RÖDEL (GERMANY), M. SCHOENBERG (USA), W. SHIPLEY (USA)

Committee 11 Non-urothelial Cancer of the Bladder

H. ABOL-ENEIN (EGYPT), B.R. KAVA (USA),A.J.K. CARMACK (USA)

7

Summary of International Consultation on Urologic Disease Modified Oxford Center for Evidence-based Medicine Grading System for Guideline Recommendations 10

Committee 1 Bladder Cancer: Epidemiology, Staging and Grading, and Diagnosis 13Z. KIRKALI (TURKEY), T. CHAN (USA), M. MANOHARAN (USA), F. ALGABA (SPAIN), C. BUSCH (SWEDEN), L. CHENG (USA), L. KIEMENEY (NETHERLANDS), M. KRIEGMAIR (GERMANY), R. MONTIRONI (ITALY, W. MURPHY (USA),

I. SESTERHENN (USA) M. TACHIBANA (JAPAN), J. WEIDER (USA)

Committee 2 Cytology And Tumor Markers:Tumor Markers Beyond Cytology 65

V. B. LOKESHWAR (USA), T. HABUCHI (JAPAN), H. B. GROSSMAN (USA),W. M. MURPHY (USA), G. P. HEMSTREET, III (USA), S. H. HAUTMANN (GERMANY), A. V. BONO (ITALY), R. H. GETZENBERG (USA), P. GOEBELL (GERMANY),B. J. SCHMITZ-DRÄGER (GERMANY), M. MARBERGER (AUSTRIA), J. A. SCHALKEN (NETHERLANDS), E. MESSING (USA), Y. FRADET (CANADA),M. J. DROLLER (USA)

Committee 3 Low Grade, Ta (Noninvasive) Urothelial Carcinoma of the Bladder 139

W. OOSTERLINCK (BELGIUM), E. SOLSONA (SPAIN), H. AKAZA (JAPAN), C. BUSCH (SWEDEN), P. J. GOEBELL (GERMANY), P.-U. MALMSTRÖM (SWEDEN),H. ÖZEN (TURKEY), P. SVED (USA)

Committee 4 High Grade Ta Urothelial Carcinoma and Carcinoma in Situ of the Bladder 165

R. SYLVESTER (BELGIUM), A. V.D. MEIJDEN (NETHERLANDS), J.A. WITJES (NETHERLANDS),G. JAKSE (GERMANY), N. NONOMURA (JAPAN), C. CHENG (SINGAPORE),A. TORRES (MEXICO), R. WATSON (AUSTRALIA), K.H. KURTH (NETHERLANDS)

Committee 5 T1 Urothelial Carcinoma of the Bladder 189

M.A.S. JEWETT (CANADA), A.M. NIEDER (USA), M. BRAUSI (ITALY), D. LAMM (USA),M. O’DONNELL (USA), K. TOMITA (JAPAN), H. WOO (AUSTRALIA)

Committee 6 Muscle-Invasive Urothelial Carcinoma of the Bladder 219

S. B. MALKOWICZ (USA), H. VAN POPPEL (BELGIUM), G. MICKISCH (GERMANY),V. PANSADORO (ITALY), J. THÜROFF (GERMANY), M. SOLOWAY (USA),S. CHANG (USA), M. BENSON (USA), I. FUKUI (JAPAN)

8

CONTENTS

Committee 7 Urinary Diversion 239

R. E. HAUTMANN (GERMANY), H. ABOL-ENEIN (EGYPT), K. HAFEZ (USA),I. HARA (JAPAN), W. MANSSON (SWEDEN), R. D. MILLS (UK), J. D. MONTIE USA), A. I. SAGALOWSKY (USA), J. P. STEIN (USA), A. STENZL (GERMANY), U. E. STUDER (SWITZERLAND), B. G. VOLKMER (GERMANY)

Committee 8 Urothelial Carcinoma of the Prostate 309

J. PALOU (SPAIN), J. BANIEL (ISRAEL), L. KLOTZ (CANADA), D. WOOD (USA),M. COOKSON (USA), S. LERNER (USA), S. HORIE (JAPAN), M. SCHOENBERG (USA),J. ANGULO (SPAIN), P. BASSI (ITALY)

Committee 9 Chemotherapy for Bladder Cancer: Treatment Guidelines for Neoadjuvant Chemotherapy, Bladder Preservation, Adjuvant Chemotherapy, and Metastatic Cancer 335

C. STERNBERG (ITALY), S. M. DONAT (USA), J. BELLMUNT (SPAIN), R. E. MILLIKAN (USA), W. STADLER (USA), P. DE MULDER (NETHERLANDS), A. SHERIF (SWEDEN), H. VON DER MAASE (DENMARK), T. TSUKAMOTO (JAPAN), M. SOLOWAY (USA)

Committee 10 Radiotherapy for Bladder Cancer 359

M. MILOSEVIC (CANADA), M. GOSPODAROWICZ (CANADA), A. ZIETMAN (USA),F. ABBAS (PAKISTAN), K. HAUSTERMANS (BELGIUM), L. MOONEN (NETHERLANDS), C. RÖDEL (GERMANY), M. SCHOENBERG (USA), W. SHIPLEY (USA)

Committee 11 Non-urothelial Cancer of the Bladder 383

H. ABOL-ENEIN (EGYPT), B.R. KAVA (USA),A.J.K. CARMACK (USA)

9

10

Levels of Evidence

Level 1 Metanalysis of RCTs or a good quality RCT

Level 2 Low quality RCT or metanalysis of good quality prospective cohort studies

Level 3 Good quality retrospective case control studies or case series

Level 4 Expert opinion based on “first principles” or bench research, not on evidence

Grades of Recommendation

Grade A Usually consistent level 1 evidence

Grade B Consistent level 2 or 3 evidence or “majority evidence” from RCTs

Grade C Level 4 evidence, “majority evidence” from level 2/3 studies, expert opinion

Grade D No recommendation possible because of inadequate or conflicting evidence

*Adapted from Evidence-based medicine: Overview of the main steps for developing and grading guideline recommendations, by P Abrams, A Grant, and S Khoury, January 2004

Summary of the International Consultation on Urologic Disease Modified Oxford Systemfor Levels of Evidence and Grades of Recommendations

11

Bladder Tumors

EDITORS

MARK SOLOWAY,

ADRIENNE CARMACK, SAAD KHOURY

12

Committee 1

Bladder Cancer: Epidemiology, Stagingand Grading, and Diagnosis

Chair

Z. KIRKALI (TURKEY)

Vice Chairs

T. CHAN (USA)

M. MANOHARAN (USA)

Members

F. ALGABA (SPAIN)

C. BUSCH (SWEDEN)

L. CHENG (USA)

L. KIEMENEY (NETHERLANDS)

M. KRIEGMAIR (GERMANY)

R. MONTIRONI (ITALY)

W. MURPHY (USA)

I. SESTERHENN (USA)

M. TACHIBANA (JAPAN)

J. WEIDER (USA)

13

1. INCIDENCE AND MORTALITY

2. ENVIRONMENTAL RISK FACTORS

3. GENETIC SUSCEPTIBILITY

4. LIFESTYLE AND PROGNOSIS

1. CLASSIFICATION OF UROTHELIALNEOPLASMS

2. STAGING OF BLADDER CANCER

1. BLADDER CANCER SCREENING

2. SIGNS AND SYMPTOMS

3. URINARY CYTOPATHOLOGY

4. IMAGING OF BLADDER CANCER AT INITIALDIAGNOSIS

5. CYSTOSCOPY

6. TRANSURETHRAL RESECTION OF BLADDERTUMORS (TURBT)

RECOMMENDATIONS

REFERENCES

III. DIAGNOSIS

II. STAGING AND GRADING OFBLADDER CANCER

I. EPIDEMIOLOGY OF BLADDERCANCER

14

CONTENTS

Bladder cancer is a heterogeneous disease with avariable natural history. At one end of the spectrum,low grade Ta tumors have a low progression rate andrequire initial endoscopic treatment and surveillance,but rarely present a threat to the patient. At the otherextreme, high grade tumors have a high malignantpotential associated with significant progression andcancer death rates.

The true natural history of untreated noninvasive dis-ease is not fully known. Seventy percent of bladdertumors present as superficial disease and the rest asmuscle-invasive disease. Among the superficial can-cer group, approximately 70% present as Ta lesions,20% as T1, and 10% as carcinoma in situ (CIS, Tis).

Many characteristics of urothelial (transitional cell)carcinoma have been studied in an attempt to predictthe variable tumor behavior. These include patholog-ic features, cytologic analysis, and molecular mark-ers. Accurate staging and grading of the disease isimportant to decide the optimal treatment. An under-standing of the epidemiology and bladder screeningstrategies helps in the prevention and early detectionof the disease.

1. INCIDENCE AND MORTALITY

Bladder cancer is the fourth most common ma-lig-nancy among Western men, following prostate, lung,and colon cancer. In Europe and the United States,bladder cancer accounts for 5% to 10% of all malig-nancies among males (Figure 1). The risk of devel-oping bladder cancer before the age of 75 years is2% to 4% for males and 0.5% to 1% for females,compared to, for example, 8% and 2% for lung can-

cer, respectively (Figure 2) [1]. The median age atdiagnosis is 65 to 70 years.

Some of the differences between countries are caused by differences in registration or reporting of(low grade) pTa tumors. Unfortunately, this makesthe comparison between countries very difficult.Age-standardized (world) mortality rates vary from2 to 10 per 100,000 per year for males and 0.5 to 4per 100,000 per year for females (Figure 3) [2].

Bladder cancer is 3 to 4 times more common amongmales than females. On the other hand, it has beensuggested that the stage-adjusted survival of bladdercancer among women is worse than among men [3].The excess of bladder cancer in males is not fullyexplained by gender differences in smoking habitsand occupation (the 2 most well-known risk factorsfor bladder cancer). Surveys of cancer incidence andmortality suggest that parous women have a lowerrisk of bladder cancer than nulliparous women, probably due to hormonal changes related to pregnancy, and that the risk may decrease withincreasing parity [4-7]. Furthermore, in animal experiments, rats treated with androgenic hormonesdeveloped more bladder tumors than animals treatedwith estrogenic hormones [8]. It is therefore suggest-ed that at least some androgenic hormones stimulate(or do not inhibit) oncogenesis while estrogenic hor-mones do the opposite.

For yet undetermined reasons, blacks experienceonly half the risk of whites (Figure 4), but the overall survival among blacks seem to be worse. Thehigher incidence among whites compared to blacksis limited to superficial tumors, with blacks andwhites having a similar risk of more invasive tumors[9,10] . This suggests that some low stage, low gradetumors among blacks remain undetected. The higherrisk in whites may also be due to different risk fac-

I. EPIDEMIOLOGY OF BLADDERCANCER

15

Bladder Cancer: Epidemiology, Stagingand Grading, and Diagnosis

Z. KIRKALI

T. CHAN, M. MANOHARAN

F. ALGABA, C. BUSCH, L. CHENG, L. KIEMENEY, M. KRIEGMAIR, R. MONTIRONI, W. MURPHY, I. SESTERHENN, M. TACHIBANA, J. WEIDER

16

Figure 1. Age-standardized (World) Incidence Rates (per 100,000) of Bladder Cancer [1]

Figure 2. Cumulative Risk (%) of Bladder Cancer Before the Age of 75 Years [1]

tors operating in the black and white populations.Also, racial biologic variations and within-race indi-vidual differences may modify various phases of car-cinogenesis such as the capacity to convert procar-cinogens to carcinogens, to detoxify carcinogens,and to repair DNA.

2. ENVIRONMENTAL RISK FACTORS

a) Smoking

The most well-established risk factor for bladdercancer is cigarette smoking (Table 1), although theassociation is not as strong as that observed for

smoking and respiratory tract cancers.

Although it is well-recognized that cigarette smokingis the most important risk factor for bladder cancer ona population basis, additional factors have to play arole in modifying the risk of smoking-related bladdercancer. Populations exist with high smoking rates butlow bladder cancer rates (for example, Polynesianmen including native Hawaiians and New ZealandMaoris) [12]. This suggests differences in themetabolism of smoking-related carcinogens. Forexample, N-acetyltransferase 2 slow acetylators areknown to have a higher risk of bladder cancer fromsmoking than rapid acetylators [13]. Exogenousagents (such as vitamin intake) may modify the sus-ceptibility to smoking-induced bladder cancer aswell.

All over the world, time trends in bladder cancer fol-low trends in smoking behavior, comparable to timetrends in lung cancer but with a longer delay. In mostWestern communities, bladder cancer incidence andmortality in men has decreased in the last decade.

b) Occupation

Occupation is the first known and, on a population

17

Figure 3. Age-standardized (World) Mortality Rates (per 100,000/year) of Bladder Cancer [2]

Figure 4. Bladder Cancer Incidence in the US / 105 Person-years by Race [11]

level, second most important risk factor for bladdercancer. It has been estimated that occupational expo-sures may account for as much as 20% of all bladdercancer [14]. Exposure to beta-naphthylamine, 4-aminobiphenyl (ABP), and benzidine, principallyamong workers in the textile dye and rubber tireindustries, are the only specific agents that have beenassociated with bladder cancer unequivocally. Due tostrict regulations, these specific chemicals are nowbanned from the workplace and contribute minimal-ly to the current incidence of bladder cancer in West-ern countries. However, many other strong candi-dates for bladder carcinogens still exist such asortho-tolui-dine, which is used now in the manufac-ture of dyes, rubber chemicals, pharmaceuticals, andpesticides [15]. In fact, many occupations have beenmarked as potentially high-risk occupations. Asalready pointed out, a strongly increased risk of blad-der cancer may still exist for (ex-)workers in the dye,rubber, and chemical industries, as a result of (his-torical) exposure to aromatic amines (arylamines)like benzidine, 2-naphthylamine, 4-ABP, 4,4’-methylene-dianiline, 4,4’-methylene-bis (2-chloroaniline), o-toluidine, 4,4’-methylene-bis (2-methylaniline), and 4-chloro-o-toluidine [16,17].The risk of bladder cancer among workers in suchindustries should therefore be monitored continuous-ly. If specific plants are suspected, the identificationof the causative agent should be started immediately,preventive measures should be taken, and exposedworkers may have to be screened for bladder cancerfor at least 2 decades.

Excess risks have been frequently observed amongpainters, which is thought to be due to exposure topossible carcinogenic constituents of paints like ben-zidine, polychlorinated biphenyls, formaldehyde,and asbestos and solvents like benzene, dioxane, andmethylene chloride [18]. A moderately increased riskis also found among leather workers and shoe mak-ers, although the responsible agent is still un-known[19]. The workers are exposed to leather dust, dyes,and solvents. Therefore, it is imaginable that the ear-lier mentioned aromatic amines play an importantrole.

An excess risk of bladder cancer is also observed inaluminum, iron, and steelworkers, which may be theresult of exposure to aromatic amines and polycyclicaromatic hydrocarbons (PAHs) in coal-tar pitchvolatiles [20-22].

Furthermore, many studies have assessed the relationbetween bladder cancer and diesel exhaust exposure,and evidence is accumulating that diesel exhaustmoderately increases the risk of bladder cancer [23].Garage mechanics; drivers of trucks, buses, andcabs; and other maintenance workers in transportcompanies appear to have an increased risk of blad-der cancer. A positive trend in risk with increasingduration of employment seems to be present (Table2).

Exhaust emissions contain PAHs and nitro-PAHs.The PAHs are formed mainly as a result of pyrolyticprocesses, in particular the incomplete combustionof organic materials. Diesel engines emit at least 10times more nitro-PAHs than gasoline engines.

Although an increased risk of bladder cancer hasbeen reported for many other occupations, findingsfor most of these occupations are not consistent[25,26].

18

Table 1. Smoking and Bladder Cancer

• Cigarette smoking increases bladder cancer risk two- tofourfold

• 30% to 50% of all bladder cancer is caused by cigarettesmoking

• Latency period is approximately 20 to 30 years

• Increasing intensity and/or increasing duration → increa-sing risk

• Quitting smoking → immediate decreasing risk approa-ching baseline after 20 to 30 years

• Black tobacco threefold increased risk vs. blond tobacco

• Unfiltered cigarettes 35% to50% higher risk than filteredcigarettes

• Deep inhalation 30% to 40% higher risk than no inhala-tion

• Pipe/cigar smoking → higher risk? (inhalation pattern)

• Snuff or chewing tobacco → no elevated risk

Table 2. Motor Exhaust and Bladder Cancer [24]

Duration of Employment (years) Cases Controls Odds Ratio

Never any motor exhaust-related occupation 1353 2724 1.0

< 5 74 129 1.2

5-9 32 45 1.4

10-24 33 31 2.1

25+ 22 19 2.2

Number adjusted for age and smoking

c) Drinking Water Quality

1. CHLORINATED DRINKING WATER

In the United States and many other countries, drinking water is disinfected with chlorine. The chlo-rination is important for the microbiologic safety ofdrinking water. During the chlorination process,chlorine reacts with organics in water resulting inhalogenated organic compounds (mainly tri-halomethanes such as chloroform and bromoform).Bio-assays and in vitro studies suggest that some ofthese halogenated compounds are mutagenic or car-cinogenic. On the other hand, a recent study fromAustralia found that the trihalomethane concentra-tion in chlorinated drinking water was not related toDNA damage in bladder cells [27].

Several studies have been performed on chlorinateddrinking water and bladder cancer, and all of thesereported increased risks [28-35]. The (smoking-adjusted) risks varied from 1.4 to 2.2 for both sexescombined (exposure time varied from 20 to over 60years). In most studies, the risks tend to increase withduration of exposure. Despite these studies, a reportof the International Agency for Research on Cancer(IARC) from 1999 concluded that there was inade-quate evidence that individual chlorination by-products such as chloroform and other tri-halomethanes were carcinogenic [36]. Althoughsome studies showed an association of chlorinateddrinking water intake with cancer, it was argued thatsingle compounds could not be evaluated becausethese compounds occur in mixtures. A report of theWorld Health Organization, published in 2000, con-cluded that the evidence was insufficient to deter-mine whether observed associations were causal orto determine which specific byproduct or contami-nant plays a role.

In 2004, Villanueva and colleagues reported a pooledanalysis in which primary data from 6 case-controlstudies with individual-based exposure assessmentswere pooled [37]. These studies were conducted in 5countries (Italy [unpublished data 2003], Canada,Finland, USA, and France) using trihalomethanes asa marker for the total mixture of chlorination byprod-ucts [29,31,32,38,39]. Exposure information andcovariates were extracted from the original databas-es including age, sex, smoking, occupation, coffee,fluid consumption, and socioeconomic status. Thefinal pooled dataset comprised 3419 cases and 6077controls. All cases were histologically confirmed. Allstudies followed similar approaches to estimate thetrihalomethane levels in the water source. The expo-

sure-related variables were the amount of daily tapwater consumption and yearly average tri-halomethane concentration. The average tri-halomethane levels between the studies varied from10 to 30 µg/L. An exposure window of 40 years wasdefined, extending from 45 to 5 years before theinterview. Exposure to trihalomethanes was associat-ed with an excess risk among ever-exposed men(odds ratio [OR] 1.32, 95% confidence interval [CI]1.10 to 1.59). The risk increased with increasingexposure. Among women, no increased risk wasfound. This discrepancy between men and women isstill unclear [40]. It may be due to several mecha-nisms, for example, the role of sex hormones in themetabolization of chlorination byproducts [41,42] orthe activity of CYP2E1, which is important in themetabolism of chloroform to active metabolites, andappears to be higher in men compared to women [43-46].

In conclusion, exposure to chlorinated drinkingwater probably increases the risk of bladder cancer.Although the observed risks are relatively small, theattributable risk may be considerable, given the sizeof the exposed population.

2. ARSENIC IN DRINKING WATER

Several large studies have evaluated the associationbetween ingestion of arsenic in drinking water andthe risk of bladder cancer. Most studies have beenperformed in Asia and Latin America.

Several studies have been performed in the endemicarea of Taiwan. Between 1930 and the mid-1960s,the population in this region was exposed to highlycontaminated well water (arsenic levels of 170 to800 µg/L; current regulation is maximum 10 µg/L).These studies have shown a clear dose-response rela-tion with bladder cancer. The age- and sex-adjustedodds ratios of developing bladder, lung, and livercancers for those who had used well water for 40 ormore years were 3.9, 3.4, and 2.7, respectively, ascompared with those who never used well water[47]. In contrast to Taiwan, Bangladesh and WestBengal experience an ongoing problem with veryhigh concentrations of arsenic in drinking water thatexceed in some sources 2000 to 4000 µg/L [48].Arsenic levels in drinking water in the United Statesand Europe are much lower than reported in Taiwan.A recent case-control study by Steinmaus et al. in-cluded 7 counties in the Western United States inwhich the levels of arsenic in drinking water varyfrom 10 to 100 µg/L [49]. All odds ratios were near1.0 when an exposure window of 5 to 20 years was

19

defined. When an exposure window of 40 years ormore was used, increased odds ratios were found forarsenic intakes greater than 80 µg/day, although nonewere statistically significant. Overall, no clear asso-ciation is found between low to intermediate expo-sure to arsenic in drinking water and the risk of blad-der cancer.

d) Medical History

1. CHRONIC URINARY TRACT INFECTION

Chronic urinary tract infection is associated with thedevelopment of bladder cancer, especially invasivesquamous cell carcinoma [50]. This type of cancermay occur in patients with spinal cord injury inwhom chronic cystitis is inevitable. This may be theresult of formation of nitrites and nitrosamines bybacterial flora and/or the inflammatory process,which leads to an increased cell proliferation, pro-viding more opportunities for spontaneous geneticmistakes.

2. PHENACETIN

Heavy consumption of phenacetin-containing anal-gesics (not sold anymore) increases the risk of upperurinary tract cancer, but has only a marginal effect onbladder cancer risk [51,52].

3. CYCLOPHOSPHAMIDE

Cyclophosphamide, an alkylating agent used in thetreatment of malignant neoplasms, particularly lym-phoproliferative and myeloproliferative diseases,increases the risk of bladder cancer (mainly urothe-lial carcinoma) with a clear dose-response relation-ship (Figure 5).

Cyclophosphamide is acutely toxic to the bladdermucosa and produces cellular abnormalities in the

epithelium. Most cyclophosphamide-induced tumorspresent themselves as muscle-infiltrating lesions atthe time of diagnosis with a relatively short latencyperiod (6-13 years). It is unclear whether cyclophos-phamide-induced urothelial malignancies are due toits immunosuppressive or inherent carcinogenic properties, but it is likely that the 2 factors worktogether. Of 4 known metabolites of cyclophos-phamide, acrolein and phosphamide mustard havebeen demonstrated to bind to DNA, and acrolein isknown to be responsible for its bladder toxicity.

4. RADIOTHERAPY

Radiotherapy is also a known risk factor for bladdercancer. Kaldor et al. carried out a case-control studyof tumors of the bladder in women who had previ-ously been treated for ovarian cancer [54]. The riskof bladder tumors was increased for patients who hadbeen treated with radiotherapy or chemotherapy(thiotepa and melphalan) compared to patients treated with surgery. Moreover, the risk seemed to bemuch higher in patients who received both.

e) Schistosomiasis

Squamous cell carcinoma of the urinary bladder hasbeen known to be associated with Schistosomahaematobium infection for many years. The epi-demio-logic association is based both on case-con-trol studies and on the close correlation of bladdercancer incidence with the prevalence of S. haemato-bium infection within different geographic areas [55-58]. S. haematobium is found throughout much ofAfrica and the Middle East. The life cycle of S.haematobium requires waterborne transmission ofinfection between man and snail. Water becomescontami-nated by Schistosoma eggs when peopleurinate in the water. The eggs release larval forms(miracidia) that have to penetrate appropriate snailhosts within 16 to 32 hours. Inside the snail, the lar-vae will grow and develop into cercariae over 6weeks. After this period, the cercariae leave the snailand are able to survive in water for 2 to 3 days. Inthis period, the parasites have to find a human hostand penetrate the skin by mechanical action andenzymatic secretions of the cephalic glands. The par-asites migrate by the blood vessels where theymature to male and female worms. These mate andtravel to the vesical veins to deposit eggs in the blad-der wall. Finally, the eggs are extruded into the urine(Figure 6).

It is not the worms but the eggs that cause the dis-ease in humans; the egg deposit causes a responsethat results in cystitis and hematuria. Over time, the

20

Figure 5. Risk of Bladder Cancer According to Cumulati-ve Dose of Cyclophosphamide and Administration ofRadiotherapy [53]

chronic inflammatory response from schistosomiasisleads to changes in the urothelium resulting in squa-mous metaplasia of the epithelium. Individuals withchronic schistosomiasis may eventually developsquamous cell carcinoma, probably as a result of ahigher amount of carcinogenic nitroso compounds inthe urine and/or a depressed immunocompetence ofinfected patients [59,60].

f) Diet

In observational studies, high vegetable and fruitconsumption has been associated with a decrease inthe risk of almost all cancers including bladder,although several large cohort studies report no pro-tective effect [62,63]. A role of diet and nutrition inbladder carcinogenesis is plausible since most sub-stances and metabolites (including precarcinogens)are excreted by the urinary tract. A possible biologicmechanism is that several antioxidants (vitamins A,C, and E; retinol; selenium; and folate) detoxify freeradicals and thereby decrease cancer risk. Surpris-

ingly, large intervention trials did not find clear pro-tective effects of antioxidants. The Physicians’Health Study (PHS), involving 22 000 male physi-cians in the United States, and the Alpha-Tocopherol,Beta-Carotene Cancer Prevention Study (ATBCcohort study), involving 29 000 Finnish male smok-ers, investigated supplements of antioxidant nutri-ents in different doses and combinations and pro-duced surprising results [64,65]. One arm of theATBC cohort study involved daily supplementationwith beta-carotene and, after 5 to 8 years of supple-mentation, an unexpected 18% higher incidence oflung cancer and an 8% higher overall mortality wasfound. In the PHS, no effect of beta-carotene on dis-ease and death was observed. Concerning alpha-tocopherol, a 34% lower incidence of prostate can-cer, but no effect on lung cancer, was found in theATBC cohort study.

During the follow-up period of the ATBC cohortstudy, 344 men developed bladder cancer. Consump-tion of fruits and vegetables was not associated withthe risk of bladder cancer [66]. Similarly, no associ-ations were observed for specific groups of fruits orvegetables. Dietary intake of alpha-carotene; beta-carotene; lycopene; lutein/zeaxanthin; beta-cryptox-anthin; vitamins A, E, and C; and folate were notrelated to the risk of bladder cancer. This study sug-gests that fruit and vegetable intakes are not likely tobe associated with bladder cancer risk. However, ina meta-analysis by Steinmaus et al., a risk of 1.4(95% CI 1.08 to 1.83) was found for a diet low infruit intake [67].

Evidence for an increase in bladder cancer riskbecause of a high total fat intake is insufficient. Thesame holds for the relation with diets high in choles-terol or saturated (animal) fat. People with a high in-take of fried food were found to have 2 to 3 times therisk of bladder cancer than people who did not con-sume fried foods. Although this association wasfound in several studies, evidence remains insuffi-cient.

Fairly convincing evidence exists with regard to arisk-increasing effect of food pyrolysis products(mutagenic and carcinogenic heterocyclic amines),which are formed during cooking, broiling, grilling,or barbecuing of meat and fish.

1. ALCOHOL

Although consumption of low levels of alcohol (1 or2 drinks a day) is believed to reduce cardiovasculardisease, positive associations have been reportedwith the risk of several types of cancer (such as oral

21

Figure 6. Life Cycle and Geographic Distribution of Schis-tosoma haematobium [61]

cavity and pharynx, larynx, liver, esophagus, colon,and breast). Alcohol consumption has often beenstudied as a possible risk factor for bladder cancer.The results are not consistent but point in the direc-tion of no association. For example, in a large studyamong Danish brewery workers with a very highaverage beer intake of 2 to 2.5 liters/day, no elevatedrisk was found [68]. The positive findings in somestudies may be the result of residual confounding bysmoking (several case control studies did not adjustfor smoking habits) or chance.

2. COFFEE

Similar to the above mentioned association, coffeeand the risk of bladder cancer has often been studiedbut the results are very inconsistent, although pointing in the direction of a weak positive associa-tion. In the United States bladder cancer study with2982 cases and 5782 controls, the largest population-based case-control study on bladder cancer, the riskfor ever versus never coffee drinkers (adjusted forage, race, geographical area, and tobacco consump-tion) was found to be 1.6 (95% CI 1.2 to 2.2) formales and 1.3 (95% CI 0.8 to 1.7) for females [69].However, in a Danish study no elevated risk wasfound for coffee drinking and bladder cancer, thoughin Scandinavian countries coffee consumption isabout 2.5 times as high as in the United States [70].In a recent pooled analysis of several case-controlstudies in Europe, 564 cases and 2929 controls wereincluded who had never smoked [71]. No excess riskwas found in ever versus never coffee drinkers. Forcoffee drinkers with more than 10 cups per day, anincreased risk was found in both men and women(OR 1.0; 95% CI 1.0 to 3.3).

Most inconsistencies in the reported data are proba-bly due to the effect of residual confounding by smoking or by other correlates of coffee drinking.Because coffee drinking and smoking are stronglycorrelated, an incomplete controlling of smoking(because of misclassification of smoking habits) canresult in an apparent relation of coffee drinking andbladder cancer. Another, more theoretical, possibilityis that people with a relatively high bladder cancerrisk (slow acetylators) drink more coffee [72].

3. ARTIFICIAL SWEETENERS

No consistent positive association between bladdercancer risk and the use of artificial sweeteners (likesaccharine and cyclamate) has been found, althoughanimal models (rats) suggested a strong relation.This may be due to the fact that the specific biologicmechanism in some animal models does not exist in

man. In studies among diabetics in the United Statesand United Kingdom who have substantial use ofartificial sweeteners, no elevated bladder cancermortality rates were found [73]. In a study byTakayama et al., 20 monkeys were treated with sodi-um saccharine (5-10 times the daily intake forhumans) for 5 days a week from 24 hours after birthuntil the age of 24 years [74]. Sixteen monkeysserved as controls. No evidence was found of forma-tion of solid material (crystals) in the urine, provid-ing additional evidence that sodium saccharine hasno carcinogenic effect on the primate urinary tract.The most recent study among humans on this subjectwas conducted by Sturgeon et al [75]. Bladder can-cer patients (1860) and controls (3934) weregrouped into low (1.68mg/day) sweetener consumption. High consumptionof sweeteners was correlated with a modestincreased risk of developing bladder cancer (OR 1.3,95% CI 0.9 to 2.1).

4. TOTAL FLUID CONSUMPTION

Several observational studies have assessed the rela-tionship between total fluid intake and the risk ofbladder cancer, but the results of these studies arevery inconsistent. Positive, negative, and no associa-tion have been reported [28,76-78].

g) Hair Dyes

The risk of bladder cancer through the use of haircoloring products has been studied since the late1970s but received more interest in the last fewyears. Hair dyes are widely used in Western commu-nities. It is estimated that more than one-third ofwomen above the age of 18 and more than 10% ofmen above the age of 40 use some type of hair dye.Some hair dyes contain aromatic amines such as 4-aminobiphenyl, of which small amounts may beabsorbed percutaneously [79]. The levels of sus-pected carcinogens differ between the different typesof dyes (permanent, semi-permanent, and temporaryrinse dyes) and between the different colors. Occu-pational exposure to hair dyes by hairdressers isbelieved to moderately increase the risk of bladdercancer (combined analysis of studies yielded OR1.4) [80]. Since the publication of a study in 2001,the personal use of hair dyes has received muchattention. After 2 cohort studies from the UnitedStates (the Nurses Health Study and the AmericanCancer Society CSP-II study) did not find an associ-ation between personal use of permanent dyes andbladder cancer, Gago-Dominguez and colleaguesreported a case-control study from Los Angeles thatinvolved almost 900 incident cases of bladder cancer

22

and an equal number of age-, sex-, neighborhood-,and ethnicity-matched controls [81-83]. Overall, reg-ular use of hair dyes did not increase the risk of blad-der cancer. However, women who used permanenthair dyes at least once a month experienced acigarette-smoking adjusted 2.1-fold risk of bladdercancer relative to non-users. The use of this type ofhair dye among men was too rare (about 2% amongcontrols) for meaningful analyses. Semi-permanentand temporary dyes did not increase the risk of blad-der cancer. In further analyses, women who usedthese types of dyes were combined with the non-users. The relative risk of permanent dyes amongwomen showed dose-response relationships withduration of use, frequency of use, and cumulativenumber of times used over lifetime [83]. In responseto this study, 1 of the 2 earlier studies with negativeresults, the American Cancer Society CSP-II study,published an update. After 16 years of follow-up of acohort of more than 500 000 women, the death rateof bladder cancer among women (N=336) was notrelated to the use of permanent hair dyes. Evenamong women who used these dyes for more than 20years, the relative risk for bladder cancer mortalitywas not increased [84].

In a case-control study from New Hampshire, in-cluding 98 women with bladder cancer and 238 con-trols, the use of permanent hair dyes was associatedwith a 1.5-fold increased risk of bladder cancer.However, the use of rinse dyes was associated withan increased risk as well (OR 1.7). Both associationswere not statistically significant and clear dose-response effects were not observed [85]. The differ-ent results between these studies are difficult toexplain. At the end of 2001, Gago-Dominguez et al.reported an additional analysis of their data. Theyphenotyped 61% of the female cases and 60% of thecontrols for N-acetyltransferase, an important en-zyme in the detoxification of aromatic amines.

They found that the increased risk of the use of per-manent hair dyes was restricted to N-acetyltransfer-ase 2 slow acetylators [86]. Recently, this study wasextended with the genotyping of GSTM1, GSTT1,GSTP1, and NAT1, and the phenotyping of NAT2and CYP1A2. Again, it was demonstrated that dif-ferences in arylamine activation and detoxificationpathways (especially NAT2, NAT1, and CYP1A2)modify the relationship between permanent hair dyesand bladder cancer in women [87]. These findingssupport a causal role of permanent hair dyes butleave the discrepancy with earlier studies unan-swered.

3. GENETIC SUSCEPTIBILITY

a) Gene-Environment Interactions

The Los Angeles County bladder cancer study onhair dyes and bladder cancer is a good example ofthe possible modifying role of constitution on theeffect of environmental or lifestyle factors [87].Although many environmental risk factors have beenimplicated in the etiology of bladder cancer, onlyoccupation and smoking are responsible for a con-siderable part of all new cases. Without any doubt,genetic susceptibility will prove to be responsible foranother considerable part, most importantly by gene-environment interactions. Most environmental car-cinogens are activated inside the body into reactiveoxygenated intermediates (which may interact withDNA) by phase I enzymes. Subsequently, these reac-tive intermediates are detoxified into conjugatedwater-soluble products by phase II enzymes. Forexample, oxidation of arylamines in the liver bycytochrome P450 1A2 leads to N-hydroxylatedmetabolites. These may enter the circulation andreact covalently with hemoglobin in erythrocytes.After filtration into the bladder lumen the metabo-lites may react with urothelial nucleic acids (Figure7). Detoxification of the reactive metabolites intonon-carcinogenic N-acetylated arylamides may take

23

Figure 7. Biochemical Pathway for Urinary Bladder Can-cer Induction by Aromatic Amines [88]

place by N-acetylation, which is catalyzed by the N-acetyltransferases, preferentially by NAT2 in theliver and by NAT1 in skin cells.

The genes that code for these enzymes showinterindividual variability leading to differences incatalytic activity. Most of this variability comes fromsingle nucleotide polymorphisms (SNPs) and tan-dem repeats, but small homozygous deletions in spe-cific genes may even lead to a total absence ofenzyme activity. Numerous studies have examinedthe role of metabolic gene polymorphisms in thedevelopment of bladder cancer. Genes that have beenstudied include NAT1, NAT2, GSTM1, GSTT1,GSTP1, CYP1A1, CYP1A2, CYP1B1, CYP2D6,CYP2E1, CYP3E4, ADH3, NQO1, SULT1A1,MPO, COMT, and MnSOD. The role of NAT2 slowacetylation phenotype or genotype and bladder can-cer is most frequently studied. In a meta-analysis of22 studies, it appeared that slow acetylators have a40% (95% CI 1.2 to 1.6) increased risk of bladdercancer. This increased risk, however, may be differ-ent in different populations (higher in Asians andlower or even absent in United States Caucasians)[13].

The phase II glutathione-S-transferases are classifiedinto at least 4 genetically distinct classes of enzymes(alpha, theta, mu, and pi), which conjugate reactiveentities with glutathione. In humans, mu and thetaclass of GST isoenzymes, GSTM1 and GSTT1 dis-play different phenotypes due to deletions of thegenes. GSTM1 is highly efficient in conjugating arylepoxides, which are formed by phase I cytochromeP450 enzymes after exposure to certain PAHs.GSTM1 may also detoxify aromatic amines. Enzymeactivity is found in several organ systems such as theliver and bowel, in lymphocytes, and also in thebladder endothelium. GSTM1 activity is absent inapproximately 40% of the Western populationbecause of a homozygous deletion of the GSTM1locus on chromosome 1p13 [89]. In a meta-analysisof 17 studies, a lack of GSTM1 activity was associ-ated with bladder cancer with an odds ratio of 1.44(95% CI 1.23 to 1.68) [90]. For specific variants inmost other metabolic genes that were studied untilnow, the results are fairly inconsistent or were re-ported in 1 study only.

Recently, attention has also been paid to polymor-phisms in genes with other functions, such as DNArepair, cell cycle control, immune response, folatemetabolism, and cell adhesion (for example XRCC1,XPD, XRCC3, CDH1, TP53, TNF, HER-2, HRAS1,CCND1, MTHFR, and MS). Many positive findings

have been reported. It is too early, however, to assessthe relevance of these findings. Again, inconsistentresults, potential (or perhaps even probable) publica-tion bias, and the absence of validation studies sug-gest that false positive findings are a real problem.This research area would benefit greatly from inter-national collaborations where large-scale projectsstudy the role of many polymorphisms simultane-ously using high-throughput array technology, whilevalidating the findings in collaborative projects. The present non-standardized study-by-study and gene-by-gene approach will only lead to alot of confusion. Furthermore, perhaps the mostimportant aim of studies on gene polymorphismsshould be to identify subgroups in which environ-mental risk factors can be evaluated with increasedpower. It is therefore advised to give as muchemphasis with data collection on hypothesized envi-ronmental risk factors as on genetic polymorphisms.

b) Familial Bladder Cancer

Urothelial carcinoma of the ureter and renal pelvis isknown to be part of the Hereditary Non-polyposisColon Cancer (HNPCC) syndrome, which is causedby mutations in DNA mismatch repair genes. For yetunknown reasons, urothelial carcinoma of the blad-der is not part of this syndrome [91]. In fact, familialbladder cancer is a fairly rare phenomenon comparedto the familial occurrence of many other tumor sites.Nevertheless, numerous case reports describe fa-mil-ial clustering of urothelial carcinoma. Severaldemonstrate an extremely early age at onset, sug-gesting a genetic component [92]. Only a few epi-demiological studies specifically addressed familialbladder cancer. Goldgar and colleagues studied therisk of cancer in first-degree relatives of cancerprobands, yielding an increased risk for bladder can-cer among first-degree relatives of 1.5 (95% CI 1.0to 2.2) [93]. When only young probands (age < 60years) were considered, the familial risk was 5.1(95% CI 1.0 to 12.5). In a study from New York,demographic data and cigarette smoking status on allthe first degree relatives of 319 male bladder cancercases and 319 neighborhood controls were collected.The cohorts were linked to the cancer registry, yielding an almost twofold risk (hazard ratio [HR]1.9, 90% CI 0.9 to 4.1) of bladder cancer for first-degree relatives [94]. By contrast, a comparablestudy among the population of Iceland reported onlya slightly elevated risk of urothelial carcinomaamong relatives of 190 bladder cancer patients withan observed-to-expected ratio of only 1.2 (95% CI0.9 to 1.7) [95]. The largest study on familial bladder

24

cancer was conducted by the same group but in theNetherlands. Using a family case-control design,1193 patients newly diagnosed with urothelial carci-noma of the bladder, ureter, renal pelvis, or urethrawere contacted. Information on the patients’ first-degree relatives was collected by postal question-naire and subsequent telephone calls. The patients’partners filled out a similar questionnaire on theirrelatives. All reported occurrences of urothelial car-cinoma among the 8014 first-degree case-relativesand 5673 control-relatives were verified using medi-cal records. Disease occurrence among case-relativesand control-relatives was compared with randomeffect proportional hazards regression analyses whileadjusting for age, sex, and smoking behavior. Amongthe case-relatives, 101 individuals were diagnosedwith cancer of the bladder (97), ureter (3), and renalpelvis (1), compared to 38 individuals among thecontrol-relatives (bladder 36, ureter 1, and urethra 1).In 6 case-families and 2 control-families, 2 affectedfirst-degree relatives were found. Overall, 8% of thepatients had a positive family history of urothelialcarcinoma compared to 4% of the controls. Themean age at diagnosis of patients with a positivefamily history was similar to that of patients with anegative family history (62 years). The mean age atdiagnosis of urothelial carcinoma among affectedcase-relatives was only slightly lower than that ofaffected control-relatives (64 vs. 66 years). Thecumulative risk of urothelial carcinoma among case-relatives was 3.8% compared to 2.1% among con-trol-relatives. The age- sex-, and smoking-adjustedhazard ratio of urothelial carcinoma for case-rela-tives compared to control-relatives was 1.8 (95% CI1.3 to 2.7). This risk appeared to be higher amongwomen (HR 3.2) and among nonsmokers (HR 3.9).When only parents were included in the analyses, thehazard ratio increased to 2.2, while it decreased to1.5 when only siblings were included. After stratifi-cation by tumor site in the probands (upper vs. lowerurinary tract), the adjusted HR was 1.8 among rela-tives of probands with upper urinary tract urothelialcarcinoma and 1.9 among relatives of probands withbladder urothelial carcinoma. When all the relativesof probands with a pTa tumor were excluded fromthe analyses, the HR increased only slightly to 2.0.The same was found when the relatives of probandsolder than 60 years were excluded from the analyses(HR 2.4). A striking clustering of tumors at othersites among the case-relatives was not foundalthough an increased risk was observed for tumorsof the hematolymphopoietic system (HR 1.9, 95% CI1.2 to 3.0). Familial clustering with other tumors was

also evaluated among the relatives of probands witha positive family history of urothelial carcinoma.There was some suggestion (nonsignificant) of aclustering of tumors of the female genital organs,non-urothelial urinary tract tumors, and cancer of thehematolymphopoietic system [96].

From these epidemiological studies, it can be con-cluded that the risk of bladder cancer is increasedapproximately twofold with a positive family historyof bladder cancer. The next question is whether thisis caused by genetic susceptibility or shared environ-ment. In a twin study from Scandinavia (Denmark,Sweden, and Finland), Lichtenstein et al. reported 5concordant and 146 discordant pairs of bladder can-cers among 7231 monozygotic male twin pairs.Among 13 769 dizygotic male twin pairs, 2 concor-dant pairs and 253 discordant pairs were found. Therelative risk of bladder cancer among monozygotictwins was 6.6 (95% CI 2.6 to 16.9). The relative riskamong dizygotic twins was 1.7 (95% CI 0.4 to 6.9).The concordance rate was 3 times higher amongmonozygotic twins (6%) than among dizygotic twins(2%). Assuming that the correlation in environmen-tal risk factors is similar among monozygotic anddizygotic twins, this finding suggests the existenceof a genetic etiology of bladder cancer [97]. In astudy on the Swedish Family-Cancer Database, itwas estimated that 7% of the occurrence of bladdercancer is due to genetic effects, 12% to shared envi-ronmental effects, 4% to childhood environmentaleffects, and 77% to nonshared environmental effects[98].

The group from the Netherlands evaluated whethermutagen sensitivity plays a role in developingurothelial carcinoma and whether this sensitivity isdifferent in familial and nonfamilial cases. Intrinsicsusceptibility was quantified by a mutagen sensitivi-ty assay (mean number of chromatid breaks per cell(PBLs) after damage induction with bleomycin in thelate S-G2 phase of the cell cycle). Twenty-five spo-radic patients, 23 familial patients (2 patients in 1nuclear family), and 13 hereditary patients (2patients < 60 years or 3 patients in 1 nuclear family)were selected and compared with control subjectswithout a history of cancer. Patients with urothelialcarcinoma showed a higher mutagen sensitivityscore compared to control subjects (mean number ofchromatid breaks per cell 0.91, 95% CI 0.84 to 0.97,and 0.74, 95% CI 0.69 to 0.79, respectively; P =0.001), suggesting a genetic origin [99].

Recently, a new bladder cancer gene was discoveredby the collaborative group of Schoenberg and

25

Sidransky at Johns Hopkins in Baltimore [100]. In1996, this group identified a family in which a malewas diagnosed with grade 2 superficial urothelialcarcinoma of the bladder at the age of 29 years. Hesubsequently developed renal pelvis carcinoma. Hismother died of metastatic urothelial carcinoma of thebladder at the age of 65. Because both the proband’swife and his mother had a history of miscarriages, akaryogram was made which showed a balancedgermline translocation t(5;20)(p15;q11) [101]. Dr.Sidransky’s lab zoomed in at the breakpoints of thistranslocation, which finally resulted in the discoveryof a new bladder cancer gene at 20q11 [100]. Thisgene, CDC91L1, encoding CDC91L1, which is alsocalled phosphatidylinositol glycan class U (PIG-U),has a role in the glycosylphosphatidylinositol (GPI)anchoring pathway. Further research suggested thatthe gene is amplified and overexpressed in as manyas one-third of bladder cancers and primary tumors.CDC91L1 should therefore be regarded as an onco-gene. The translocation led to overexpression of thegene and, probably, to both bladder cancers in thispedigree. Carriers of the translocation in this familywere therefore genetically susceptible for bladdercancer. Because the exact translocation site shouldbe regarded as an extremely rare phenomenon, thisgene should not be considered as a candidate for thegenetic cause of many patients with hereditary blad-der cancer. For that, tumor suppressor or DNA mis-match repair genes have yet to be discovered.Recently, it was shown that inherited mutations inthe retinoblastoma tumor suppressor gene may causebladder cancer. In a long-term follow-up study fromthe UK, 5 bladder cancer cases were found among144 hereditary retinoblastoma cases, an observed toexpected ratio of 26.3 (95% CI 8.5 to 61.4) [102].This study shows that bladder cancer should be puton the list of tumors for which hereditary retinoblas-toma patients should be checked during lifetime fol-low-up.

4. LIFESTYLE AND PROGNOSIS

Although much research has been done and is stillongoing on prognostic factors in invasive and super-ficial bladder cancer (such as angiogenesis, cellcycle control proteins, and cell adhesion molecules)the management of patients is usually based on clin-ical characteristics only (such as tumor stage, grade,and lymph node involvement). Based on these characteristics, it appears to be very difficult to pre-dict prognosis on an individual level. This appliesboth to the risks of recurrence and progression in

superficial disease and to the risk of metastases anddeath in invasive disease. Surprisingly, little atten-tion has been paid to the role of lifestyle and consti-tution on prognosis. With the current knowledge, it isimpossible for clinicians to advise patients about theprognostic benefit of lifestyle changes. Bladder can-cer prognosis is related to age (younger patients hav-ing a better prognosis), gender (males having a bet-ter prognosis), and race (whites having a better prog-nosis), but all of these differences can be largelyexplained by a different stage at diagnosis.

In a systematic review by Aveyard et al., 15 studieswere identified in which the prognostic role of smok-ing and stopping smoking was reported [103]. Over-all, the results suggested that continued smoking or alifetime of smoking constitutes a moderate risk factor for recurrence and death. Also, stopping smok-ing probably results in a clinically relevant prognos-tic improvement. Unfortunately, most of the studieshad major methodological shortcomings. Firm evi-dence-based conclusions can therefore not be drawn.In a study from Japan of 258 patients, drinking ofalcoholic beverages was significantly associatedwith a better survival (HR 0.46, 95% CI 0.3 to 0.8),but a clear dose-response relationship was notobserved. No prognostic significance was found forthe use of artificial sweeteners, coffee, powderedgreen tea, and cola consumption [104]. Without anydoubt, a large body of data still exists from studies onthe etiology of bladder cancer. These data may besupplemented quite easily with follow-up data inorder to learn more about the prognostic value oflifestyle. Also, future epidemiological studies shouldconsider the collection of follow-up data for this pur-pose.

1. CLASSIFICATION OF UROTHELIALNEOPLASMS

In December 1998, members of the World HealthOrganization (WHO) and the International Societyof Urologic Pathologists (ISUP) published theWHO/ISUP consensus classification of urothelial(transitional cell) neoplasms of the urinary bladder(Level 4, [105]). This new classification system aroseout of the need to develop a universally acceptableclassification system for bladder neoplasia that couldbe used effectively by pathologists, urologists, and

II. STAGING AND GRADING OFBLADDER CANCER

26

oncologists. Prior to this classification system,numerous diverse grading schemes for bladder can-cer existed whereby the same lesion seen by differ-ent pathologists would result in very different diag-noses solely based on definitional differences oflesions. Another strength of the consensus classifica-tion system is that it provides detailed histologicalcriteria for papillary urothelial lesions. In contrast,prior grading systems for bladder tumors were vagueand subjective. This classification system not onlycovered neoplastic conditions, but also the nomen-clature of preneoplastic lesions .

a) Normal and Hyperplastic Urothelium

Many pathologists overuse the diagnosis of “milddysplasia” to describe flat lesions with minimallydisordered growth pattern or cellular hyperchroma-sia due to variation in tissue fixation, staining, orspecimen orientation. The consensus classificationstates that the term “mild dysplasia” should not beused and that flat lesions with minimal cytologic a-typia and architectural disorder should be diagnosedas “normal.”

Flat urothelial hyperplasia consists of a markedlythickened mucosa without cytologic atypia, and maybe seen adjacent to low grade papillary urothelialneoplasms, but there is no data as to its premalignantpotential when seen by itself.

Papillary urothelial hyperplasia is characterized byurothelium of variable thickness with undulatinggrowth. In contrast to papillary urothelial tumors,these lesions lack distinct fibrovascular cores. Papil-lary urothelial hyperplasia without cytologic atypia,because of its frequent association with either a prioror concurrent history of a low grade papillary blad-der neoplasm, is thought to be a precursor lesion ofthese neoplasms (Level 3, [106]). Papillary urothelialhyperplasia may also be lined by cytologically atyp-ical urothelium ranging from dysplasia to flat CIS,which are often associated with and thought to be aprecursor of high grade papillary urothelial carcino-mas (Level 3, [107]).

b) Flat Lesions With Atypia

Prior to the consensus classification, different pa-thologists variably used the terms “atypia” and “dys-plasia” to denote either inflammatory atypia or a pre-neoplastic condition. The WHO/ISUP systemdescribed the histological findings associated withinflammatory atypia and designated these lesions as“reactive atypia,” which should not be consideredneoplastic.

Dysplasia (intraurothelial neoplasia) was defined asa lesion with appreciable cytologic and architecturalabnormalities felt to be neoplastic, yet falling shortof the diagnostic threshold for carcinoma in situ.There is evidence along several lines of investigationthat dysplasia may be a precursor of invasive carci-noma (108-111, Level 3; 112, Level 2; 113, Level 3).

CIS is a flat lesion of the urothelium, and docu-mented precursor of invasive cancer in many cases(Figure 8A). Prior to the consensus classification,CIS was frequently underdiagnosed and described asmoderate dysplasia or atypia. The WHO/ISUP sys-tem described the key histologic features of CIS,including its more subtle variations that were oftenunderrecognized. Under the WHO/ISUP system, allCIS are by definition high grade lesions.

c) Papillary Urothelial Neoplasms: Classification

The classification of papillary urothelial neoplasmshas been a long-standing source of controversy(Level 4, [114]). One source of controversy lies inthe grading of papillary urothelial carcinomas.There are numerous grading systems, all of whichhave poor interobserver reproducibility, with mostcases falling into the intermediate category (115,Level 4; 116, Level 2; 117, Level 4; 118,119, Level 3;120, Level 2; 121, Level 3; 122, Level 2). TheWHO/ISUP system is a modified version of thescheme proposed by Malmströ m et al. [120]. Amajor limitation of the WHO 1973 grading system isthe vague definition and lack of specific histologicalcriteria of the various grades. The following state-ment is the sole description of the difference betweenWHO grades 1, 2, and 3, as written in the originalWHO 1973 book: “Grade 1 tumors have the least

27

Figure 8A. Flat Urothelial Carcinoma in Situ

degree of anaplasia compatible with the diagnosis ofmalignancy. Grade 3 applies to tumors with the mostsevere degrees of cellular anaplasia, and Grade 2 liesin between” [117]. Detailed histological descriptionof the various grades, employing specific cytologicand architectural criteria is one of the major contri-butions of the WHO/ISUP system. These criteria arebased on the architectural features of the papillae andthe overall organization of the cells. Cytologic fea-tures encompassed in the WHO/ISUP system in-clude nuclear size, nuclear shape, chromatin content,nucleoli, mitoses, and umbrella cells. The terminol-ogy used in the WHO/ISUP system parallels thatused in urine cytology. Having a consensus classifi-cation between cytology and histopathology is alsoadvantageous. A website (www. pathology. jhu. edu/bladder) illustrating examples of the various gradeswas developed to further improved accuracy in usingthe WHO/ISUP system.

1. RELATION OF WHO 1973 TO WHO/ISUP

A major misconception is that there is a one-to-onetranslation between the WHO/ISUP and the WHO1973 classification systems. Only at the extremes ofgrades in the WHO 1973 classification, does this cor-relation hold true. Lesions called papilloma in theWHO classification system would also be calledpapilloma in the WHO/ISUP system. At the otherend of the grading extreme, lesions called WHOgrade 3 are by definition high grade carcinoma in theWHO/ISUP system. However, for WHO grades 1and 2, there is no direct translation to the WHO/ISUPsystem. Some lesions classified as WHO grade 1 inthe 1973 system that upon review show no cytologicatypia, some nuclear enlargement, and merely thick-ened urothelium are papillary urothelial neoplasmsof low malignant potential (PUNLMP) in theWHO/ISUP system. However, other WHO grade 1lesions showing slight cytologic atypia and mitosesare diagnosed in the WHO/ISUP system as low gradepapillary urothelial carcinomas. WHO grade 2 is avery broad category. It includes lesions that are rela-tively bland, which in some places are diagnosed asWHO grade 1 to 2; these lesions in the WHO/ISUPsystem would be called low grade papillary urothe-lial carcinoma. In other cases, WHO grade 2 lesionsborder on higher grade lesions, which in many insti-tutions are called WHO grade 2 to 3; these lesions inthe WHO/ISUP classification system would be called high grade papillary urothelial carcinoma.

2. PAPILLOMA

Using the WHO 1973 system, some experts applied

very restrictive criteria for the diagnosis of urothelialpapilloma, in part based on the number of cell layers,and regarded all other papillary neoplasms as carci-nomas. Others applied a broader definition of“urothelial papilloma” so as not to label all patientswith papillary lesions with minimal cytologic andarchitectural atypia as having carcinoma. TheWHO/ISUP system has very restrictive histologicfeatures for the diagnosis of papilloma, where nor-mal appearing urothelium lines papillary fronds.Defined as such, it is a rare benign condition typical-ly occurring as a small, isolated growth seen primar-ily in younger patients. The majority of these lesionsonce excised will not recur (Level 3, [123]).

3. PAPILLARY UROTHELIAL NEOPLASMS OF LOWMALIGNANT POTENTIAL (PUNLMP)

The category of PUNLMP was derived to describelesions that do not have cytologic features of ma-lig-nancy, yet have thickened urothelium as compared topapilloma (Figure 8B). There is no or very little

28

Figure 8B. Papillary Urothelial Neoplasm of Low Malignant Potential (PUNLMP)

variation of nuclear features or the pattern of organi-zation. Having a category of PUNLMP avoids label-ing a patient as having cancer with its psychosocialand financial (i.e., insurance) implications, althoughthey are not diagnosed as having a benign lesion(i.e., papilloma), whereby they might not be fol-lowed as closely. The prognosis of these lesions andother papillary tumors in the WHO/ISUP system willbe discussed later in this work. The current classifi-cation system allows for designation of a lesion (pap-illary urothelial neoplasm of low malignant poten-tial), that biologically has a very low risk of progres-sion, yet is not entirely benign. In the past, theselesions were a source of controversy, as some expertswould label such lesions as malignant, while others,not wanting to label a patient with such a low gradepapillary lesion as having carcinoma, would diag-nose these lesions as papilloma. This inter-mediatecategory allows both schools of thought to diagnosea lesion as not fully malignant, yet still documentsneed for additional follow-up.

4. LOW AND HIGH GRADE PAPILLARY CARCINOMA

In an attempt to simplify the WHO 1973 system andavoid an intermediate cancer grade group (WHOgrade 2), which is often the default diagnosis formany pathologists, the WHO/ISUP system classifiespapillary urothelial carcinoma into only 2 grades.Low grade papillary urothelial carcinoma exhibits anoverall orderly appearance but has minimal varia-bility in architecture and/or cytologic features, whichare easily recognizable at scanning magnification(Figure 8C 1&2). High grade papillary urothelialcarcinomas are characterized by a disorderly appear-ance due to marked architectural and cytologicabnormalities, recognizable at low magnification(Figure 8D). It is important to remember that a sin-gle papillary urothelial neoplasm may contain aspectrum of cytologic and architectural abnormali-ties. In tumors with variable histology, the tumorshould be graded according to the highest grade,although current practice is to ignore minuscule

29

Figure 8C1. Papillary Carcinoma, Low Grade Figure 8C2. Papillary Carcinoma, Low Grade

areas of higher grade tumor. Studies are needed todetermine how significant a minor component mustbe in order to have an impact on prognosis.

d) Papillary Urothelial Neoplasms: Prognosis

One of the earliest papers reported to have used theWHO/ISUP system has led to misconceptions re-garding the classification system. In October 1998, amanuscript was submitted for publication and even-tually published in Cancer entitled “PapillaryUrothelial Neoplasms of Low Malignant Potential”(Level 3, [124]). This article was submitted beforethe WHO/ISUP classification system was even pub-lished and before there were detailed illustrations ordescriptions on how to classify tumors using the newsystem. The authors took lesions that were formerlycalled WHO grade 1 and designated them asPUNLMP. As described above, there is not a one-to-one translation between the WHO/ISUP and WHO1973 classification systems. If these lesions wereanalyzed correctly using the WHO/ISUP classifica-tion system, many of these lesions would not be clas-sified as PUNLMP, but would be diagnosed as lowgrade urothelial carcinomas. It is no surprise that thecases they classified as PUNLMP had an increased

risk of recurrence, progression, and death from blad-der cancer. This article has been cited as an argumentagainst the use of the WHO/ISUP system. However,given that WHO grade 1 cancers were simplyrenamed as PUNLMP in that study, it should noteven be considered as having used the WHO/ISUPsystem.

Before discussing the prognosis of tumors using theWHO/ISUP system, it is worthwhile to briefly high-light differences among studies in terms of theirinclusion criteria and definition of progression.Some reports restrict their study to only cases with-out invasion (pTa). Although stage pT1 disease isstill considered to be “superficial bladder cancer,”once lamina propria invasion is identified, thesepatients are at an increased risk of subsequentlydeveloping detrusor muscle (muscularis propria)invasion (stage pT2). Consequently, including pTa(noninvasive) and pT1 (superficially invasive)tumors leads to a heterogeneous group of patients.Another difference in inclusion criteria is that somestudies include patients with CIS, while others donot. CIS is one of the more aggressive lesions in thebladder despite its flat morphology. Includingpatients with CIS in a study of noninvasive or mini-mally invasive papillary carcinomas adds an addi-tional variable, which must be taken into accountwhen analyzing prognosis. The definition of pro-gression also varies among studies. Some includeprogression from pTa to pT1, while others requireevolution to pT2. In some but not other studies, achange in grade or the development of CIS is con-sidered progression.

The first article to use the WHO/ISUP system as itwas meant to be used and to correlate its lesions withprognosis was published by Desai et al. in 2000(Level 2, [125]). The authors examined 120 pTa andpT1 tumors, including patients even if they had CIS.They showed significant differences in prognosisamong the various categories. While papillomas didnot recur or progress, and PUNLMP tumors recurredbut did not progress, low grade and, to a greaterextent, high grade carcinomas experienced progres-sion and, in some cases, death (Table 3).

In 2001, Alsheikh et al. examined 49 patients withpTa tumors who did not receive any additional treat-ment after an initial transurethral resection (Level 2,[126]). The authors focused on the differences be-tween the 20 PUNLMPs and 29 low grade carcino-mas. Twenty-five percent of the PUNLMPs recurred,in contrast to 48% of the low grade carcinomas. Ofthe 2 patients who progressed to high grade muscle-

30

Figure 8D. Papillary Carcinoma, High Grade

invasive carcinoma, both were initially low gradecarcinomas. There was also one patient who pro-gressed to CIS who also initially had low grade car-cinoma.

The largest study to date looking at the WHO/ISUPclassification system is that by Holmang et al. (Level2, [127]). Of the 363 pTa tumors evaluated, 83% ofthe patients received no additional treatment untillater. Progression was defined as tumors that devel-oped lamina propria invasion beyond the muscularismucosa or metastatic disease. Most patients withPUNLMP had no evidence of disease, and only asmall percentage of patients had tumor at last follow-up, but no one was dying of disease. In contrast,patients with low grade carcinoma had an increasedrisk of tumor being present at last follow-up, in addi-tion to a small percentage of patients dying of dise-ase. Patients with high grade carcinoma had a largerrisk (16%) of dying of disease (Table 4). Low andhigh grade carcinomas had similar risks of recur-rence, in contrast to a lower risk with PUNLMPs. Interms of progression, PUNLMPs and low grade car-cinomas had similar risks compared to an increasedrisk with high grade carcinomas. When the authorscompared the risk of progression between WHOgrade 2 and grade 3 lesions, there was a greater dif-ference in terms of the risk of progression comparedto the difference between low grade and high grade

carcinomas using the WHO/ISUP classification sys-tem. Since most WHO grade 3 cases are aggressivetumors and already have coexisting invasive cancer,very few patients with WHO grade 3 tumors initial-ly satisfied the criteria of having noninvasive papil-lary carcinoma; of the 363 noninvasive papillary car-cinomas, only 3.6% were classified as WHO grade 3.Furthermore, some of the WHO grade 3 noninvasivepapillary cancers had coexistent CIS. With theWHO/ISUP system, Holmang et al. classified 28%of the cases as high grade carcinoma, with an in-creased risk of progression.

Pich et al. also focused their investigations on dif-ferences between PUNLMPs and low grade carcino-mas (Level 2, [128]). In addition to recurrence andprogression, p53 expression and proliferation mark-ers were analyzed. Sixty-two pTa tumors were studied. No patients received adjuvant therapy untilrecurrence. Progression was defined as any invasionor metastases. Differences in recurrence were notedbetween PUNLMPs and low grade cancers withrecurrence rates of 47.4% and 76.7%, respectively.While none of the PUNLMPs progressed, 11.6% ofthe low grade carcinomas progressed. There was alsoa difference in the recurrence-free interval betweenPUNLMPs and low grade carcinomas, with 76- and15-month recurrence-free intervals, respectively.Differences were also noted between the 2

31

Table 3. Relation of WHO/ISUP Grades to Progression [125]

Papilloma PUNLMP Low Grade High Grade(n=8) (n=8) (n=42) (n=62)

Recurrence 0% 33.3% 64.1% 56.4%

Any Stage Progression 0% 0% 10.5% 27.1%

Lamina Propria Invasion 0% 0% 2.6% 8.3%

Detrusor Muscle Invasion 0% 0% 5.3% 6.3%

Metastases/Death 0% 0% 10.6% 25.0%

Table 4. Relation of WHO/ISUP Grades to Progression [127]

PUNLMP Low Grade High Grade(n=95) (n=160) (n=108)

No Evidence of Disease 94% 76% 67%

Alive With Disease 3% 10% 9%

Dead With Disease 1% 6% 7%

Dead of Disease 0% 4% 16%

No Follow-up 2% 4% 1%

WHO/ISUP grades in their p53 expression, mitoses,and MIB1 (a proliferation marker) activity.

Cina et al. have also analyzed tumors using theWHO/ISUP classification system for p53 expressionand proliferation as measured by KI67 (Level 3,[129]). Increases in p53 expression of 0.4%, 2.9%,and 25.7% were documented in cases of PUNLMP,low grade carcinoma, and high grade carcinoma, re-spectively. Proliferation also increased among the 3grades: 2.5%, 7.3%, and 15.7%, respectively. In aseparate study, 134 patients with pTa tumors withoutprior or concurrent CIS or invasion were analyzed(Level 2, [130]). Progression was defined as tumorrecurrence with any invasion (pT1 or pT2) or CIS.The 90-month actuarial risks of progression forWHO papilloma and carcinomas grade 1, grade 2,and grade 3 were 0%, 11%, 24%, and 60%, respec-tively. The corresponding progression rates forWHO/ISUP papilloma, PUNLMP, low grade andhigh grade carcinomas were 0%, 8%, 13%, and 51%,respectively. WHO grade (P = 0.003) and tumor size(P = 0.03) and WHO/ISUP (P = 0.002) and tumorsize (P = 0.04) independently predicted progression.Although WHO grade 3 cancers had a more rapidprogression rate and a slightly worse long-term pro-gression rate compared to WHO/ISUP high gradecancer, only 4.5% of tumors were WHO grade 3 ascompared to 21.6% classified as WHO/ISUP highgrade. These findings are similar to Holmang’sstudy, where only 3.6% of tumors were classified asWHO (1973) grade 3, as compared to 28% classifiedas WHO/ISUP high grade carcinoma (Level 2,[127]). As patients with high grade noninvasive pap-illary carcinoma are not treated with definitive ther-apy (such as cystectomy), the goal should not be torestrict this high risk group to a very small popula-tion, but to expand it to include all patients at signif-icant risk of progression for close monitoring.

e) Invasive Urothelial Carcinoma

Confusion in terminology is not limited to the diag-nostic entities in the various classification schemes,as it also exists for the descriptive terminologyapplied to invasive urothelial lesions. Various termsinclude “superficial muscle invasion,” “deep muscleinvasion,” “muscle invasion (not otherwise speci-fied),” and “superficial bladder cancer.” The latterterm is particularly confusing as it could be appliedto CIS, noninvasive papillary neoplasms, or trulyinvasive urothelial carcinoma. Due to variations intreatment and prognostic significance related to thedepth of invasion of bladder tumors, the consensusgroup developed several recommendations to pro-

vide clinicians with this e