4510 Background: First-line high-dose chemotherapy (HDCT) for poor-prognosis GCT PTS demonstrated improved relapse-free and overall survival compared to historical controls receiving conventional-dose regimens (JCO 1996;14:2546). Retrospective studies also showed correlations between long-term outcome and initial declines of alpha-fetoprotein (AFP) and/or human chorionic gonadotropin (HCG). This trial sought to determine: 1) if early intervention with HDCT resulted in an outcome superior to standard-dose chemotherapy and 2) whether AFP and HCG declines during cycles 1 & 2 correlated with long-term treatment outcome.
METHODS
PTS with untreated intermediate- or poor-risk GCT by International criteria (IGCCCG) were randomized to either 2 cycles of standard BEP (bleomycin, etoposide & cisplatin) followed by 2 cycles of HDCT (cyclophosphamide, etoposide, carboplatin) plus stem-cell rescue (BEP + HDCT) or to 4 cycles of BEP. The primary endpoint was the percent of PTS in complete response at 1 year (CR-1 yr). Based on an historical CR-1 yr of 45% for BEP, targeted accrual was 109 PTS/arm to detect a 20% improvement in PTS receiving BEP + HDCT with an alpha of 5% and 80% power. Randomization used random permuted blocks; strata were risk status (poor/intermediate) and treatment center. An independent DSMB performed one interim analysis of CR-1yr and survival in May 2000.
RESULTS
219 PTS were randomized; 108 to BEP + HDCT and 111 to BEP alone. Final analysis demonstrated a CR-1yr of 52% for BEP + HDCT and 48% for BEP alone (P = .53 via actuarial methods). Slow marker decline PTS (AFP and/or HCG) during cycles 1 & 2 of BEP had a shorter progression-free and overall survival compared to satisfactory decline PTS (P ≤ .02) Among 70 PTS with unsatisfactory marker decline in cycles 1 & 2, the CR-1 yr was 61% for PTS receiving HDCT for cycles 3 & 4 versus 31% for those receiving 2 more cycles of BEP (P = .008).
CONCLUSIONS
The routine inclusion of HDCT for intermediate- and poor-risk GCT does not improve treatment outcome. Serum marker decline during the first 2 cycles of BEP chemotherapy provides a clinically useful estimate of outcome. Research support provided by the NIH. [Table: see text].