Progression free survival (PFS) as a surrogate endpoint for overall survival (OS) in first-line advanced ovarian cancer (AOC) therapy: A Gynecologic Cancer InterGroup individual (GCIG) patient-level (IPD) analysis of multiple randomized trials.

5576 Background: OS is considered the gold standard endpoint for controlled clinical trials but it requires extended follow-up (median OS > 40 months for first line therapy) and large sample sizes. Our objective was to evaluate whether PFS based on CA125 measurements confirmed by radiological exam or on combined GCIG criteria is a surrogate endpoint for OS in AOC. Methods: Using the meta-analytic approach on trials published after 2000, correlations between PFS and OS at the individual level, and between treatment effects on PFS and on OS at the trial level, were estimated using Kendall’ tau and copula bivariate (R2Copula) models respectively. Criteria for PFS surrogacy required R2Copula ≥ 0.80. Results: We analyzed IPD from 8,973 patients enrolled in 15 randomized first line trials of standard (n = 7), intensification (n = 5) and maintenance (n = 3) chemotherapies or targeted treatments. No heterogeneity in the treatment effects across trials was detected. High correlations were found at the individual level (tau = 0.69) but low correlation at the trial level (R2Copula = 0.03). Sub-group analyses led to similar results. Conclusions: This large IPD meta-analysis did not establish PFS as a surrogate endpoint for OS in first line treatment of AOC, but the analysis was limited by the narrow range of treatment effects observed and post study treatment. Additional trials are being collected that will enable us to refine the sub-groups analyses. Table of results Endpoint / Trial type Trials N (pts) taua R2Copulaa Overall 15 (8973) 0.69 0.03 CA125 confirmed by radiological exam 9 (5126) 0.67 0.16 GCIG criteria 4 (2548) 0.7 0.14 Carbo-Tax as control 9 (5807) 0.72 0.05 Standard or intensification 12 (7703) 0.71 0.10 Maintenance 3 (1270) 0.64 0.14 atau and R2 values range from 0 (no association) to 1 (perfect correlation).