Chemotherapy Options for Good-Risk Germ Cell Tumors: Focus on Post-Chemotherapy Retroperitoneal Histology

Chemotherapy Options for Good-Risk Germ Cell Tumors: Focus on Post-Chemotherapy Retroperitoneal Histology


Dr. Aditya Bagrodia, credit UT Southwestern

Dr. Darren R. Feldman

By Aditya Bagrodia, MD, and Darren R. Feldman, MD

Article Highlights

  • Given the excellent clinical outcomes in patients with International Germ Cell Cancer Collaborative Group good-risk germ cell tumors (GCT), research efforts have focused on minimizing treatment-related toxicity while maintaining or improving upon historic cure rates. 
  • The literature considering optimal treatment of patients with good-risk nonseminoma emphasizes that chemotherapy should be administered in an effective manner by an oncologist with expertise in this area and that the surgical expertise in performing PC-RPLND is critical to outcomes.
  • Longer-term follow-up data are needed to elucidate the differential impact of an additional cycle of etoposide and cisplatin (in patients receiving four-cycle EP) or of the addition of bleomycin (in patients receiving three-cycle BEP) on late adverse consequences such as cardiovascular toxicities and secondary malignancies in GCT survivors.

Patients with International Germ Cell Cancer Collaborative Group (IGCCCG) good-risk germ cell tumors (GCT) have cure rates exceeding 90% when appropriate chemotherapy is administered and post-chemotherapy surgery is utilized.1 Given the excellent clinical outcomes in this group of patients, research efforts have focused on minimizing treatment-related toxicity while maintaining or improving upon historic cure rates. 

For patients with good-risk disease, multiple strategies for lessening adverse effects have been investigated. In a prospective phase III multicenter trial, three cycles of bleomycin (B), etoposide (E), and cisplatin (P) resulted in similar efficacy but less toxicity than four cycles of BEP, demonstrating that the fourth cycle of BEP could be safely eliminated and establishing three cycles of BEP as a standard of care for good-risk disease.2

Investigators sought to further improve chemotherapy tolerability by substituting carboplatin (C), which is more myelosuppressive but less emetogenic, ototoxic, neurotoxic, and nephrotoxic than cisplatin, into multidrug regimens. Two prospective studies revealed inferior outcomes for patients when carboplatin was substituted for cisplatin: (1) 76% disease-free survival with four-cycle EC compared with 87% with four-cycle EP, and (2) 91% failure-free survival with four-cycle BEP compared with 77% with four-cycle BEC.3,4 These trials definitively demonstrated that cisplatin could not be replaced by carboplatin in patients with good-risk nonseminomatous GCT.

Investigating the Elimination of Bleomycin

Significant attention has also been devoted to eliminating bleomycin from chemotherapy regimens for good-risk disease because of bleomycin-specific adverse effects, which can include rare but occasionally fatal pulmonary toxicity as well as Raynaud phenomenon. Investigators at Memorial Sloan Kettering Cancer Center (MSKCC) reported similar efficacy with an improved toxicity profile among patients with good-risk disease treated with four cycles of EP compared with those treated with four cycles of vinblastine, actinomycin D, bleomycin, and cisplatin.5 The same group reported excellent long-term results with four cycles of EP in patients with good-risk disease, including 94% relapse-free and 96% overall survival rates at a median follow-up of 7.7 years.1 Notably, 62 of 204 (30%) patients with nonseminoma histology had teratoma (25%) or viable GCT (5%) at post-chemotherapy surgery, emphasizing the role of complete surgical resection of residual disease.

Three Cycles of BEP Versus Four Cycles of EP

One randomized trial, conducted by the French Federation of Cancer Centers’ Genito-Urinary Group from 1993 to 1999, has directly compared three cycles of BEP with four cycles of EP in patients with good-risk metastatic GCT.6 The trial enrolled 270 patients, and 257 were evaluable for efficacy. Approximately one-half of the patients in each arm were counseled to have all residual disease greater than 1 cm resected after chemotherapy, and patients with viable GCT elements received additional chemotherapy. Favorable response rate comprised the primary endpoint and included any of the following: complete clinical response (serologic and radiographic complete response), pathologic complete response (normalization of tumor markers and fibrosis, necrosis, or teratoma on resected specimens), surgical complete response (complete resection of viable GCT components and normal tumor markers), and partial response (complete serologic response with residual subcentimeter radiographic masses).

The primary endpoint was achieved in 94.7% of patients receiving three cycles of BEP and 96.8% of patients receiving four cycles of EP. Event-free survival was 91% in the three-cycle BEP arm and 86% in the four-cycle EP arm for the intention-to-treat analysis, and the 4-year overall survival in the good-risk intention-to-treat population was 97% and 93% for three-cycle BEP and four-cycle EP, respectively. There were no statistically significant differences in the various survival analyses. It should be noted that the cumulative total doses of etoposide and cisplatin were lower than planned, probably due to the protocol’s dose-modification guidelines. In addition, only 40% of patients with residual masses above 1 cm had surgery.  The impact of these deviations from standard GCT management on the trial’s outcome is unknown.

From a toxicity perspective, neutropenia (but not neutropenic fever) was significantly more common in the four-cycle EP arm (90% vs. 72%; p = 0.0002), whereas grade 1 neurotoxicity (16% vs. 5%; p = 0.006) and dermatologic toxicity (29% vs. 8%; p < 0.0001), including Raynaud phenomenon, were more common in patients receiving three-cycle BEP. No differences in pulmonary toxicity were noted on a patient-level analysis (increased toxicity when comparing by cycles), although only clinical assessment was employed as opposed to pulmonary function tests.

The study authors ultimately recommended that three-cycle BEP should be the treatment of choice for good-risk nonseminoma based on the nonsignificant differences in clinical outcomes. However, this trial has been criticized because of conclusions based on secondary endpoints, testing of multiple endpoints, post-hoc analyses, suboptimal trial design, and imbalance in treatment intensity and treatment delays between the two arms favoring BEP.1 Because of the inconclusive nature of the available evidence supporting one regimen over the other, the National Comprehensive Cancer Network Guidelines recommend either three cycles of BEP or four cycles of EP for patients with good-risk GCT.

The Impact of Treatment Regimen on Post-Treatment Histology

Most of the literature comparing chemotherapy regimens for patients with good-risk nonseminoma have included only a limited analysis of the chemotherapy regimen’s impact on post-chemotherapy histology of resected residual masses. Such investigation is important, as it may help clarify discrepancies in rates of response and relapse reported in randomized trials.6

Two large surgical series from Indiana University and MSKCC provide provocative data regarding the impact of treatment regimen on post-treatment histology.7,8 The group from Indiana, where three-cycle BEP is favored, reported a retrospective review of 226 patients who received either three-cycle BEP (179 patients) or four-cycle EP (47 patients) and then subsequently received post-chemotherapy retroperitoneal lymph node dissection (PC-RPLND) from 1985 to 2011. Differences in the rate of viable GCT elements in PC-RPLND specimens were compared, and propensity matching was performed to identify risk factors for viable GCT elements on pathologic examination. Of note, the size of residual masses after chemotherapy and rate of teratoma in PC-RPLND specimens were not reported.

The two cohorts had several prominent differences. In the group receiving four-cycle EP, only one patient (2%) received chemotherapy at Indiana University, whereas 48 patients (27%) who received three-cycle BEP were treated at Indiana. This disparity introduces a referral bias as a potential significant confounding factor. For example, suboptimal doses or delays in administration of cisplatin and etoposide at outside institutions (which have been demonstrated to lead to inferior outcomes in randomized trials) may have affected results. Another important difference between the EP and BEP cohorts was the proportion of patients with seminoma in their primary tumor histology (15% vs. 4%, respectively; p = 0.012). Following completion of chemotherapy, patients with seminoma are typically managed with observation unless active disease is suspected based on PET scan findings or enlarging retroperitoneal lymph nodes. The higher proportion of seminoma in the four-cycle EP arm biased toward having viable GCT in the PC-RPLND specimen. Overall, seven of 14 patients with primary seminoma had seminoma in PC-RPLND specimen (five of seven in the EP arm, and two of seven in the BEP group). The overall incidence of viable seminoma in the EP group was 11% compared with 1% in the BEP arm.1 Further, the proportion of patients with pure yolk sac tumor or yolk sac tumor components was not reported, which may have implications with respect to chemotherapy response and presence of teratoma in the PC-RPLND specimen.9 Information regarding the size of residual nodes would also have been beneficial.

Ultimately 29 (12.8%) patients in the entire cohort had active cancer on PC-RPLND, including 8% in the BEP group and 32% in the EP cohort (p < 0.01). This rate of viable tumor in the four-cycle EP group was significantly higher than the 5% rate in patients who received four-cycle EP for nonseminoma at MSKCC and underwent PC-RPLND.1 The rate of teratoma in the two groups was also not reported. The odds ratios—unadjusted (OR 5.5, 95% CI [1.3, 20.5]), multivariable (OR 5.3, 95% CI [1.3, 22.6]), and propensity score adjusted (OR 3.8, 95% CI [1.3, 11.2])—were in favor of three-cycle BEP (p < 0.05 for all). The authors suggested that in the absence of conclusive data comparing four-cycle EP with three-cycle BEP, tumor treatment response is an appropriate surrogate endpoint and favors three-cycle BEP.

The MSKCC experience, where four-cycle EP was favored, included 579 patients with good-risk nonseminomatous GCT who were treated with four-cycle EP (505 patients), three-cycle BEP (46 patients), or four-cycle BEP (28 patients).8 Most patients who received bleomycin-containing regimens were treated at outside facilities, including all 28 who received four-cycle BEP. For unexplained reasons, more patients in the BEP arm had a teratomatous component in their orchiectomy specimens (50% vs. 41%) and larger preoperative lymph node size (35% with < 2 cm lymph node and 27% with > 5 cm lymph node in the BEP group vs. 70% with < 2 cm lymph node and 8% with > 5 cm lymph node in the EP group). Fibrosis was more common in the EP group (60% vs. 39%; p = 0.001), whereas pure teratoma (53% vs. 32%; p = 0.001) and the presence of any teratoma (57% vs. 34%; p < 0.001) were more prevalent in the BEP arms, respectively. Importantly, there were no significant differences in the rates of viable GCT at PC-RPLND (BEP 5%, EP 6%).

On multivariable analysis, receipt of BEP was associated with increased risk of teratoma on PC-RPLND histology (OR 2.0, 95% CI [1, 4]; p = 0.04) and also when patients receiving four-cycle BEP were excluded (OR 3.7, 95% CI [1.5, 8.9]; p = 0.004). The authors also identified bleomycin-containing regimens to be associated with higher rates of teratoma in patients with residual masses smaller than 1.1 cm whether three or four cycles of chemotherapy were administered. The 5- and 10-year relapse-free survival estimates were nearly identical between patients receiving EP and BEP (95%). The authors concluded that histologic findings at PC-RPLND should be considered when making treatment decisions. As with the Indiana series, referral bias is a potential confounding factor in interpreting the MSKCC study. 

The amalgamation of literature considering optimal treatment of patients with good-risk nonseminoma emphasizes the following critical concepts: (1) chemotherapy should be administered in an effective manner by an oncologist with expertise in this area, and (2) the surgical expertise in performing PC-RPLND is critical to outcomes in this group of patients. Further, longer-term follow-up data are needed to elucidate the differential impact of an additional cycle of etoposide and cisplatin (in patients receiving four-cycle EP) or potentially synergistic adverse consequences of bleomycin and cisplatin (in patients receiving three-cycle BEP) on cardiovascular toxicities and secondary malignancies in GCT survivors. Additional follow-up will also reveal if there is any clinical significance of occult, untreated teratoma in patients who did not receive PC-RPLND. In the absence of an appropriately powered randomized clinical trial to answer these questions, it is mandatory that patients with GCT are treated with the utmost medical and surgical expertise to maximize oncologic outcomes while minimizing treatment morbidity.

About the Authors: Dr. Bagrodia is an assistant professor with the Department of Urology at the University of Texas Southwestern Medical Center. Dr. Feldman is an assistant attending within the Department of Medicine at Memorial Sloan Kettering Cancer Center and Weill Cornell Medical College.