Surveillance Alone Versus Radiotherapy After Orchiectomy for Clinical Stage I Nonseminomatous Testicular Cancer

Transcription

1 Surveillance Alone Versus Radiotherapy After Orchiectomy for Clinical Stage I Nonseminomatous Testicular Cancer By Mikael Rorth, Grethe Krag Jacobsen, Hans von der Maase, Ebbe Lindegdrd Madsen, Ole Steen Nielsen, Mogens Pedersen, Henrik Schultz, and the Danish Testicular Cancer Study Group From December 1980 to January 1984, all patients with stage I nonseminomatous testicular cancer in Denmark entered a randomized trial comparing surveillance only with radiotherapy after orchiectomy. One hundred fifty patients were assessable for the final analysis. Relapse occurred in 23 patients in the surveillance group and in 11 patients in the radiotherapy group. Radiotherapy completely prevented retroperitoneal relapse; 14 retroperitoneal relapses occurred in the surveillance-only group. All relapsing patients in the surveillance-only group are without evidence of disease with a median observation time after chemotherapy of 67 months. Two of the patients with relapse in the radiotherapy group died with disease; the others are alive without evidence of BEFORE 1980, the standard treatment in Denmark for patients with clinical stage I testicular cancer was high-voltage radiotherapy after orchiectomy. Patients with nonseminomatous germ cell tumors of the testicle (NSGCT) received approximately 40 Gy radiation to paraaortic and ipsilateral pelvic lymph nodes.' Before the introduction of cisplatin-based chemotherapy for salvage treatment of recurrences, the survival rates for clinical stage I NSGCT patients ranged from 70% to 90%. After the introduction of cisplatin, the survival rates have approached 100%.2 With the advent of successful chemotherapy and more sophisticated staging procedures, a treatment strategy of surveillance-only after orchiectomy was introduced, primarily in Great Britain. 3 The basic assumption was that the majority of patients with clinical stage I disease (60% to 80%) are cured by orchiectomy, and by careful surveillance, the other patients can be salvaged by applying chemotherapy at a relatively early stage of dissemination. In Denmark, staging, treatment, and follow-up of virtually all patients with testicular cancer were organized by a country-wide study group in In 1980 a randomized study was undertaken comparing the surveillance strategy with the standard radiotherapy of patients with clinical stage I NS- GCT. A report of the patients' histopathologic features and the possible prognostic significance disease, with a median observation time after relapse treatment of 72 months. In the surveillance group, four relapses occurred later than 2 years after orchiectomy; only one such late relapse occurred in the radiotherapy group. Four of the retroperitoneal relapses occurred without concomitant increase in the serum marker levels (alpha-fetoprotein [AFP] and human chorionic gonadotropin [HCG]). It is concluded that surveillance only should replace radiotherapy after orchiectomy as standard treatment for clinical stage I nonseminomatous testicular cancer. Improved methods for control of retroperitoneal relapses, especially of embryonal carcinomas, are needed. J Clin Oncol 9: by American Society of Clinical Oncology. of these features has been published separately. 5 We report here the clinical aspects of the trial. MATERIALS AND METHODS Virtually all patients with testicular cancer in Denmark are referred to one of five oncology centers for staging and treatment after orchiectomy. All five centers participated in the Danish Testicular Cancer Study Group trial; thus all clinical stage I NSGCT patients in Denmark were included in the trial, which took place from December 1980 to January After referral, the patients were staged according to standard procedures, including consecutive alpha-fetoprotein (AFP) and human chorionic gonadotropin (HCG) measurements in serum, computed tomographic (CT) scans of the abdomen, and bipedal lymphangiograms, as previously described. 4 Histology was originally reviewed by pathologists at the oncology centers, and the tumors were classified according to the World Health Organization criteria with slight modifications.' In mixed tumors, each tumor component was recorded separately. For the present From the Department of Oncology, Rigshospitalet, Copenhagen; Department of Oncology, Herlev Hospital, Herlev; Department of Oncology, Odense Hospital, Odense; Department of Oncology, Aalborg Hospital, Aalborg; Department of Oncology, Aarhus Kommunehospital, Aarhus; and the Institute of Pathological Anatomy, Gentofte Hospital, Gentofte, Denmark. Submitted June 20, 1990; accepted March 27, Address reprint requests to Mikael ROrth, MD, Department of Oncology 5074, Finsen Institute Rigshospitalet, 9 Blegdamsvej, DK-2100 Copenhagen, Denmark byamerican Society of Clinical Oncology X $3.00/0 Journal of Clinical Oncology, Vol 9, No 9 (September), 1991: pp

2 1544 evaluation, all tumor specimens except three were retrieved and reevaluated by a trained pathologist (G.K.J.), and a standardized recording of components and features was performed. 5 This reanalysis was performed blind; ie, the investigator was not informed about the treatment, allocation, or the clinical outcome. Clinical stage I was defined as no evidence of disease at lymphangiogram, CT scan, or chest x-ray. If tumor markers were elevated before orchiectomy, the decrease after orchiectomy should follow expected half-lives of less than 7 days for AFP and less than 2 days for HCG, and marker levels should be normalized before randomization. All patients were randomly assigned by a central closed envelope system. Substratification was not performed. Patients allocated to radiotherapy received 40 Gy in 25 fractions, five fractions per week to the paraaortic and ipsilateral pelvic lymph nodes. All patients were followed-up with monthly examinations the first year, every second or third month for the second year, and thereafter, at increasing intervals. Follow-up included chest roentgenograms and serum tumor markers. CT scan of the abdomen was repeated two to three times the first year and one to two times the second year. At most centers the standard follow-up for these patients lasted approximately 5 years after achievement of disease-free status. If the patients did not attend follow-up examinations, information concerning their health status was sought from general practitioners, local hospitals, or central registers. Relapse-free survival in the two treatment groups was analyzed by Kaplan-Meier plots and tested for statistical significant difference by the log-rank test. 7 A statistical difference ofp <.05 was regarded as significant. The risk of a type II error was not considered when the trial was planned; therefore, the study was not designed to meet a certain beta value. The clinical data in the present analysis were collected in November From 1980 to 1985, relapsing patients were treated with six series of cisplatin, vinblastine and bleomycin (PVB), according to the method of Einhorn and Donohue. 8 After that period, the standard relapse treatment was changed to four series of cisplatin, etoposide, and bleomycin (PEB), according to the method of Peckham et al. 9 Histopathologic features together with other variables, such as age, tumor size, and preorchiectomy levels of AFP and HCG, were analyzed for prognostic significance (ie, prediction of relapse) in a multiple regression model. The Cox proportional hazards model was used. RESULTS From December 1980 to January 1984, 156 patients entered the trial. Seventy-nine patients were allocated to surveillance only, and 77 to radiotherapy. At the reevaluation of tumor specimens, three specimens were lost and three tumors were found not to fulfill the criteria for NSGCT (two seminomas, one adenocarcinoma of rete testis), leaving 150 patients for the final evaluation. The radiotherapy group consisted of 73 patients, six of whom refused the treatment and RORTH ET AL were subjected to surveillance only. The surveillance group consisted of 77 patients. In the surveillance group one patient was lost to follow-up 28 months after orchiectomy. In the radiotherapy group one patient was lost to follow-up 11 months after orchiectomy. Both patients are known to be alive more than 5 years after orchiectomy and, according to information from local hospitals and general practitioners, neither one has shown any clinical signs of relapsing tumor. Median age and median observation time are given in Table 1. Details of histopathologic features of the primary tumors have been reported elsewhere. 5 Selected characteristics of the primary tumor of possible prognostic significance, ie, the presence of embryonal carcinoma (EC), seminoma (S), and/or yolk sac tumor (YST), are given in Table 1. Before orchiectomy, 53 of the patients in both groups had elevated AFP and/or HCG in serum. All patients were observed for more than 5 years. Median follow-up was 64 months in both groups. Altogether, three patients died of causes unrelated to testicular cancer or the treatment. One patient (age 64 years at orchiectomy) in the surveillance group died of pulmonary embolism 33 months after orchiectomy. In the radiotherapy group one patient (age 66 years at orchiectomy) with diabetes mellitus died of cardiovascular disease 44 months after orchiectomy, and one patient (age 25 years at orchiectomy) died of Romano- Ward syndrome (congenital cardiac arrythmia) 34 months after orchiectomy. None of the three Table 1. Selected Characteristics of the Study Population Radiotherapy Surveillance (n 73) (n =77) n % n % Primary tumor EC S YST Vascular invasion Preorchiectomy serum marker levels "+ AFP "+HCG Negative Unknown Median age, years (range) 30 (16-66) 30 (17-64) Median observation time, months (range) 64 (34-95) 64 (33-103)

3 SURVEILLANCE v RADIOTHERAPY POSTORCHIECTOMY patients experienced relapses, nor did autopsy in any case show malignant disease. Relapses Characteristics of the relapsing patients with regard to histology and marker levels at primary diagnosis are given in Table 2. Twenty-three patients relapsed in the surveillance group. Median time to relapse was 4 months (range, 2 to 62 months). Four relapses occurred later than 2 years after orchiectomy (25, 35, 47, 62 months, respectively). All four of these patients had elevated markers at relapse (two AFP, two HCG). In all four cases, the primary tumors were without vascular invasion: two had pure EC, one had EC plus endodermal sinus tumor, and one had EC plus S. One patient who relapsed after 25 months and went into complete remission after PVB chemotherapy had a second relapse 13 months later. Complete remission was achieved by PEB chemotherapy, and the patient has now been observed for 32 months with no evidence of disease. All relapses were treated with combination chemotherapy, and all patients are now without evidence of disease. Median observation time after relapse treatment was 67 months (range, 30 to 97 months). Sites of relapse are given in Table 3. Sixteen of the 23 relapses were accompanied by an increase in serum marker levels. In three of these cases, the preorchiectomy levels were normal. One patient with elevated preorchiectomy markers had a retroperitoneal relapse 3 months later with normal serum markers. Three other patients had retroperitoneal relapses without significant increase in the serum levels of AFP and/or HCG (2, 5, and 14 months after orchiectomy, respectively). Table 2. Selected Characteristics of the Relapsing Patients Radiotherapy Surveillance n % n % Primary tumor EC S YST Vascular invasion Preorchiectomy serum marker levels "+ AFP "+ HCG Negative Unknown Table 3. Sites of Relapse Radiotherapy Surveillance 1545 Marker- Marker- Total Positive Total Positive Retroperitoneum Lung Retroperitoneum + lung Mediastinum Inguinal nodes Local Markers only 3* 3 2t 2 *Three AFP. tone AFP, one AFP + HCG. Eleven patients relapsed in the radiotherapy group. Median time to relapse was 6 months (range, 5 to 42 months). One relapse occurred later than 2 years after orchiectomy (42 months). Three patients had a second relapse. One patient relapsed 9 months after the first relapse and treatment with PEB led to complete remission that has been sustained for 64 months. One patient had a second relapse 32 months after the first relapse. He went into complete remission with PEB. However, 41 months later a new relapse occurred and, despite intensive chemotherapy, the patient died of disseminated disease 12 months later. The patient with late relapse (42 months) had a second relapse 53 months later. The second relapse was accompanied by increasing AFP and a biopsy-verified EC in the retroperitoneal area. Because of alcoholic cirrhosis with ascites and severe intestinal radiation fibrosis, his general condition was very poor. He died shortly after initiation of PEB treatment. Autopsy was not performed. Altogether, nine of the 11 patients are alive without evidence of disease. Median observation time after the first relapse treatment is 72 months (range, 53 to 90 months). Sites of first relapse are listed in Table 3. There were no first relapses in the retroperitoneal area. Eight of 11 patients with relapse had an increase in one or two serum marker levels. Two of these had normal serum levels before orchiectomy. Three patients had marker-negative relapses in their lungs, and in all three cases, AFP and/or HCG were elevated in serum before orchiectomy. In the first year after orchiectomy, the difference in the relapse rate between the two groups (18 of 77 v 10 of 73) was not statistically significant. However, there was a significant reduction of

4 1546 retroperitoneal primary relapses in the radiotherapy group (zero of 73 v 14 of 77). Analysis by the log-rank test yielded a P value of.0506 (Mantel- Cox). The relapse-free survival curves are shown in Fig 1. The prognostic significance of age, tumor size, and tumor marker levels was analyzed by a multiple regression model. A Cox proportional hazards model was used. None of these variables showed any correlation with the risk of relapse. DISCUSSION As stated in several recent reviews,1' 0 11 the current management methods should cure virtually all patients with stage I NSGCT. In the present study, the standard radiotherapy treatment after orchiectomy used in Denmark before 1980 was compared with surveillance only. Radiotherapy effectively abolished retroperitoneal relapses and thereby reduced the overall relapse rate. Only one patient had a (second) relapse in the retroperitoneal area. Disease-related deaths occurred in this group of patients, and, since there is considerable long-term toxicity in patients treated with 40 Gy of high-voltage radiotherapy-- especially with respect to gastrointestinal problems,1 2 radiotherapy after orchiectomy should no longer be used as standard treatment for patients with stage I NSGCT. This conclusion is drawn on z 0 w r Y YEARS Fig 1. Relapse-free survival: ----, patients from the radiotherapy group (n = 73); -, patients from the surveillance group (n = 77). RORTH ET AL the basis of the excellent survival results of the surveillance strategy and of treatment with retroperitoneal lymph node dissection (RPLND). The relapse rate experienced in the surveillance group of this study, with an observation time of greater than 5 years, is exactly as would be predicted from the experience with RPLND. About 20% of RPLND patients with clinical stage I NSGCT have positive lymph nodes, and about 10% eventually have relapses outside the retroperitoneal area. Thus, 70% of the patients can be cured with orchiectomy alone. The remaining 30% are apparently cured with cisplatin-based chemotherapy. In the present study, only one patient had a second relapse. He was treated primarily with PVB and was later salvaged by PEB, with an observation time after the second relapse of 32 months. None of the patients in this group died due to testicular cancer. A major concern with the surveillance strategy is the possibility of late relapse in the retroperitoneal area, which did occur in this study. However, in most cases late relapse was accompanied by a concomitant increase in serum markers, rendering the diagnosis of relapse fairly easy. Nonetheless, marker-negative relapses in the retroperitoneal area also occurred, even in patients who were primarily marker-positive. Therefore, an efficient strategy for the diagnosis of retroperitoneal tumor is still a major problem. In our series, markernegative relapses have invariably been found in patients with EC components in the primary tumor. It is thus worthwhile to search for a serum marker for this component. Visualization of the retroperitoneal area is now routinely done by CT scan. This procedure is hampered by high costs and difficulties with obtaining concomitant biopsies, and its reliability is qualified by high rates of both false-positive and false-negative results. With a size criteria for positive nodes of 5 mm, the false-positive rate approaches 40% and with a size criteria of 15 mm, the false-negative rate exceeds 50%. This problem can be solved only by combining the imaging technique with a biopsy procedure. A comparison of CT scanning with ultrasonography in these patients should be done to establish if the less expensive and easier procedure of ultrasonography with concomitant biopsy could replace CT scan at follow-up. Based on our results, the occurrence of late relapses indicates that

5 SURVEILLANCE v RADIOTHERAPY POSTORCHIECTOMY follow-up examinations, at least in some patients, should be extended beyond the 5 years routinely used. The frequency and duration of follow-up should be subjected to cost-benefit analysis to define the optimal schedule. Another major concern has been patient compliance. 13 This concern is relevant with all types of management strategies including RPLND, although, admittedly, the problem is greater in surveillance-only studies. In the present series, comprising all patients in Denmark for a certain period of time, the problem occurred in two cases. Due to an elaborate health system and effective registration in Denmark, it was possible to keep track of these patients, although this is not an optimal way to conduct long-term follow-up. In other settings, this problem may be much greater. Follow-up after RPLND is believed to be easier and it is suggested that CT scan of the retroperitoneal area can be omitted. However, it should be remembered that the modified RPLND procedures designed to preserve fertility in these young patients should be followed by observation of the retroperitoneal area, as the modified procedure leaves some lymph nodes behind. The reported experience so far indicates that the relapse rate in the retroperitoneum for a well-performed modified dissection is very small.' 4 In view of the considerable number of relapses, it is appealing to consider a strategy with adjuvant chemotherapy. By applying early chemotherapy it is hoped that the overall intensity of the treatment could be reduced and the problem of lack of compliance overcome. Candidates for adjuvant therapy should be those with a high risk of relapse. In our study, 5 a high risk of relapse was significantly correlated to the presence of EC and vascular invasion in the primary tumor. In patients 1547 with EC, however, the presence of either YST or S components indicated a decreased risk of relapse. These findings are in agreement with those of several other groups."'" Pont et al"' administered two series of PEB to patients with EC and vascular invasion of the primary tumor (risk-adapted chemotherapy). Of 40 patients, 18 were initially treated with chemotherapy. Three relapses occurred, with a median follow-up of 27 months. The relapse rate in patients with vascular invasion in this (small) study was reduced to 11% (two of 18), compared with 31% in the present study (25 of 79). A similar trial is ongoing in the Medical Research Council in Great Britain. Although these trials are likely to give important information, they will not provide the definite answer to the question of ideal management of stage I NSGCT. The adjuvant strategy should be compared with a strategy in which chemotherapy is applied at first recurrence. According to recent results reported by Einhorn et al,' 9 three series of PEB may be sufficient to cure almost all relapsing patients. Therefore, the advantage of the adjuvant strategy is limited to the decrease in late toxicity when using two instead of three series of PEB and to the possibly improved compliance of the patients. The disadvantage will obviously be unnecessary cytotoxic treatment of patients who would not relapse anyway. In conclusion, in the process of improving the treatment of patients with stage I NSGCT, radiotherapy can no longer be recommended as part of the treatment strategy. Surveillance-only is a feasible strategy when performed at oncology centers. Improvements should be sought in the follow-up strategy, especially with regard to retroperitoneal relapses. The length of follow-up should be extended beyond 5 years. 1. Von der Maase H, Engelholm SA, Rorth M, et al: Non-seminomatous testicular germ cell tumours in Denmark Results of treatment. Acta Radiol Oncol 23: , Peckham MJ, Barrett A, McElwain TJ, et al: Combination management of malignant teratoma of the testis. Lancet 2: , Peckham MJ, Husband JE, Barrett A, et al: Orchidectomy alone in testicular stage I non-seminomatous germcell tumours. Lancet 2: , Schultz H, The Danish Testicular Carcinoma Study Group: Testicular carcinoma in Denmark Stage REFERENCES and selected clinical parameters at presentation. Acta Radiol Oncol 23: , Jacobsen GK, Rorth M, Osterlind K, et al: Histopathological features in stage I non-seminomatous testicular germ cell tumours correlated to relapse. APMIS 98: , Jacobsen GK, Barlebo H, Olsen J, et al: Testicular germ cell tumors in Denmark Pathology of 1058 consecutive cases. Acta Radiol Oncol 23: , Peto R, Rike MC, Armitage P, et al: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 35:1-47, 1977

6 Einhorn LH, Donohue JP: Cis-diaminodichloroplatinum, vinblastin and bleomycin combination chemotherapy in disseminated testicular cancer. Ann Intern Med 87: , Peckham MJ, Barrett A, Liew KH, et al: The treatment of metastatic germ cell testicular tumours with bleomycin, etoposide and cis-platin (BEP). Br J Cancer 47: , Gressler VH, Levine LA, Vogelzang NJ: A third option in the management of patients with clinical stage I nonseminomatous germ cell tumor? J Clin Oncol 8:4-8, Rorth M, Cullen MH, Horwich A, et al: Management of patients with non-seminomatous germ cell tumors stage I, in Newling DWW, Jones WG (eds): EORTC Genitourinary Group Monograph 7: Prostata Cancer and Testicular Cancer. New York, NY, Wiley, 1990, pp Hamilton C, Horwich A, Easton D, et al: Radiotherapy stage I seminoma testis: Results of treatment and complications. Radiother Oncol 6: , Pizzocaro G, Zanoni F, Salvioni R, et al: Difficulties of a surveillance study omitting retroperitoneal lymphadenectomy in clinical stage I nonseminomatous germ cell tumors of the testes. J Urol 138: , 1987 RORTH ET AL 14. Donohue JP: What are the surgical limits of retroperitoneal lymph adenectomy? Proceedings of the International Testicular and Prostatic Cancer Conference, Toronto, Canada, October 1990, pp Freedman LS, Jones WG, Peckham MJ, et al: Histopathology in the prediction of relapse of patients with stage I testicular teratoma treated by orchidectomy alone. Lancet 2: , Dewar JM, Spagnolo DV, Jamrozik KD, et al: Predicting relapse in stage I non-seminomatous germ cell tumor of the testis. Lancet 1:454, Fung CY, Kalish LA, Brodsky GL, et al: Staging nonseminomatous germ cell testicular tumor: Prediction of metastatic potential by primary histopathology. J Clin Oncol 6: , Pont J, H61tl W, Kosak D, et al: Risk-adapted treatment choice in stage I nonseminomatous testicular germ cell cancer by regarding vascular invasion in the primary tumor: A prospective trial. J Clin Oncol 8:16-20, Einhorn LH, Williams SD, Loehrer PJ, et al: Evaluation of optimal duration of chemotherapy in favorableprognosis disseminated germ cell tumors: A Southeastern Cancer Study Group protocol. J Clin Oncol 7: , 1989