Serum tumor marker (STM) decline rates during high-dose chemotherapy (HDCT) to predict outcome for germ cell tumor (GCT) patients (pts).

Abstract

4532 Background: STM decline rates predict outcome for GCT pts during 1st-line conventional-dose chemotherapy. We investigated the importance of STM decline rates during salvage HDCT. Methods: HDCT included an initial low dose (LD) portion (1-2 cycles of paclitaxel plus ifosfamide [TI] + G-CSF) for stem cell collection followed by a high-dose (HD) portion (3 cycles of carboplatin + etoposide or TI + carboplatin). Eligible pts had elevated STM (AFP >15 and/or HCG >2.2) at the start of LD and completed ≥1 HD cycle. Slope and t1/2 were calculated for HCG and AFP for cycles 1-2 of LD and HD starting with the peak value during days (d) 1-6 of LD and HD to avoid interference from lysis. T1/2 for AFP ≤7d and HCG ≤3.5d were categorized as satisfactory (SAT); normalization within 7d was also considered SAT, regardless of t1/2. Pts with elevated AFP and HCG had to have adequate decline in both to be considered SAT. Progression-free (PFS) and overall survival (OS) were compared for SAT vs. unsatisfactory (UNSAT) pts using the log-rank test. Results: Of 117 pts (115 male, 105 nonseminoma), the most common primary sites were testis (81) and mediastinum (26). 93, 19, and 5 pts completed 3, 2, and 1 HD cycles, respectively. Overall, 19/117 had SAT decline during LD (19/77 for HCG; 7/57 for AFP). For LD, there was no significant difference in PFS or OS for SAT vs. UNSAT decline. For HD, 17/96 had SAT decline (11/59 for HCG; 9/47 for AFP). SAT overall decline (HCG and AFP) during HD predicted superior PFS (p=0.0025) and OS (p=0.0046) with 2-year (yr) PFS 0.88 (0.59, 0.97) vs. 0.37 (0.26, 0.47) and 3yr OS 0.82 (0.55, 0.94) vs. 0.37 (0.26, 0.48). SAT HCG decline alone also predicted superior PFS (p=0.0018) and OS (p=0.004) with 2yr PFS 0.91 (0.51, 0.99) vs. 0.32 (0.19, 0.46) and 3yr OS 0.91 (0.51, 0.99) vs. 0.35 (0.22, 0.50) whereas only a non-significant trend for improved PFS (p=0.14) and OS (p=0.16) was seen for AFP alone. Conclusions: STM decline rates during the HD portion of salvage HDCT predicts favorable PFS and OS, mostly driven by HCG. However, UNSAT STM decline does not exclude long-term PFS/OS. STM decline rates during the initial LD portion of HDCT programs do not predict outcome; thus, UNSAT LD decline should not prohibit subsequent HDCT.