Surrogacy of time to prgression for overall survival in advanced hepatocellular carcinoma treated with systemic therapy: A systematic review and meta-analysis of randomized controlled trials.

Abstract

403 Background: Time to progression (TTP) is widely used as the endpoint in early-phase trials for advanced hepatocellular carcinoma (HCC). However, the relevance of using TTP as a surrogate marker for overall survival (OS) in pivotal trials remains uncertain. Methods: The PubMed database and ASCO meeting library were searched for reports of randomized controlled trials that investigated patients with advanced HCC, included data for both OS and TTP, and were launched between 2009 and 2016. Correlation between hazard ratios (HRs) for TTP and OS was determined using weighted linear regression. Correlations between median OS and TTP, and between median OS and post-progression survival (PPS), defined as the period obtained by subtracting median TTP from median OS, were also evaluated. Results: The database search yielded 24 trials with 50 arms. Overall, TTP HR correlated with OS HR (R = 0.73); however, the coefficient in the regression equation was 0.48. When trials were stratified by treatment line, TTP HR was more strongly correlated with OS HR in second-line (R = 0.91) than in first-line (R = 0.77) settings. Correlation between median OS and median TTP was weak (R = 0.50), whereas the correlation between median OS and median PPS was strong (R = 0.78). Conclusions: In advanced HCC, the OS HR can be predicted from the TTP HR, especially in second-line trials, which is useful when considering whether to proceed to a pivotal trial based on the results of early-phase trials. However, TTP may not be appropriate as a primary endpoint in a pivotal trial if the trial aims to evaluate the survival benefit of novel agents because improvement of TTP cannot fully reflect improvement of OS because of the impact of PPS on OS.