UNDIFFERENTIATED SARCOMA OF T H E KIDNEY A Tumor of Childhood with Histopathologic and Clinical

Characteristics Distinct from Wilms’ Tumor

ELAINE MORGAN, MD, AND JAMES M. KIDD, M D

The pathological and clinical features of 123 patients with Wilms’ tumor diag- nosed between 1961 and 1977 were analyzed. A subgroup of nine patients with distinctive histology characterized by the presence of exclusively sar- comatous elements was identified. These patients were indistinguishable clinically from the other patients with Wilms’ tumor and were treated as Wilms’ tumor. The rate of recurrence in this group of sarcomatous Wilms’ tumor was very high (7/9) and long-term disease-free survival correspondingly low (3/9 free of disease from 15 months to nine years) when compared to the patients with classical Wilms’ tumor (50% recurrence rate and 71% currently free of disease). In addition, patients with metastases (seven of seven patients with recurrence had skeletal involvement as contrasted to one of 50 patients with metastatic classical Wilms’ tumor). We conclude that the “sarcomatous Wilms’ tumor” represents a distinct clinicopathological entity identifiable morphologically at diagnosis, best described as renal sarcoma and that in the future optimal treatment of this disease may vary somewhat from the treat- ment of classical Wilms’ tumor.

Cancer 42: 1916- 1921, 1978.

HE TERM WILMS’ TUMOR or nephro- T blastoma has been used to encompass a wide diversity of malignant renal tumors of childhood which are similar in gross patho- logic appearance but whose histology may be variable. Morphologically they may contain both mesenchymal and epithelial elements in differing ratios and in differing stages of m a t ~ r i t y . ~ , ’ ~ Poorly to well-differentiated glomerular and tubular structures are fre- quently but not universally present in the tumor t i s ~ u e . ~ Mesenchymal elements such as muscle, cartilage, adipose tissue and occa- sionally bone may also be identified.*.l9

At one time it was felt that there was no relationship between histology and prognosisz0 and therefore no need to subclassify these tumors. However, more recently, renewed interest in histopathologic correlations has led to the identification of at least one distinct tumor type which is no longer included in the category of Wilms’ tumor, namely the

From the Department of Pediatrics, Northwestern University Medical School, Children’s Memorial Hospi- tal, Chicago, Illinois.

Address for reprints: Elaine Morgan, MD, Dept. of Pediatrics, Northwestern University Medical School, Children’s Memorial Hospital, 2300 Children’s Plaza, Chicago, IL 60614.

Accepted for publication December 2 1 , 1977.

infantile mesoblastic n e p h r ~ m a , ~ . ~ , ~ ~ which has little or no metastatic potential. In addi- tion, several studies have indicated a relation- ship between histology and prognosis within the broad category of nephroblastoma6 with more well-differentiated tumors containing a predominance of epithelial elements having in general a more favorable o u t c ~ r n e . ~ ~ ~ ~ A subgroup of tumors containing largely or solely sarcomatous elements has been identi- fied by several investigators and has been associated with a high metastatic rate and a low long term (3-5 year) survival rate1,3’6,15”8 (Table 1).

In the past, little attention has been paid to differing patterns of metastatic spread among the various histologic subclassifications of nephroblastoma. However, we have sus- pected for some time that there was a small nucleus of tumors, histologically distinct, with a marked propensity to metastasize to bone.12

The present study describes this pathologi- cal entity previously included in the overall category of Wilms’ tumors. The tumor is com- posed solely of sarcomatous elements with no convincing nephroblastic features. Clini- cally, the group of patients with this type of tumor follow a course which is atypical for Wilms’ tumor with a high rate of relapse, a

0008-543X/78/1000/1916 $0.75 0 American Cancer Society

1916

No. 4 UNDIFFERENTIATED RENAL SARCOMA . Morgan and Kidd 1917

TABLE 1. Influence of Cell Type

Number of patients

Series Total Sarcoma Classic

Relapse rate (%) Survival (%)

Total Sarcoma Classic Total Sarcoma Classic

NWTS 13 429 22 407 Currie et al. 19736 21 1 o* 11 Perez et al. 197318 47 24* 23 Jereb 19731° 112 31* 81 Lemerle 19761g 195 63* 132 Makinen 197516 58 16 42

26.3 72.7 23.8 85.11 36.45. 87.7t 47.61 10**t 81.91

51 72* 30.4 69.6 81* 34.6 46 66* 40 56.9f 40**f 65.9f

50t 18.9f 61.9:

* Includes anaplastic and undifferentiated tumors. t 2 year plus. f 5 year.

marked tendency to metastasize to bony sites and, correspondingly, a low incidence of pul- monary metastases and an overall poor prog- nosis for long term survival. Recognition of this form of kidney tumor as a clinicopatho- logic entity may be of importance prognosti- cally and may ultimately influence treatment programs.

MATERIALS AND METHODS

From 1961 to 1977, 123 children were seen at Children’s Memorial Hospital in Chicago with a diagnosis of Wilms’ tumor. A review of the pathological findings revealed a sub- group of nine patients whose histology was distinctive and characterized by the presence of exclusively sarcomatous elements, showing no discernible differentiation to any particu- lar cell type. An additional five patients with malignant renal tumors (four leiomyosarcomas and one rhabdomyosarcoma) were excluded from the present study as they were felt to represent distinct pathologic entities charac- teristic of neither classical Wilms’ tumor nor the undifferentiated sarcoma herein described.

The records of the 123 patients were ex- amined in a retrospective manner to deter- mine whether they were clinically different from the overall group of Wilms’ tumor pa- tients with regard to age, sex, stage, sites of metastases, treatment and outcome.

Patients

Patients having this undifferentiated sar- coma of kidney were similar to the overall group of patients with Wilms’ tumor in re- spect to age distribution, sex, and stage of disease. The median age of this group at diagnosis was 2% years (range 1% years to 5% years). The ratio of maledfemales was

5/4. There were no cases of bilateral tumor in this group. Distribution is illustrated in Table 2. Only one patient was stage IV at outset. This distribution is similar to the pa- tient characteristics of children with Wilms’ tumor as reported in the literature and is not different from the group of patients seen at Children’s Memorial Hospital with classical Wilms’ tumor.

Patients with Wilms’ tumor have been treated by a variety of methods over the years encom- passed by this survey, with a gradual evolution to the current program of combined modality treatment. Earlier patients were treated with a combination of surgery, radiation therapy and actinomycin D. Later, vincristine was added to the treatment of Wilms’ tumor and recently adriamycin” has also been used in these patients.21 Patients with the sarcomatous type of Wilms’ tumor here referred to as un- differentiated sarcoma of kidney have been treated at the Children’s Memorial Hospital according to the prevailing concept of therapy for Wilms’ tumor at the time of their diag- nosis. Thus, while the therapy has not been uniform, it has been identical with the treat- ment of the children with classical Wilms’ tumor treated here at this hospital at the same time, and thus is generally compatible.

PATHOLOGICAL DESCRIPTION

Gross Features

The majority of , the tumors weighed from 300-800 g; however, two tumors were excep- tionally large, weighing over 1,500 g. Extra- renal extension was grossly seen in five cases and microscopically in one further case. The

* Adriamycin-doxorubicin hydrochloride, Adria Lab- oratories, Wilmington, Delaware.

1918 CANCER October 1978 Vol. 42

TABLE 2. Patient Characteristics Microscopic Features

Total patients Sex (MIF) Median age Group

I I1 111 IV V

Unknown

Wilms’ tumors

114 49/65

3

16 (14%) 62 (54%) 16 (14%) 13 ( 1 1 % ) 3 (3%) 4 (4%)

Undiffer- entiated sarcoma

9 514 2%

2 (22%) 5 (67%) 1 ( 1 1 % )

All patients

123 53/70

18 (15%) 67 (54%) 17 (14%) 13 (10%) 3 (3%) 4 (4%)

tumors grossly were similar to the classic Wilms’ tumor with the tumor mass involving the body or body and one pole of kidney and presenting on cut surface a slightly lobulated gray-white to pink-white sarcomatous pouting appearance often with focal areas of cystic degeneration and hemorrhage. The tumor was generally of firmer rubbery consistency and less lobulated than the classical Wilms’ tumor. The residual renal parenchyma was stretched over the tumor with apparent good demarcation of kidney and tumor. In two cases, large tumorous renal hilar masses were noted.

The above group of tumors presented a histologic picture quite distinct from Wilms’ tumor. The tumor was highly cellular and composed of loosely aggregated cells with oval-to-polygonal pale nuclei with sparsely scattered particulate chromatin and no well defined nucleoli (Fig. 1). Many nuclei ap- peared naked without cytoplasm; however, some had scanty eosinophilic feathery cyto- plasm extending from one pole without form- ing a cytoplasmic border. The tumor cells lay haphazardly in aggregates between a uniformly distributed delicate fibrovascular stroma to which the peripheral tumor cells often adhered (Fig. 2). Only occasionally did strap-like extensions of cytoplasm extend from the nuclei. PTAH and reticulum stains were negative. There was no evidence of lym- phatic or venous invasion. Occasional medium sized vessels seen within the tumor were sur- rounded by a thicker cuff of fibrous tissue. In one case there was in addition a distinct perpendicular orientation of nuclei to the delicate fibrous stroma lending a trabecular pattern to some parts of this tumor. Mod- erate mitotic activity was seen throughout all tumors. Of particular importance was the

FIG. 1. Haphazardly arranged tumor cells with characteristic pale ovoid to polygonal nuclei with sparsely dis- tributed chromatin; moderate mitotic activity (~400).

No. 4 UNDIFFERENTIATED RENAL SARCOMA . Morgan and Kidd

FIG. 2. Highly cellular sarcoma supported by uniformly distrihuted delicate fibrovascular strotna ( X 100)

absence of any nephroblastic elements. Al- though occasional tubules were noted, these were interpreted as entrapped mature tubules within invading tumor tissue. The overall picture was that of an undifferentiated sar- coma of nonnephroblastic potentiality. The clear nuclei of the haphazardly arranged tumor cells and evenly dispersed fibrovascular stroma contrasts with the polarization and hyperchromatisin of the undifferentiated ele- ments seen in a poorly differentiated Wilms’ tumor.

RE s u LTS

The subgroup of nine patients with undif- ferentiated sarcoma of kidney clinically had a decreased overall survival and a high rate of metastasis with a specific predilection to metastasize to bone.

As can be seen (Table 3) 44% of patients in this group have died; 23% are alive with disease and 33% are alive with no evidence of disease from 15 months to nine years. Of the patients, 78% have had metastases (Table 4). All (seven patients) have had metastatic involvement of the skeleton at a median time of 11 months from diagnosis. In four patients (44%) the first site of metastasis was skeletal. Only one patient developed pulmonary me- tastases.

1919

In contrast to this, 71% of patients with classical Wilms’ tumor are alive with no evi- dence of-disease from four months to 16 years after diagnosis. Death has occurred in 23% (27 patients)-24 with disease and three of other causes. Metastases developed in 50% (57 patients) at a median time from diagnosis of three months. Of these, 11 % (13 patients) were stage IV at diagnosis. In patients with metastases 44% (50 patients) metastasized to lung, 23% (27 patients) to other sites, most commonly abdomen, contralateral kidney and liver. One patient (0.9%) with multiple sites of metastases developed lesions in two ribs overlying a paravertebral mass. This was diagnosed only by a positive bone scan in an

A ABLE 3. Clinical Course

Undiffer- Wilms’ entiated tumors sarcoma Overall No. of No. of No. of

patients patients patients

Alive-NED 81 (71%) 3 (33%) 84 (67%) Alive with

disease 2 (2%) 2 (23%) 4 (4%) Exp. with

Exp.-NED 3 (2%) 0 (0%) 3 (2%) Lost to E’/U 4 (4%) 0 (0%) 4 (4%)

disease 24 (21%) 4 (44%) 28 (23%)

Median survival 5 + years 30 months

1920 CANCER October 1978 Vol. 42

TABLE 4. Metastases

Undiffer- Wilms’ entiated tumors sarcoma No. of No. of

patients patients Overall

No recurrence 57 (50%) 2 (22%) 59 (48%) Metastases 57 (50%) 7 (78%) 64 (52%) Median time to

recurrence 3 mos. 11 mos.

Sites of metastases*

No. % o f No. % o f No. % o f pts. mets. pts. mets. pts. mets.

Lung 50 (44%) 1 (11%) 51 (42%) Bone 1 (0.9%) 7 (78%) 8 (6%) Other 27 (23%) 3 (33%) 30 (24%)

*Some patients may have more than one site of involvement.

area of bone pain. Routine radiographs of the area revealed no abnormality.

DISCUSSION

In reviewing all patients seen at this insti- tution in the past 16 years with malignant tumors of the kidney we have identified sev- eral different pathologic groups. The majority of tumors were classified as Wilms’ tumor showing varying amounts of mesenchymal and epithelial elements. A small group of tumors initially classified as Wilms’ tumor but composed of undifferentiated sarcoma- tous elements has been diagnosed as “sar- comatous Wilms’ tumors.” Although the total number of patients with such tumors seen at our institution is small, they are clinically different from patients with classical Wilms’ tumor in several important characteristics.

First, as noted by other investigator^',^^^^^^^- 15,16~18 their incidence of metastasis is higher and percent of long-term survival significantly less than that of patients with classical Wilms’ tumor (Table 1). This overall poor prognosis persists despite standard combined modality treatment with radiation therapy, aggressive surgery and combination chemotherapy with vincristine and actinomycin D, which cur- rently yields an overall two years disease-free rate of 80% (3.7) to 94% (9). In addition to the higher incidence of metastases in the sar- comatous group, the ability to salvage patients in this group once relapse has occurred seems less than in the classical Wilms’ tumor patients. Thus of the seven patients with undifferen- tiated sarcoma who have developed metastases,

only one is currently alive without evidence of disease 2% years following the recurrence. The other patients have either died of disease or have had multiple relapses and/or progres- sive disease. In contrast to this, 28 of 57 pa- tients with classical Wilms’ tumor who have had metastases are currently free of disease from 2% to 15 years. Thus appearance of metastases in the patient with undifferen- tiated sarcoma of kidney carries an extremely poor prognosis for long-term survival.

Second, the median time of first recurrence appears to be somewhat longer in the sar- comatous group than in the nonsarcomatous group (1 1 months vs three months). This is suggestive of a trend toward later metastases in these children but is not statistically sig- nificant because of the small number of pa- tients.

Third, and perhaps most significant, the pattern of metastases was different in patients with sarcoma from patients with Wilms’ tumor. In classical Wilms’ tumor patients, the pattern of metastatic sites is generally similar, with 63-81% involving the lungs, 37-52% involv- ing intra-abdominal sites and 23-33% involv- ing other sites (brain, skin, bone, e t ~ . ) . ~ , ~ * * , ~ ~ * - 15~17323 Only 10- 13% of metastases from Wilms’ tumor involve the bony ~kele ton .~ , ’~ All seven of the patients with undifferentiated sarcoma who recurred had bony lesions. In four of the seven, the first recurrence involved the skele- ton. Of greater significance is that only one of seven patients developed pulmonary lesions. Furthermore, only one of the patients with classical Wilms’ tumor developed skeletal involvement and no patient with classical Wilms’ tumor developed radiologic bone lesions. This difference is statistically signifi- cant (p<O.Ol). The overall incidence of skeletal involvement in the 123 patients with malignant renal tumors was 8/123 or 6.5%. This is comparable to the incidence of skeletal metastases in Wilms’ tumor reported by other ~ b ~ e r ~ e r ~ . ~ ’ ~ ~ , ~ ~ , ~ ~ ~ ~ ~ In previous reports, how- ever, no mention has been made of any dis- tinguishing pathologic features of the primary tumors in these patients. It is possible that these patients were unrecognized cases of non-Wilms’ sarcoma, although we cannot sup- port this hypothesis with data. Certainly at this institution skeletal metastases were seen almost exclusively in patients with sarcomas.

In summary, by retrospective analysis we have identified a group of patients with undif- ferentiated sarcoma of the kidney. Their tumors, previously diagnosed as Wilms’ tu- mors, are pathologically different from Wilms’

No. 4 UNDIFFERENTIATED RENAL SARCOMA * Morgan and Kidd 1921

tumors and are probably identical with the entity of sarcomatous Wilms’ tumors described by other ~bservers . ’ ,~ ,~*’~,’~ They are also dis- tinguished clinically from patients with Wilms’ tumors by several features including a high rate of metastases, a low salvage rate once metastasis has occurred and a definite tend- ency for skeletal involvement with a corre- spondingly poor long term prognosis. Thus in terms of natural history of disease these patients resemble patients with other sarco- mas more than patients with Wilms’ tumors and therefore might more accurately be de- scribed pathologically as non-Wilms’ sarcoma of the kidney. The high incidence of recur- rence and low rate of long-term survival in patients with sarcomatous Wilms’ tumor has been noted by other investigators. Many of

these have combined the anaplastic poorly differentiated and sarcomatous tumors to- gether as a single pathologic subgroup of Wilms’ tumor having a worse prognosis than other types of Wilms’ tumors.6~10~11”8 Ho wever, the tendency for skeletal metastases seen in our patients has not been previously reported. This clinical observation combined with the previously cited course of disease seems to support the designation of the “sarcomatous Wilms’ tumor” as a distinctive clinicopatho- logical entity, namely that of renal sarcoma. It is hoped that by identifying these patients at the time of diagnosis we may be able to devise more appropriate chemotherapy for them and thereby improve their chances for a long term survival.

REFERENCES

1. Beckwith, J. B., and Palmer, N. F.: National Wilms’ tumor study I1 informational bulletin 5, June 15, 1977.

2. Bolande, R. P., Brough, A. J.. and Izant, R. J.: Congenital mesoblastic nephroma of infancy. A report of eight cases and the relationship to Wilms’ tumor. Pediatrics 40:272-278, 1907.

3. Breslow, N. E., Palmer, N. F., and Hill, L. R.: Wilms’ tumor: Prognostic factors for patients without metas- tasis at diagnosis-Results of the national Wilms’ tumor study. Cancer 41:1577-1589, 1978.

4. Cassady, J. R., Tefft, M., Filler, R. M., Jaffe, N., Pald, N., and Hellman, S.: Considerations in the radiation therapy of Wilms’ tumor. Cancer 32:598-608, 1973.

5. Chatten, J.: Epithelial differentiation in Wilms’ tumor. A clinicopathologic appraisal. Perspect. Pediatr. Pathol. 3:225-254, 1976.

6. Currie, D. P., Daly, J. T., Grimes, J. H., and Ander- son, E. E.: Wilms’ tumor: A clinical pathological correla- tion.]. Urol. 109:495-500, 1973.

7. D’Angio, G. H., Evans, A. E., Breslow, N., Beckwith, B., Bishop, H., Feigl, P., Goodwin, W., Leape, L. L., Sinks, I,. F., Sutow, W. W., Tefft, M., and Wolff, J.: The treatment of Wilms’ tumor-Results of the national Wilms’ tumor study. Cancer 38:633-646, 1976.

8. Hardwick, D. F., and Stowens, D.: Wilms’ tumors. J. Urol. 85:903-910, 1961.

9. Harrison, J., Myers, M., Wiley, A., and Vermund, H.: Management of Wilms’ tumor with particular emphasis on combination treatment with radiation, actinornycin D and vincristine. Int. J. Radiat. Oncol. Biol. Phys. 1:227-234, 1976.

10. Jereb, B., and Sangstedt, B.: Structure and size vs. prognosis in nephroblastoma. Cancer 3 1: 1473- 148 1, 1973.

1 1 . Kenny, G. M., Webster, J. H., Sinks, L. M., Gaeta, J. F., Staubitz, W. J., and Murphy, G. P.: Results from treatment of Wilms’ tumor at Roswell Park 1927-28. J . Surg. Oncol. 1:49-61, 1969.

12. Kidd, J. M.: Exclusion of certain renal neoplasms from the category of Wilms’ tumor, Pediatric Pathology Club Spring Meeting, St. Louis, March 7, 1970.

13. Kumari, S.: Wilms’ tumor with bony metastases. South. Med. J. 69:812-813, 1976.

14. Lawler, W., Marsden, H. B., and Palmer, M. K.: Wilms’ tumor-histologic variation and prognosis. Cancer 36:1122-1126, 1975.

15. Lemerle, J., Tournade, M., Gerard-Marchant, R., Flamant, R., Sarrazin, D., Flamant, F., Lemerle, M., Jundt, S., Zucker, J., and Schweisguth, 0.: Wilms’ tumor: Natural history and prognostic factors. Cancer 37:2557- 2566, 1976.

16. Makinen, J., and Rapola, J.: Renal tumours in children. A histological evaluation. Acta Pathol. Micro- bid . Scand. 83:237-244, 1975.

17. Movassaghi, N., Leikin, S., and Chandra, R.: Wilms’ tumor metastasis to uncommon sites. J.Pediatr. 84:416-417, 1974.

18. Perez, C. A,, Kaiman, H. A,, Kerth, J., Mill, W. B., Vietti, T. J., and Powers, W. E.: Treatment of Wilms’ tumor and factors affecting prognosis. Cancer 32: 609-617, 1973.

19. Sullivan, M. P., Hussey, D. H., and Ayala, A. G.: Wilms’ tumor. In Clinical Pediatric Oncology, W. W. Sutow, T. J. Vietti, and D. Fernbach, Eds. St. Louis, C. V.

20. Swenson, 0.: Wilms’ tumor and neuroblastoma. In Pediatric Surgery, Vol. 1 , Edition 111, 0. Swenson, Ed. New York, Meredith Publishing Go., 1969; pp. 874-895.

21. Tan, C., Rosen, G . , Chavemi, F., Haghbin, M., Helson, L., Wollner, N., and Murphy, M. L.: Adriamycin (NSC-123127) in pediatric malignancies. Cancer Chemo- ther. Rep. 6:259-266, 1975.

22. Wigger, H. J.: Fetal rhabdomyomatous nephro- blastoma-a variant of Wilms’ tumor. Hum. Pathol. 7:613-623, 1976.

23. Wolff, J. A,, D’Angio, T., Hartmann, J., Krivit, W., and Newton, W. A,: Long term evaluation of single vs. multiple courses of actinomycin D therapy of Wilms’ tumor. N . Engl. J . Med. 290:84-86, 1974.

Mosby, 1973; pp. 359-384.