Surrogate endpoints for overall survival in combined chemotherapy and radiotherapy trials in nasopharyngeal carcinoma: Meta-analysis of randomised controlled trials.

Abstract

We used a literature-based meta-analysis to assess whether failure-free survival (FFS) or progression-free survival (PFS) could be reliable surrogate endpoints for overall survival (OS) in trials of combined chemotherapy and radiotherapy for nasopharyngeal carcinoma (NPC). We identified randomised trials that evaluated combined chemoradiotherapy strategies, and reported FFS or PFS and OS in NPC. We analysed the treatment effects on FFS or PFS, and OS. We used the coefficient of determination (R(2)), and the surrogate threshold effect (STE) to assess the trial-level correlation. Twenty-one trials (5212 patients), with sixteen treatment-control comparisons for FFS, and nine for PFS, were analysed. FFS was strongly correlated with OS (R(2)=0.88, STE=0.84), as was PFS (R(2)=0.90, STE=0.88). Moreover, FFS and PFS at 3 years were still strongly correlated with 5-year OS (R(2)=0.80, STE=0.83; R(2)=0.85, STE=0.84). Both FFS and PFS could be valid surrogate endpoints for OS in trials of combined chemotherapy and radiotherapy for NPC; PFS may be a more acceptable surrogate endpoint compared with FFS.