Radiological progression-free survival as a surrogate for overall survival in patients with metastatic hormone-sensitive prostate cancer: A bivariate meta-analysis.
Background: Overall survival (OS) is the standard efficacy endpoint in various solid tumor trials; however, it requires longer follow-up time for assessment than potential intermediate endpoints. This study evaluated radiological progression-free survival (rPFS) as a surrogate for OS in metastatic hormone-sensitive prostate cancer (mHSPC) using aggregate-level data from randomized controlled trials (RCTs).
Methods: A systematic literature review identified mHSPC RCTs published through December 2023, reporting hazard ratios for rPFS (HRrPFS) and OS (HROS). Correlation between HRrPFS and HROS was assessed using bivariate random-effects meta-analysis (BRMA). Predictive validity was assessed with leave-one-out cross-validation (LOOCV). The surrogate threshold effect (STE), or minimum rPFS benefit predicting an OS benefit, was estimated using recent mHSPC trial sample sizes. Sensitivity analyses (1) omitted trials that had only one of the endpoints reported, (2) omitted HRs that violated proportional hazards assumptions, (3) omitted trials that allowed cross-over and (4) investigated different assumed values of the within-study correlation.
Results: The primary analysis included 35 treatment comparisons from 31 trials. The estimated rPFS-OS correlation was 0.95 (95 % CrI: 0.75, 1.00). LOOCV confirmed HROS were within 95 % prediction intervals. The estimated STE ranged from 0.55 to 0.71 depending on the trial size being predicted. Sensitivity analyses produced strong but slightly lower correlations (0.87, 0.89, 0.91) than the primary analysis, with full coverage of the reported HROS in cross validation. Increasing within-study correlation slightly reduced between-study correlation.
Conclusions: The derived surrogacy equation enables OS estimation based on reported rPFS benefits in mHSPC, meeting NICE's 95 % surrogate validity threshold. These findings support rPFS as a reliable surrogate for OS, facilitating prediction of OS benefits in future mHSPC trials.