Meta-analysis of surrogate endpoints for survival in patients with unresectable hepatocellular carcinoma treated with immune checkpoint inhibitor-based regimens.

Abstract

483 Background: Overall survival (OS) remains the gold standard for demonstrating clinical benefit in advanced hepatocellular carcinoma (HCC). While surrogate radiology-based endpoints for OS, such as progression-free survival (PFS) and objective response rate (ORR) are commonly used in the field of oncology, they remain controversial in the setting of HCC where antiangiogenic agents were the mainstay of treatment until recently. Immune checkpoint inhibitors (CPIs) as monotherapy or in combination have shifted the treatment paradigm in HCC. We investigated the surrogacy of radiology-based endpoints from CPI clinical trials in advanced HCC. Methods: Study level data were collected for disclosures published in EMBASE, PubMed, Clinicaltrials.gov and conference abstracts. Keywords used included (abbreviated here): immune CPIs, advanced or metastatic HCC and ORR, PFS and OS. Key inclusion criteria included: Phase I, II and III trial data, and patients with advanced or metastatic HCC who received CPI monotherapy or in a combination regimen. The endpoints investigated were median OS (mOS), median PFS (mPFS) and ORR (RECIST v1.1). A weighted Pearson correlation coefficient, with weights corresponding to the number of patients in the different treatment arms of the trials, was calculated between ORR/mOS, ORR/mPFS and mPFS/mOS in all arms, arms that included only patients with Child-Pugh A liver function, and arms in the first-line therapy setting. 95% confidence intervals (CIs) based on the empirical bootstrap re-sampling method are also reported. Results: Eighteen clinical trials (8 randomized trials and 10 multi-cohort trials) published between 2017 and 2021, including 3,246 patients across 32 treatment arms, were selected for inclusion in this meta-analysis. Ten trials included CPI monotherapy arms, 5 trials included CPI + TKI arms, 4 trials included CPI + anti-angiogenic monoclonal antibody arms, and 2 trials included CPI + CPI arms. There was 1 CPI + chemotherapy arm and 1 CPI + CPI + TKI arm in separate trials. Weighted Pearson correlation coefficients and corresponding 95% CIs are presented in the table. Conclusions: Our analysis suggests ORR/mOS, ORR/mPFS, and mPFS/mOS are positively correlated, with ORR/mPFS more strongly correlated compared to the other pairs, in patients with advanced HCC who were treated with CPI-based regimens. [Table: see text]