Histology of Tumor Residuals Following Chemotherapy in Patients with Advanced Nonseminomatous Testicular Cancer

0022-534 7/89/1425-1239$02.00/0 THE JOURNAL OF UROLOGY Copyright© 1989 by AMERICAN UROLOGICAL ASSOCIATION, INC.

Vol. 142, November Printed in U.S.A.

HISTOLOGY OF TUMOR RESIDUALS FOLLOWING CHEMOTHERAPY IN PATIENTS WITH ADVANCED NONSEMINOMATOUS TESTICULAR CANCER SOPHIE D. FOSSA, NINA AASS, SIGURD OUS, JOHAN H0IE, ANNA E. STENWIG, HANS H. LIEN, ELISABETH PAUS AND OLAV KAALHUS From the Departments of Medical Oncology and Radiotherapy, Surgical Oncology, Pathology, Diagnostic Radiology, and Biophysics, Cancer Research Institute, and Central Laboratory, The Norwegian Radium Hospital, Oslo, Norway

ABSTRACT

A total of 111 patients with advanced nonseminomatous testicular cancer underwent cisplatinbased combination chemotherapy, followed by surgical removal of residual masses in 101. Surgery included retroperitoneal lymph node dissection in 92 patients, thoracotomy in 19 and hepatic resection in 1 (11 patients underwent 2 operations). Complete necrosis and/or fibrosis was found in 52 operative specimens, mature teratoma in 37 and vital malignant tumor in 12. Of the 11 patients who underwent 2 operations 4 had complete necrosis and/or fibrosis in both histological specimens. After a median observation of 55 months 83 of 89 patients with complete necrosis and/or fibrosis or mature teratoma were without evidence of disease. Only 7 of 12 patients with vital malignant tumor in the operative specimen survived without evidence of disease. Relapses were observed in 16 patients, 4 of them in the retroperitoneal space. Of the 16 relapses 5 were in 12 patients with residual vital malignant tumor, 5 in 37 patients with post-chemotherapy mature teratoma and 4 in 52 patients with complete necrosis and/or fibrosis after chemotherapy. Two patients with recurrence did not undergo an operation. In patients in whom post-chemotherapy retroperitoneal lymph node dissection is considered complete necrosis and/or fibrosis can be predicted by the combination of several factors, including absence of teratomatous elements in the testicular tumor, complete response on post-chemotherapy computerized tomography, and normal a-fetoprotein and human chorionic gonadotropin levels after chemotherapy (sensitivity 83%, specificity 76% and correctly predicted 79%). With the knowledge of these factors it seems possible to omit post-chemotherapy retroperitoneal lymph node dissection in approximately 20% of the patients with advanced metastatic nonseminomatous testicular cancer with initial retroperitoneal tumors. (J. Ural., 142: 1239-1242, 1989) The majority of patients with metastatic testicular nonseminoma are cured with cisplatin-based chemotherapy followed by surgical removal of residual tumors. 1- 5 Clinical research concentrates on the avoidance of over-treatment in good risk patients and intensification of treatment in patients at poor risk. Avoidance of long-term side effects is of utmost importance. Dry ejaculation represents a major side effect after bilateral retroperitoneal lymph node dissection, which often is performed in testicular cancer patients after induction chemotherapy. 6· 7 However, in many patients the removed tumor tissue is completely necrotic,a... 11 indicating that tumor removal is a diagnostic rather than a therapeutic procedure. Therefore, it is in demand to define factors that after chemotherapy predict the presence of complete necrosis and/or fibrosis, thus, identifying patients in whom a post-chemotherapy operation can be omitted. PATIENTS AND METHODS

From 1980 to 1986, 111 patients with advanced metastatic testicular cancer underwent cisplatin-based induction chemotherapy at our hospital. All patients had stage IIB or greater disease according to the Royal Marsden Hospital classification system: stage I-no metastases evident outside the testis, stage II-infradiaphragmatic nodal metastases (stage IIA-metastases less than 2 cm. in diameter, stage IIB-2 to 5 c,m. in diameter, stage IIC-5 to 10 cm. in diameter and stage IIDAccepted for publication April 19, 1989. Supported by The Norwegian Cancer Society.

more than 10 cm. in diameter), stage III-supradiaphragmatic nodal metastases [abdominal status 0, normal computerized tomography (CT) scan (stages IIIA, IIIB and IIIC as for stage II], and stage IV-extranodal metastases (stage IVLl-3 or fewer pulmonary metastases, stage IVL2-multiple small pulmonary metastases less than 2 cm. in diameter, stage IVL3multiple small metastases with 1 or more greater than 2 cm. in diameter and stage IVH +-hepatic involvement, abdominal status as for stage III, table 1). 12 Of the patients 89 had been included in the Swedish-Norwegian Testicular Cancer Project.13 All patients had determinations of serum tumor markers (afetoprotein, upper normal limit less than 20 µg./1. and human chorionic gonadotropin, upper normal limit less than 10 units per 1.) before each chemotherapy cycle and before the postchemotherapy operation. At diagnosis all patients underwent abdominal CT and most also underwent thoracic CT. The CT examinations were repeated 3 to 4 weeks after cycle 4 and, if chemotherapy was continued, after completion of chemotherapy. Complete remission of any retroperitoneal tumor was defined as disappearance of the tumor or residual masses with a maximum diameter of 1 cm. 2 or less on a transverse CT plane. All histological specimens were reviewed by 1 of us (A. E. S.), and stratified according to the classification systems of Collins and Pugh, 14 and of the World Health Organization. 15 Treatment details have been reported previously. 16- 18 In summary, induction chemotherapy consisted either of 4 cycles of a slightly modified Einhorn regimen (CVB-20 mg./m. 2 cisplatin on days 1 to 5, 0.15 mg./kg. vinblastine on days 1 and 2, and

1239

,, !, I

1240

FOSSA AND ASSOCIATES TABLE

1. Patient and treatment details

No. pts. Median age (range) Stage:

IIB IIC-D

111 26 (15-59)

TABLE

2. Post-chemotherapy histology (all procedures) and treatment outcome Not N~cros!s/ Mature Maii~:~nt Totals Operated F1bros1s Teratoma Tumor (lll pts.) (10 pts.) (52 pts.) (37 pts.) (l 2 pts.)

22 16

IIIA-B IIIC-D IV OLl-3 IVA-BLl-3 IVC-DLl-3 Extrapulmonary hepatic metastasis Histology: Teratoma, differentiated Malignant teratoma, intermediate Malignant teratoma, undifferentiated Malignant teratoma, trophoblastic Seminoma* Median levels before chemotherapy (range):t a-fetoprotein (µg./1.) Human chorionic gonadotropin (U./1.) No. pts. with normal a-fetoprotein and normal human chorionic gonadotropin Median mos. observation time (range) Chemotherapy::j: Type CVB and/or BEP20 BEP40/BEP60 + CVB/BEP20 No. cycles:

<4 4

>4 Type of post-chemotherapy operation (1 procedure/2 procedures): Retroperitoneal lymph node dissection Thoracotomy Hepatic resection Not operated

6 4

12 24 24 3 9

42 42 15 3

69 (5-81,000) 27 (5-130,000) 19 55 (0.5-102) 94 17 12 86 13

90/2 10/9 1 10

* With a-fetoprotein production. t Upper normal limits less than 20 µg./1. for a-fetoprotein and less than 10 units per I. for human chorionic gonadotropin. :j: CVB-cisplatin (20 mg./m. 2 x 5), vinblastine (0.15 mg./kg. x 2) and BEP20-cisplatin (20 mg./m. 2 x 5), etoposide (100 mg./m. 2 X 5) and bleomycin (90 mg.), BEP40-cisplatin (40 mg./m. 2 X 5), etoposide (120 mg./m. 2 X 3) and bleomycin (90 mg.) and BEP60-cisplatin (60 mg./m. 2 X 3), etoposide (120 mg./ m. 2 X 3) and bleomycin (90 mg.).

90 mg. bleomycin) or BEP20 cycles (vinblastine substituted by 100 mg./m. 2 etoposide on days 1 to 5) or high dose cisplatin regimens (BEP40 or BEP60-40 mg./m. 2 cisplatin on days 1 to 5 or 60 mg./m. 2 cisplatin on days 1 to 3, 120 mg./m. 2 etoposide on days 1 to 3 and 90 mg. bleomycin, table 1). Whenever technically possible, even in patients with residual large tumors, a secondary operation was performed after chemotherapy cycle 4 if the tumor markers had normalized or at least remained at a steady state. In 13 patients more than 4 cycles were necessary before the tumor markers were normal or had reached a steady state. Four patients had elevated afetoprotein before a secondary operation (43, 70, 97 and 169 µg./1., respectively), 4 had elevated serum human chorionic gonadotropin (11, 13, 14 and 500 units per 1., respectively) and 1 had elevated a-fetoprotein (27 µg./1.) and human chorionic gonadotropin (17 units per 1.) levels. The routine was to perform bilateral retroperitoneal lymph node dissection in all patients who initially had retroperitoneal tumors but not in patients who at diagnosis had a normal abdominal CT scan. A thoracic operation (resection of residual lung densities) was done in 19 patients with lung densities remaining after chemotherapy. Eleven patients underwent a retroperitoneal and thoracic post-chemotherapy operation. Ten patients did not undergo an operation at all due to death or complications before cycle 4 in 4, no retroperitoneal tumor at the start of chemotherapy in 3, irradiation due to initial classification as pure seminoma in 2 and development of brain metastases in 1. Statistics. Medians and ranges were calculated with the statistical package Medlog, 19 which also was used to perform significance testing (chi-square and Wilcoxon tests) and to perform the logistic regression analysis for predictive factors.

Early death* Alive, no evidence of disease Alive with tumor Dead of Ca Dead with Cat Recurrent disease

3 4 2 1 2

48 4

35 1 1

5

4

5

5

7

3 94 1 12 1 16

* Before completion of cycle 2. t One patient died of chemotherapy complications after cycle 3.

Crude survival was calculated by the Kaplan-Meier method. The median observation time was 55 months (range 0.5 to 102 months). A p value ofless than 0.05 was regarded as statistically significant. RESULTS

The 5-year crude survival for all 111 patients was 86%. A total of 3 patients died before completion of cycle 2 (2 died of treatment complications and 1 died of advanced cancer). One patient died of treatment-resistant septicemia after cycle 3. Twelve patients treated according to the scheduled plan died of recurrent testicular cancer. Completely necrotic or fibrotic tumor tissue was found in 52 of 101 patients undergoing an operation after chemotherapy (in patients undergoing 2 operations the worst histological result was considered, table 2). A total of 37 patients had mature teratoma and in 12 vital malignant tumor was found. Only 4 of the 16 patients with relapse had recurrence in the retroperitoneal space. In 3 of these 4 patients the post-chemotherapy histological specimens showed vital malignant tumor, while in 1 mature teratoma was demonstrated. The rate of recurrence was significantly less in patients who after chemotherapy had completely necrotic and/or fibrotic tumors (4 of 52) compared to patients with residual vital malignant tumor (5 of 12). Of 37 patients with mature teratoma in the operative specimen 5 subsequently had relapse. Of the 101 patients undergoing a secondary operation 92 had retroperitoneal lymph node dissection. In a univariate analysis several factors were predictive for complete necrosis and/or fibrosis in the retroperitoneal lymph node dissection specimen (table 3), including absence of teratomatous elements in the primary tumor histological specimen, normal preoperative serum tumor markers, complete response of the retroperitoneal tumor as evaluated on the post-chemotherapy CT scan, low levels of pre-chemotherapy a-fetoprotein and human chorionic gonadotropin, and the initial size of the retroperitoneal tumor. In a logistic regression analysis only the presence of teratomatous elements in the primary tumor histological specimen and complete response on post-chemotherapy CT remained independent significant predictive factors for post-chemotherapy complete necrosis and/or fibrosis. The demonstration of normal preoperative tumor markers was not significant (p = 0.13). Table 4 shows the predictive ability of the primary tumor histological status combined with the CT-visible response. The addition of the criterion of normal preoperative tumor markers increased the specificity from 69 to 76%, and decreased the false positive and false negative rates by 4 and 2%, respectively. DISCUSSION

When this treatment policy was begun in 1980 it was decided that all patients with initial retroperitoneal tumors should undergo post-chemotherapy retroperitoneal lymph node dissection, even those who achieved a radiological complete response during chemotherapy. Several institutions have since changed the policy and no longer operate on patients with a complete response. Histological examination most often reveals necrosis

1241

HISTOLOGY OF TUMOR RESIDUALS AFTER CHEMOTHERAPY TABLE

3. Predictive factors and post-chemotherapy histological findings in 92 patients undergoing retroperitoneal lymph node dissection (univariable analysis) Necrosis/ Fibrosis (47 pts.)

Teratomatous elements in the primary tumor histology: Present Absent Preop. tumor markers: Normal Elevated Complete response on abdominal CT: Yes No Median values before chemotherapy (range):t a-fetoprotein Human chorionic gonadotropin Median cm. 2 initial size of the retroperitoneal tumor (range)

Mature Teratoma (34 pts.)

Vital Malignant Tumor (11 pts.)

Totals (92 pts.)

p Value*

18 29

28 6

7 4

54 38

0.0002

45 1

30 4

7 4

82 9

0.003

20 27

28 6

10 1

58 34

0.0001

70 (5-81,000) 27 (5-805,000) 14 (2-361)

0.01 0.022 0.002

29 (5-81,000) 5 (5-288,000) 10 (2-234)

176 (5-9,190) 99 (5-805,000) 26 (2-200)

180 (5-65,600) 18 (5-124,000) 56 (6-361)

* Chi-square or ~il~oxon test,

co~paring the frequency of necrosis/fibrosis with the rate of mature teratoma plus vital malignant tumor. t Upper normal hm1ts a-fetoprotem less than 20 µg.jl. and human chorionic gonadotropin less than 10 units per I.

TABLE 4.

Prediction of necrosis/fibrosis in the post-chemotherapy retroperitoneal lymphadenectomy specimen

Sensitivity Specificity Correctly predicted False pos. rate False neg. rate

Teratomatous Elements and Complete Response(%)

Teratomatous Elements, Complete Response and Normal Preop. Markers

83 69 76 26 21

83 76 79 22 19

(%)

and/or fibrosis in these completely responding patients. 8 - 11 If completely necrotic and/or fibrotic masses are present no benefit is gained from surgical removal. However, if mature teratoma is present complete removal of residual is indicated. Mature teratoma potentially may be the origin of late recurrences, 20 as also noted in 1 of our patients who had a retroperitoneal relapse 8 years after retroperitoneal lymph node dissection, when a large mature teratoma was incompletely resected. It is questionable whether patients with residual vital tumor benefit from post-chemotherapy resection. Only 50% of these patients currently can be cured,2 1 as confirmed in our study. We found vital malignant tumor in 10% of our patients, which is in agreement with most reports. 8- 11 The exact percentage of necrosis and/or fibrosis may depend on the intensity of chemotherapy. Freiha and associates, who used a median of 5.2 cycles before resection, found residual malignant tumor in only 3% of their patients. 10 Elevated tumor markers occasionally may be observed in patients with mature teratoma8 and even in those with complete necrosis histologically. This may be due to an error in histological sampling, particularly relevant in larger residuals. Minute a-fetoprotein and human chorionic gonadotropin-producing tumor foci may easily be overlooked. Furthermore, since afetoprotein has a 5 to 7-day half-life several weeks must pass until high a-fetoprotein levels become normal in patients whose tumors have become necrotic during chemotherapy. This factor probably explains the slightly elevated serum a-fetoprotein levels in 1 of our patients with complete tumor necrosis. As also shown by others, 11 the frequency of vital malignant tumor residuals is increased significantly in patients with elevated post-chemotherapy tumor markers compared to those with mature teratoma, or necrosis and/or fibrosis. 11 Increasing serum tumor markers in particular indicate active malignant tumor. In the future such patients should not be operated on routinely but they should undergo prolonged and more intensive chemotherapy.

How well can the presence of necrosis and/or fibrosis be predicted preoperatively? It has been shown previously that a complete remission on a post-chemotherapy CT scan alone is not an adequate criterion, since a third of such patients had mature teratoma or vital malignant tumor in the post-chemotherapy operative specimen. 22 Additional criteria must be considered. Husband and associates suggested a decrease in the CT attenuation values to be an additional predictive factor of post-chemotherapy necrosis,2 3 a finding that could not be confirmed by Donohue 9 and Stomper24 and their associates. The h~st?logical statu~ ?f the primary tumor may represent a predictive factor addit10nally to volume reduction as also noted by Donohue and associates, 9 and Vugrin and Whitmore 25 and confirmed in our study. If there were no teratomatous el~ments in the primary tumor, the percentage of patients in whom necrosis could be predicted was 83 % among those with a postchemotherapy complete remission on CT. Adding the results of the preoperative tumor markers did not increase the sensitivity but it improved the specificity and the percentage of correctly predicted cases. With the 3 criteria (complete remission on post-chemotherapy CT, absence ofteratomatous elements in the primary tumor a~d normal preoperative tumor markers) 16 of our 47 patients with post-chemotherapy complete necrosis and/or fibrosis could have been spared retroperitoneal lymph node dissection independent of the initial size of the retroperitoneal tumor o; the level of pre-chemotherapeutic tumor markers. Patients who fulfill these 3 criteria can be observed without an elective operation if regular followup is feasible. However, the clinician should remember that approximately 20% of the nonoperated patients will have residual malignant tumor or mature teratoma, possibly representing the origin of a subsequent clinical relapse. 20 This means that patients who do not undergo postchemotherapy retroperitoneal lymph node dissection should have frequent and long-lasting followup, including abdominal CT. An alternative approach in these patients would be to perform more limited nerve-sparing retroperitoneal lymph node dissection techniques, 26 thus, obtaining a post-chemotherapy histological status with a minimal risk of dry ejaculation. REFERENCES

1. Birch, R., Williams, S., Cone, A., Einhorn, L., Roark, P., Turner,

S. and Greco, F. A.: Prognostic factors for favorable outcome in disseminated germ cell tumors. J. Clin. Oncol., 4: 400, 1986. 2. Medical Research Council Working Party on Testicular Tumours: Prognostic factors in advanced non-seminomatous germ-cell testic!1lar tumours: result of a multicentre study. Lancet, 1: 8, 1985. 3. Fossa, S. D., Aass, N. and Kaalhus, 0.: Testicular cancer in young Norwegians. J. Surg. Oncol., 39: 43, 1988.

1242

FOSSA AND ASSOCIATES

4. Stoter, G., Sylvester, R., Sleijfer, D. T., Ten Bokkel Huinink, W. W., Kaye, S. B., Jones, W. G., Van Oosterom, A. T., Vendrik, C. P., Spaander, P. and De Pauw, M.: Multivariate analysis of prognostic factors in patients with disseminated nonseminomatous testicular cancer: results from a European Organization for Research on Treatment of Cancer multiinstitutional phase III study. Cancer Res., 47: 2714, 1987. 5. Bosl, G. J., Geller, N. L., Vogelzang, N. J., Kennedy, B. J., Whitmore, W. F., Jr., Vugrin, D., Scher, H., Nisselbaum, J. and Golbey, R. B.: Multivariate analysis of prognostic variables in patients with metastatic testicular cancer. Cancer Res., 43: 3403, 1983. 6. Whitmore, W. F., Jr.: Surgical treatment of adult germinal testis tumors. Sem. Oncol;, 6: 55, 1979. 7. Fossa, S. S., Ous, S., Abyholm, T. and Loeb, M.: Post-treatment fertility in patients with testicular cancer. I. Influence of retroperitoneal lymph node dissection on ejaculatory potency. Brit. J. Urol., 57: 204, 1985. · 8. Hendry, W. F., Goldstraw, P., Husband, J. E., Barrett, A., McElwain, T. J. and Peckham, M. J.: Elective delayed excision of bulky para-aortic lymph node metastases in advanced nonseminoma germ cell tumours of testis. Brit. J. Urol., 53: 648, 1981. 9. Donohue, J.P., Rowland, R. G., Kopecky, K., Steidle, C. P., Greier, G., Ney, K. G., Einhorn, L., Williams, S. and Loehrer, P.: Correlation of computerized tomographic changes and histological findings in 80 patients having radical retroperitoneal lymph node dissection after chemotherapy for testis cancer. J. Urol., 137: 1176, 1987. 10. Freiha, F. S., Shortliffe, L. D., Rouse, R. V., Mark, J. B. D., Hannigan, J. F., Jr., Aston, D., Spaulding, J. T., Williams, R. D. and Torti, F. M.: The extent of surgery after chemotherapy for advanced germ cell tumors. J. Urol., 132: 915, 1984. 11. Carter, G. E., Lieskovsky, G., Skinner, D. G. and Daniels, J. R.: Reassessment of the role of adjunctive surgical therapy in the treatment of advanced germ cell tumors. J. Urol., 138: 1397, 1987. 12. Peckham, M. J., Barrett, A., McElwain, T. J., Hendry, W. F. and Raghavan, D.: Non-seminoma germ cell tumours (malignant teratoma) of testis. Results of treatment and an analysis of prognostic factors. Brit. J. Urol., 53: 162, 1981. 13. Fossa, S. D.: Svensk-norsk diagnose- og behandlingsopplegg for ikke-seminomer i testis (SWENOTECA). Tidsskr. Nor. Lregeforen, 102: 371, 1982. 14. Collins, D. H. and Pugh, R. C. B.: The pathology of testicular tumours. Brit. J. Urol., suppl., 36: 1, 1964. 15. Mostofi, F. K. and Sobin, L. H.: Histological typing of testis tumours. In: International Histological Classification of Tumours. Geneva: World Health Organization, No. 16, 1977. 16. Klepp, 0., Fossa, S. D., Ous, S., Lien, H., Stenwig, J. T., Abeler, V., Eliassen, G. and H0st, H.: Multi-modality treatment of advanced malignant germ cell tumours in males. I. Experience with cis-platinum-based combination chemotherapy. Scand. J. Urol. Nephrol., 18: 13, 1984. 17. Aass, N., Fossa, S. D., Otto, F. and Ose, T.: Acute subjective morbidity after cisplatin based combination chemotherapy in patients with testicular cancer: a prospective study. Radiother. Oncol., 14: 27, 1989. 18. Hall, K. S., Fossa, S. D. and Aas, M.: High-dose cis-platinum combination chemotherapy in advanced nonseminomatous malignant germ cell tumours with emphasis on nephrotoxicity. Cancer Chemother. Pharm., 18: 74, 1986. 19. Layard, M. W. and McShane, D. J.: Applications of medlog, a microcomputer-based system for time-oriented clinical data. In: Proceedings of the Seventh Annual Symposium on Computer Applications in Medical Care. Silver Spring, Maryland: IEEE Computer Society Press, p. 731, 1983. 20. Loehrer, P. J., Sr., Hui, S., Clark, S., Seal, M., Einhorn, L. H., Williams, S. D., Ulbright, T., Mandelbaum, I., Rowland, R. and Donohue, J.P.: Teratoma following cisplatin-based combination chemotherapy for nonseminomatous germ cell tumors: a clinicopathological correlation. J. Urol., 135: 1183, 1986.

21. Hendry, W. F., Goldstraw, P. and Peckham, M. J.: The role of surgery in the combined management of metastases from malignant teratomas of testis. Brit. J. Urol., 59: 358, 1987. 22. Fossa, S. D., Ous, S., Lien, H. H. and Stenwig, A. E.: Postchemotherapy lymph node histology in radiologically normal patients with metastatic nonseminomatous testicular cancer. J. Urol., 141: 557, 198fl 23. Husband, J.E., Hawkes, D. J. and Peckham, M. J.: CT estimations of mean attenuation values and volume in testicular tumors: a comparison with surgical and histologic findings. Radiology, 144: 553, 1982. 24. Stomper, P. C., Jochelson, M. S., Garnick, M. B. and Richie, J.P.: Residual abdominal masses after chemotherapy for nonseminomatous testicular cancer: correlation of CT and histology. Amer. J. Roentgen., 145: 743, 1985. 25. Vugrin, D. and Whitmore, W. F., Jr.: The role of chemotherapy and surgery in the treatment of retroperitoneal metastases in advanced nonseminomatous testis cancer. Cancer, 55: 1874, 1985. 26. Donohue, J. P., Rowland, R. G. and Bihrle, R.: Transabdominal retroperitoneal lymph node dissection. In: Diagnosis and Management of Genitourinary Cancer. Edited by D. G. Skinner and G. Lieskovsky. Philadelphia: W. B. Saunders Co., chapt. 55, pp. 802-816, 1988.

EDITORIAL COMMENTS This retrospective review has focused upon identification of the criteria that can predict necrosis or fibrosis in the retroperitoneal lymph node dissection specimen. It is not surprising that patients with elevated post-chemotherapy tumor markers had a higher incidence of residual malignancy than those whose tumor markers had normalized Our practice for many years has been not to recommend retroperitoneal lymph node dissection if the tumor markers still are elevated but to continue with definitive chemotherapy until the markers have normalized. The authors appropriately come to this same conclusion. The question remains whether absence of teratoma in the primary tumor and complete response on a CT scan are sufficient predictive factors to obviate the need for retroperitoneal lymph node dissection. The CT scan does tend to understage or under-appreciate the extent of involvement, especially when there has been a large tumor before chemotherapy. In such circumstances critical reading of the CT scan is necessary so as not to overlook potential residual disease. It should be remembered that the correct prediction rate reported by the authors is only 79% in patients who meet all 3 criteria in the study. In general, most patients who have a large retroperitoneal mass before chemotherapy still will benefit from retroperitoneal lymph node dissection. In selected patients avoidance of a retroperitoneal lymph node dissection may be feasible but one must weigh the risks of missing residual teratoma and/or residual cancer in patients in whom observation is elected after chemotherapy. Jerome P. Richie Department of Urology Brigham-Women's Hospital Boston, Massachusetts This is a clear statement confirming the reasonable position of observation in good risk patients who have achieved a complete remission after chemotherapy for clinically advanced testis cancer. The authors have analyzed favorable prognostic features predictive of benign histological status in the retroperitoneal space, also reported in other large experiences. Such studies permit consideration of the-concept of individualization of therapy for patients with advanced testis cancer based upon variables, such as primary tumor histology, rate and completeness of response.

John P. Donohue Department of Urology Indiana University Medical Center Indianapolis, Indiana