Surrogate endpoints for overall survival in advanced hepatocellular carcinoma: A meta-analysis.

Abstract

563 Background: Recent advances in systemic therapies have resulted in improved overall survivals (OS) for patients with advanced hepatocellular carcinoma (aHCC). The most commonly utilized primary endpoint for phase III trials in aHCC is OS. A robust surrogate endpoint (SEP) can reduce time and financial costs of future trials and accelerate regulatory approvals of new therapies. We conducted a literature-based meta-analysis to evaluate SEPs for OS in aHCC. Methods: Randomized controlled trials evaluating systemic therapies in aHCC published 2007 - 2023 were identified through a systemic literature search of Cochrane and Medline databases as well as of abstracts presented at ASCO and ESMO meetings. Hazard ratios (HR) for OS and progression-free survival (PFS) and time to progression (TTP) were extracted. ΔORR was calculated as the difference in overall response rates (ORR) between control and experimental arms. Pearson correlation and mixed-effects meta-regression analyses were performed. Surrogate threshold effect (STE) was determined for each comparison when possible. A p < 0.05 was considered to be statistically significant. Results: 26 trials were identified in the first-line setting with 14,827 patients and 11 trials in the subsequent-line setting with 5,316 patients. In first-line trials, immunotherapy (IO) is associated with higher ORR than other agents (p=0.003). There are statistically significant correlations between HRs of PFS/TTP/ΔORR and OS (Table). The relationship between HR-PFS and HR-OS persists in trials with or without IO, but only in trials enrolling < 30% patients of non-viral aetiologies. In subsequent-line trials, there are statistically significant correlations between HRs of PFS/TTP and OS (Table). STE for HR PFS is 0.68 and 0.87 respectively for 1st and subsequent line trials. Conclusions: There are moderate to strong correlations between PFS and OS in aHCC in first-line and subsequent line trials. STE for HR-PFS of 0.68 in first-line and 0.87 in subsequent-line settings can guide sample size calculation in future clinical trials. [Table: see text]