Clinical outcomes of intermediate risk metastatic germ cell tumors (IRGCT): Results from a single-institution series.

Abstract

380 Background: IRGCT comprises a consistent category of metastatic patients (pts), and information on the recommended management of these pts should be updated. Usually they enter clinical trials for poor prognosis GCT. We aimed to address the heterogeneity of this category and to identify clinical prognostic factors for sub-stratification of pts. Methods: Data on consecutive pts with IRGCT and who received treatment at Fondazione INT Milano in the time-frame 02/1980-03/2014 were collected. Cox regression analyses were done evaluating potential prognostic factors for overall survival (OS, primary endpoint) to first-line therapy. Each factor was evaluated in a multivariable model. An exploratory OS comparison between outlier groups was undertaken with Kaplan Meier curves and logrank test. Results: Data on 181 pts were collected. Median age was 27 yrs (IQR 22-32), 10 pts had a retroperitoneal (RP) primary, 6 had pure seminoma. 72 (39.8%) had lung metastases and 54 (32.3%) bulky (i.e. ≥10cm) RP lymph-nodes (LN). Pts received cisplatin, bleomycin and etoposide (PEB, n=156) or vinblastine (PVB, n=23), 2 other treatments. Median follow up was 173 months (IQR: 87-237). Globally, 5-y PFS and OS were 66.8% (95%CI: 60.1-74.2) and 83.3% (77.8-89.2). However, 5-y OS for pts with AFP 5,000-10,000 IU/ml (N=19) was 61.8% (95%CI: 43.0-88.7) while it was 89.1% (95%CI: 81.2-97.7) for nonseminomas with elevated LDH only (N=57) and similar for elevated HCG only (N=22); overall p<0.001. Multivariable analysis for OS is shown in the table (c-index= 0.63). Distribution of variables over time: bulky RP LN and elevated LDH were more frequent in earlier series (p=0.003 and 0.011). Conclusions: The prognostic heterogeneity of IRGCT category is a matter of fact and should be addressed by clinical trials. Pts with highly elevated AFP have an OS similar to poor prognostic category, while those categorized by elevated HCG or LDH only are close to good risk ones. [Table: see text]