A Prospective Study of Cisplatin-Based Combination Chemotherapy in Advanced Germ Cell Malignancy: Role of Maintenance and Long-Term Follow-Up

Transcription

1 A Prospective Study of Cisplatin-Based Combination Chemotherapy in Advanced Germ Cell Malignancy: Role of Maintenance and Long-Term Follow-Up By John A. Levi, Damien Thomson, Tom Sandeman, Martin Tattersall, Derek Raghavan, Michael Byrne, Grantley Gill, Vernon Harvey, Ivon Burns, and Raymond Snyder for the Australasian Germ Cell Trial Group Two hundred fifty-three patients with advanced germ cell malignancy received initial chemotherapy with cisplatin, vinblastine, and bleomycin followed by surgical resection of residual masses if possible. Patients achieving complete remission (CR) were prospectively randomized to receive 6 months maintenance therapy with vinblastine or no further treatment. CR was achieved in 183 patients (72%) and a further eight patients (4%) had complete resection of residual viable malignancy (no evidence of disease [NED]). Pretreatment factors having a significant adverse influence on response by univariate analysis included extragonadal origin of the tumor, poor performance status, advanced lung or lung and abdominal disease, and elevated serum levels of human chorionic gonadotropin (HCG) and alpha-fetoprotein (AFP) > 1,000 ng/ml. Multivariate regression analysis indicated the independent prognostic factors of significance were advanced lung or advanced lung and abdominal disease, total tumor diameter > 10 cm, and a serum level of HCG > 1,000 ng/ml. Of the toxicities encountered, myelosuppression was significant, being exacerbated by radiotherapy, and seven deaths occurred from septicemia. Bleomycin pulmonary toxicity occurred in 46% of patients and was severe in 4%, resulting in eight deaths. With a median followup of 64 months, relapses have occurred in 25 patients with no significant difference between those patients receiving or not receiving maintenance vinblastine. Eight of these relapses occurred beyond I year and four beyond 2 years of follow-up. Presently, 68% of the total patient population is alive and disease-free, with 84% of the CR and NED patients alive and 81% alive and disease-free. It is concluded that with prolonged follow-up, vinblastine maintenance therapy does not improve treatment outcome. Moreover, late relapses occur, cautioning against premature pronouncements of cure. J Clin Oncol 1988 by American Society of Clinical Oncology. T IS NOW well established that major improvements in outcome for patients with advanced germ cell malignancies may be expected with cisplatin-based combination chemotherapy. Complete remission (CR) rates of 70% to 80% and relapse rates of 8% to 16% with 2 to 3 years follow-up have been regularly reported. 1 2 An additional 8% to 10% of patients with incomplete tumor regression are rendered disease-free by surgical resection of residual disease, with survival comparable to those achieving CR on che- From the Royal North Shore Hospital of Sydney; Princess Alexandra Hospital, Brisbane; Peter MacCallum Clinic, Melbourne; Royal Prince Alfred Hospital, Sydney; Sir Charles Gairdner Hospital, Perth; Royal Adelaide Hospital, Adelaide; Auckland Hospital, Auckland; and St. Vincents Hospital, Melbourne, Australia. See Appendix for other participating institutions and principal investigators. Submitted September 24, 1987; accepted January 26, Address reprint requests to John A. Levi, MD, Department of Clinical Oncology, Royal North Shore Hospital of Sydney, St. Leonards, 2065, NSW, Australia by American Society of Clinical Oncology X/88/ $3.00/0 motherapy alone. 3 ' 4 Furthermore, one randomized trial with a minimum follow-up of 1 year and several nonrandomized studies have reported that maintenance therapy may be unnecessary." The vast majority of relapses following chemotherapy occur within 2 years, and it is frequently assumed that long-term survival can be confidently predicted with this length of followup.'' 5 However, recent data suggest that late relapses do occur, thereby indicating a need for caution in statements of curability with these treatment programs. 2,8 In the present multiinstitutional prospective study, 253 patients with advanced-stage germ cell malignancy received a cisplatin-based combination chemotherapy regimen and patients achieving CR were randomized to maintenance therapy or no further treatment. The median duration of follow-up is in excess of 5 years, and the results of this mature study are now reported. PATIENTS AND METHODS From May 1979 to February 1983, 260 patients from 23 institutions in Australia and New Zealand were entered onto 1154 Journal of Clinical Oncology, Vol 6, No 7 (July), 1988: pp

2 CHEMOTHERAPY IN ADVANCED GERM CELL MALIGNANCY 1155 study. All patients had a proven diagnosis of germ cell carcinoma and had either inoperable stage II or stage III disease. Histopathological review was undertaken by a central panel in 79% of cases. None of the patients had received prior chemotherapy with any of the study agents. Before each cycle of chemotherapy, all patients had a complete history and physical examination, chest x-ray, full blood count including differential and platelets, biochemical profile, creatinine and creatinine clearance, serum alpha-fetoprotein (AFP), and human chorionic gonadotropin (HCG). Pulmonary function studies including vital capacity and diffusion capacity of carbon monoxide (DLCO) were performed when possible and in all patients after the first 100 patients entered on trial. Retroperitoneal tumor was evaluated every 6 weeks by computerized tomography (CT). Thoracic CT, intravenous (IV) pyelogram, lymphangiography, and radionuclide scans were performed as necessary. Audiograms were obtained pretreatment and after alternate cycles of chemotherapy whenever possible and when clinically indicated. All patients received induction chemotherapy with cisplatin, 100 mg/m 2 on day 1 as a two-hour infusion accompanied by IV hydration and mannitol diuresis, vinblastine, 6 mg/m 2 days 1I and 2 by IV injection, and bleomycin, 30 mg by intramuscular (IM) injection on day 1. Cisplatin and vinblastine were repeated at three weekly intervals and bleomycin administered weekly for a maximum of 12 weeks (360 mg). Courses were delayed by 1I week if the WBC count was <3,000/pL and/or the platelet count <100,000/pL. Doses of vinblastine were modified according to the degree of myelosuppression or if significant peripheral and/or autonomic neuropathy developed. If the serum creatinine level increased to > 0.20 mmol/l, further doses of cisplatin were withheld until renal function improved. If the DLCO decreased by > 25% baseline levels or if radiographic evidence of pulmonary toxicity developed, bleomycin was withdrawn from treatment. Bleomycin doses were also modified if moderate or severe mucosal toxicity occurred. The number of courses of induction chemotherapy administered was determined by clinical and biochemical evidence of tumor response without any empirical limit. Patients were reevaluated clinically, with tumor markers and radiographically as indicated after every course. If CR was achieved as defined by complete disappearance of all evidence of disease and return to normal of elevated tumor markers for at least 1 month, then two further courses of induction chemotherapy were administered before consideration of maintenance therapy. Patients with residual disease, providing previously elevated tumor markers had returned to normal for at least I month, underwent surgical exploration wherever possible and resection of all residual abnormalities was attempted. If the resected specimens contained fibrosis or benign teratoma, the patients were considered to be in CR and eligible for randomization. If viable carcinoma was found and considered completely removed, the patients were classified as having no evidence of disease (NED), received two further courses of induction chemotherapy, and were observed thereafter. Patients in whom surgical resection of residual disease was not undertaken or was incomplete, and who had continued elevation of tumor markers or obvious progressive disease while on induction chemotherapy were classified as incomplete responders and removed from study. Patients who achieved CR, with or without surgery, were allocated by a random number system to receive either no further treatment or 6 months maintenance chemotherapy with vinblastine, 10 mg/m 2 by IV injection at 28-day intervals for a total of six courses. The definitions of CR, NED, and incomplete response used in this study are in general usage and indicated above. Relapse was defined as evidence of recurrent disease clinically or radiographically or if a persistent increase in HCG and AFP by more than 1 log above normal values occurred after CR of NED. Toxicities encountered with therapy were classified according to World Health Organization (WHO) criteria. 9 For statistical purposes, the durations of response and survival were calculated from the date of initiation of chemotherapy until relapse, death, or the date of last follow-up. Response data were compared by chi-square analysis. Durations of response and survival were compared by log rank analysis and survival curves constructed according to the method of Kaplan-Meier. 10 The Cox method of multivariate regression analysis was used to determine the influence of individual prognostic variables on outcome. RESULTS Pretreatment characteristics of the patient population are indicated in Table 1. Of the 260 patients who entered the study, seven were inevaluable: four had no follow-up after entry, two were major protocol violations receiving different chemotherapy, and one patient proved not to have germ cell carcinoma on histological review. Response Rates Of the 253 evaluable patients, 183 (72%) achieved CR. Surgical resection of residual masses was performed in 63 of these patients (34%) and revealed necrosis and fibrosis in 37 (20%) and benign teratoma in 26 (14%). In a further eight patients viable carcinoma within the residual masses was completely resected, render- Table 1. Total patient entry No. of evaluable patients Median age (range) Origin of primary tumour Testicular Ovarian Extragonadal Prior therapy None/surgical reduction Radiotherapy Chemotherapy Performance status* *ECOG criteria. Pretreatment Characteristics (15-65)

3 1156 ing these patients NED. Thus, the combined CR plus NED rate was 76%. The median number of chemotherapy courses administered to all patients was four (range, three to ten). Of the 39 patients receiving five or more courses, 82% entered CR with NED, compared with 71% of 214 patients receiving fewer courses. There was no significant difference in the proportion of patients entering CR with NED irrespective of the number of courses required (P =.16). Prognostic Factors Univariate analysis showed that various pretreatment characteristics influenced response (Table 2). These included extragonadal origin of the primary tumor, poor performance status (Eastern Cooperative Oncology Group [ECOGI 2 to 4), and extensive disease as defined by Samuels' criteria, 12 involving lung or both lung and abdomen (all P =.0001). Patients with advanced abdominal disease only did not have a significantly reduced probability of achieving CR plus NED. Prior radiotherapy and the histological type of the tumor were not significant variables. However, the small number of patients with pure yolk sac tumors (five) and pure choriocarcinomas (six) achieved CR with NED in only one instance each. There was no significant difference in response for those patients with an elevated pretreatment level of either HCG or AFP compared with those with levels within normal limits. However, when analysis was undertaken of levels of either HCG or AFP > 1,000 ng/ml, then these patients had a significantly worse outcome (P =.0001 and P =.003, respectively). When these various factors were evaluated by Cox's multivariate regression analysis to determine the independent contribution of each factor to remission potential, extent of disease proved to be the most influential, with advanced lung and combined advanced lung and abdomen being the dominant adverse factors (P =.0006). Total tumor size, defined by summing the maximum diameters of all sites of measurable tumor, was also significant with total tumor diameters > 10 cm reducing remission probability (P =.01). The only other pretreatment characteristic that had an independent significant adverse influence on remission was an initial level of HCG > 1,000 Table 2. Factor Site of primary Testicular Extragonadal Ovary Prior therapy None/surgery Radiotherapy Chemotherapy Performance status Histology* T±E±C±Y Y±E ±T±S C±E±T Extent of disease Elevated markers only Cervival nodes only ML±MA MA AA ± ML AL ± MA AA ± AL Bone Marker status HCG AFP LEVI ET AL Influence of Pretreatment Factors on Response No. CR+NED(%) 180 (81) 6 (26) 5 (71) 167 (77) 24 (73) 0 (0) (81) (47) (33) (81) (73) (91) (53) (50) 11 (92) 4 (100) 24 (89) 19 (100) 96 (77) 16 (50) 20 (50) 1 (33) Normal 82 (76) Elevated 105 (75) Normal 72 (78) Elevated 118 (74) *E, embryonal; T, teratocarcinoma; S, seminoma; C, choriocarcinoma; Y, yolk sac. Only those cases with full histological review were included. tusing Samuels' criteria: ML, minimal lung; MA, minimal abdomen; AA, advanced abdomen; AL, advanced lung. ng/ml (P =.004). Other prognostic factors on univariate analysis, including extragonadal origin of the primary tumor, poor performance status, and initial level of AFP, were not independent prognostic factors with the multivariate regression model. Toxicities Toxicities encountered with this program of induction chemotherapy were considerable. Leukopenia was the principal hematological toxicity, with nadir WBC counts of < 3,000/LL oc-

4 CHEMOTHERAPY IN ADVANCED GERM CELL MALIGNANCY 1157 curring in 81% of patients and < 1,000//L in 23%. Seven deaths from septicemia occurred in patients with leukopenia. Chronic normocytic, normochromic anemia (hemoglobin < 10 g/dl) developed in 49% of patients, particularly those who received more than four courses of induction therapy. Thrombocytopenia (platelet nadir < 100,000/gL) occurred in 32% of patients, but no significant hemorrhagic episodes occurred. These hematological toxicities were all significantly worse in patients who had received prior radiotherapy compared with those who had only surgery (P =.01). Table 3 indicates the incidence and severity of other toxicities accompanying induction chemotherapy. As expected, moderate to severe nausea and vomiting occurred in 97% of patients and was not well controlled with antiemetics. While skin toxicity was not a major concern, at least six patients developed Raynaud's syndrome as a treatment sequela. Peripheral and autonomic neuropathy were reported in 48% of patients and tended to be a cumulative phenomenon. Renal dysfunction occurred in 32%, but was severe (serum creatinine > 0.40 mmol/l) in only 3% of patients and in all instances occurred in conjunction with septicemia and shock. The other major toxicity was bleomycin-related pulmonary toxicity, which occurred in 46% of patients, was severe in 4% (nine patients), and the direct cause of death in eight patients. Seven of these deaths occurred in the first 100 patients on study. By contrast in the latter 153 patients on study, administration of bleomycin was ceased when serial pulmonary function studies showed a > 25% decrease in DLCO, and only one death occurred. In relation to these toxicities, reductions in the dose or withdrawal of the drug from the treatment program was necessary with cisplatin in 8% of patients, in 37% with vinblastine, and in 35% of patients with bleomycin. There was no apparent effect of these dose modifications on the CR plus NED rate. Maintenance Therapy Of the 183 patients who achieved CR and who were eligible for randomization to maintenance therapy, 88 were randomized (43 to maintenance therapy and 45 to no maintenance). For various reasons, most often patient refusal, the remaining 95 patients were not randomized but were followed regularly without further treatment. Of the 43 patients randomized to maintenance vinblastine, 16 did not complete the full six courses, seven because of relapse and nine because of unacceptable toxicity. At a median follow-up of 64 months, relapses have occurred in 11 patients who received maintenance (26%), seven patients randomized to no maintenance (16%), and seven patients not randomized but just followed (7%). These relapse rates were not significantly different (P =.08). Figure 1 illustrates the time to relapse of the patient population and shows that 81% of the 191 patients achieving CR with NED and 61% of the total patients treated are alive and disease-free. No pretreatment factor or the number of induction chemotherapy courses administered significantly influenced the relapse statistics. Twenty-five CR patients have relapsed, while none of the eight NED patients have done Table 3. Nonhematological Toxicity With Induction Chemotherapy No. of Toxicity Grade Patients (as a percent of patients)* Toxicity Evaluable Nausea/vomiting Stomatitis Skin Alopecia Neuropathy Auditory Renal Pulmonary *WHO criteria. 0 P _x Relapse Time (Months) Fig 1. Time to relapse. A, CR and NED (N = 191); 0, total patient population (N = 253); B, incomplete responders (N = 62). Log rank analysis P<

5 1158 so. Of these relapses, eight (32%) occurred at > 1 year and four (16%) at > 2 years (28, 29, 38, and 60 months) posttreatment. Of the eight relapses after 1 year, four patients had received maintenance therapy and four had not. Survival For the duration of follow-up, no patient has developed a second testicular carcinoma in the contralateral testis and only one other malignancy occurred in a patient who developed acute myeloblastic leukemia 6 months after the initial diagnosis of metastatic testicular cancer. Figure 2 shows the survival curves for the patient population; 68% of all patients treated (172) are long-term survivors and 84% of patients (161) entering CR plus NED are alive. Six of the 11 surviving patients who had an incomplete response to induction chemotherapy had a subsequent CR with second-line chemotherapy; but five received no further therapy, suggesting that the residual abnormalities did not contain viable malignancy. Of the eight patients relapsing after 12 months, three are alive and presently disease-free (two of these relapsed after 24 months) achieving a second CR with salvage chemotherapy. DISCUSSION Response Rates In this large study, 76% of patients achieved CR and NED with a cisplatin-based combination chemotherapy program and resection of residual masses wherever possible, confirming reports Survival Time (Months) Fig 2. Duration of survival. A, CR and NED (N = 191); 0, total patient population (N = 253); B, incomplete responders (N = 62). Log rank analysis P= LEVI ET AL from other generally smaller studies that this combined modality approach is highly successful in the treatment of advanced-stage germ cell malignancies.1,4,13 It has been stated that four courses of induction chemotherapy is optimal in this setting and further treatment is unlikely to increase the proportion of remissions. Our data do not support this contention, as the remission rate was 82% in patients who required five or more courses to achieve remission and their outcome was comparable to those requiring fewer courses. This emphasizes the point that induction chemotherapy should continue in responding patients until maximum response is achieved without restriction to an arbitrary number of courses. Newlands et al also noted similar findings in their studies. 14 The fact that only eight patients (4%) had evidence of viable malignancy and were rendered disease-free with surgery postchemotherapy, a figure somewhat lower than that reported in other series, probably reflects the prolongation of chemotherapy and the concomitantly reduced need for major surgery in a number of patients."' 3 " 1 Prognostic Factors In this study the pretreatment factors that proved to be independently significant in determining remission potential were the extent of disease-patients with advanced lung or combined advanced lung and abdominal disease having the worst prognosis-a total tumor size in excess of 10 cm, and an initial level of HCG > 1,000 ng/ml. No differences were observed among the small number of patients with ovarian origin of tumor when compared with origin from the testis. It is of some interest that tumors arising from extragonadal sites did not emerge as an independent factor, presumably because these patients often presented with very advanced disease, which is the principal determinant of subsequent outcome. These findings are essentially in accord with other reports. 1 "' 8 This consistency among several studies indicates that tumor volume and levels of tumor markers are the principal determinants in defining good and bad prognostic groups at presentation. This stratification of patients is now being used to develop newer programs of therapy aimed at improving outcome in the bad-prognosis patients with a more aggressive approach and improving the therapeutic in-

6 CHEMOTHERAPY IN ADVANCED GERM CELL MALIGNANCY 1159 dex while maintaining the excellent results in the "good-prognosis" categories Toxicities The spectrum and degree of toxicities encountered in this study are generally in accord with that expected of high-dose cisplatin-based combinations for advanced germ cell carcinoma. 2, 7 ' 5 Myelosuppression was exaggerated in patients who had received prior radiotherapy, although this did not compromise the potential for remission. Nevertheless, seven patients died from septicemia while leukopenic (four having prior radiotherapy), emphasizing the risk associated with prior therapy that compromises bone marrow function. The current "watch policy" studies in early stage testicular cancer, if successful, may substantially reduce the number of previously irradiated patients who require subsequent chemotherapy. The 48% incidence of peripheral and autonomic neuropathy in this series, which was moderate to severe in 18%, was frequently troublesome because of its prolonged duration. Both cisplatin and vinblastine probably contributed to this complication and the substitution of etoposide for vinblastine in some recent studies may reduce this problem. 20, 2 Of particular concern in this study was the high incidence of bleomycin-induced pulmonary toxicity, with eight deaths occurring. This complication was more severe during the earlier phases of the study. When serial pulmonary function studies including measurements of DLCO were obligatory, withdrawing bleomycin if the DLCO decreased by > 25%, the extent of the toxicity was markedly reduced. Nevertheless, this potentially fatal toxicity of bleomycin together with a significant potential for skin and mucosal toxicity, all of which have been shown to be worse when combined with cisplatin, make it desirable to omit this agent, providing treatment efficacy is not compromised. 23 Preliminary results of current studies by our group suggest that bleomycin is unnecessary, at least for the treatment of good-prognosis patients.21 Maintenance Therapy and Survival The data available from this mature study convincingly show that vinblastine maintenance therapy does not improve the outcome in advanced-stage germ cell carcinomas treated with cisplatin-based therapy. With a median followup of > 5 years, maintenance vinblastine did not significantly influence the number of relapses or relapses beyond 1 and 2 years. Nevertheless, our study shows that late relapses can occur, cautioning against claims of cure when follow-up is of only 1 to 2 years' duration. In fact, isolated relapses occurred up to 5 years after therapy, emphasizing the need to maintain indefinite followup of these patients. When relapses occur salvage chemotherapy has proven to be effective when the extent of disease is limited and can result in subsequent long-term survival. Analysis of patients who relapsed has failed to show that any pretreatment factor, number of courses of induction therapy, or findings at surgery, be it fibrosis or benign teratoma, significantly influenced relapse potential. In conclusion, this study has shown that despite the success of this combined chemotherapy and surgical approach in advanced germ cell carcinoma, toxicity remains a major concern and certain independent prognostic factors including extent of disease and initial plasma level of HCG are predictive of likely outcome. Newer treatment programs are being evaluated using these findings to lessen toxicity in those patients with good prognostic characteristics and increase efficacy for the poor-prognosis patients. Maintenance therapy has not been proven useful, but prolonged follow-up is important to ensure maximum potential to salvage any patients with late relapse. APPENDIX Other institutions and principal investigators of the Australasian Germ Cell Trial Group 1979 to 1983 include Alan Gray, Wellington Hospital, Wellington, New Zealand; Albert Freedman, Prince of Wales Hospital, Sydney, Australia; Keiren Phadke, St. George Hospital, Sydney, Australia; Max Schwarz, Alfred Hospital, Melbourne; David Dalley, St. Vincents Hospital, Sydney, Australia; Bert Sundstrup, Peter MacCallum Clinic, Launceston, Australia; Ken MacMillan, Peter MacCallum Clinic, Hobart, Australia; Ray Lowenthal, Royal Hobart Hospital, Hobart, Australia; Robert Woods, Concord Repatriation Hospital, Sydney, Australia; Geoff Hawson, Prince Charles Hospital, Brisbane, Australia, Pathology Review Panel included Richard Hollis, Royal North Shore Hospital, Sydney, Australia; John Bell, Princess Alexandra Hospital, Brisbane, Australia; and Richard Sinclair, Peter MacCallum Clinic, Melbourne.

7 1160 LEVI ET AL 1. Einhorn LH: Testicular cancer as a model for a curable neoplasm: The Richard and Hinda Rosenthal Foundation Award Lecture. Cancer Res 41: , Bosl GJ, Gluckman R, Geller NL, et al: VAB-6: An effective chemotherapy regimen for patients with germ-cell tumors. J Clin Oncol 4: , Stahel RA, Von Hochstetter AR, Largiader F, et al: Surgical resection of residual tumor after chemotherapy in non-seminomatous testicular cancer. Eur J Cancer Clin Oncol 18: , Bracken RB, Johnson DE, Frazier HO, et al: The role of surgery following chemotherapy in stage III germ cell neoplasms. J Urol 129:39-43, Einhorn LH, Williams SD, Troner M, et al: The role of maintenance therapy in disseminated testicular cancer. N Engl J Med 305: , Vugrin D, Whitmore WF, Golbey RB: VAB-6 combination chemotherapy without maintenance in treatment of disseminated cancer of the testis. Cancer 51: , Logothetis CJ, Samuels ML, Selig D, et al: Improved survival with cyclic chemotherapy for nonseminomatous germ cell tumors of the testis. J Clin Oncol 3: , Terebelo HR, Taylor HG, Brown A, et al: Late relapse of testicular cancer. J Clin Oncol 1: , Miller AB, Hoogstraten B, Stacquet M, et al: Reporting results of cancer treatment. Cancer 47: , Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53: , Cox DR: Analyses of binary data. London, Chapman and Hall, 1970, pp Samuels ML, Johnson DE, Holoye PY: Continuous intravenous bleomycin (NSC ) therapy with vinblastine (NSC 49842) in stage III testicular neoplasia. Cancer Chemother Rep 59: , Vugrin D, Whitmore WF, Sogani PE, et al: Combined chemotherapy and surgery in treatment of advanced germ cell tumours. Cancer 47: , 1981 REFERENCES 14. Newlands ES, Begent RHJ, Rustin GJS, et al: Further advanced in the management of malignant teratomas of the testis and other sites. Lancet 1: , Einhorn LH, Williams SD, Mandelbaum I, et al: Surgical resection in disseminated testicular cancer following chemotherapeutic cytoreduction. Cancer 48: , Bosl GJ, Geller NL, Cirrincione C, et al: Multivariate analysis of prognostic variables in patients with metastatic testicular cancer. Cancer Res 43: , Birch R, Williams S, Cone A, et al: Prognostic factors for favorable outcome in disseminated germ cell tumors. J Clin Oncol 4: , Medical Research Council Working Party on Testicular Tumors: Prognostic factors in advanced non-seminomatous testicular tumors: Results of a multicentre study. Lancet 1:8-11, Ozols RF, Deisseroth AB, Javadpour N, et al: Treatment of poor prognosis nonseminomatous testicular cancer with a "high dose" platinum combination chemotherapy regimen. Cancer 51: , Stoter G, Kaye S, Sleyfer D, et al: Preliminary results of BEP (bleomycin, etoposide, cisplatin) versus an alternating regimen of BEP and PVB (cisplatin, vinblastine, bleomycin) in high volume metastatic testicular non-seminomas: An EORTC study, Proc Am Soc Clin Oncol 5:106, 1986 (abstr) 21. Levi J, Raghavan D, Harvey V, et al: Deletion of bleomycin from therapy for good prognosis advanced testicular cancer: A prospective randomised study. Proc Am Soc Clin Oncol 5:97, 1986 (abstr) 22. 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: , Dalgleish AG, Woods RL, Levi JA: Bleomycin pulmonary toxicity: Its relationship to renal dysfunction. Med Pediatr Oncol 12: , 1984