Supplementary Material for: Surrogate endpoints for overall survival in advanced hepatocelluar carcinoma in the era of immunotherapy: a trial level meta-analysis

Background: Systemic therapy containing immune checkpoint inhibitors (ICIs) has become the standard of care for patients with advanced hepatocellular carcinoma (HCC). A prior analysis from the pre -ICI era demonstrated a moderate correlation between progression-free survival (PFS) and overall survival (OS). We performed a literature-based meta-analysis to include randomized phase III trials (RCTs) of ICIs to evaluate surrogate endpoints for OS in advanced HCC.
Methods: RCTs evaluating systemic therapies in advanced HCC published/presented between 2007-2024 were identified through a systemic literature search. Hazard ratios (HRs) for OS and PFS were extracted. The change in the overall response rates (ΔORR) was calculated as the difference between experimental and control arms. Pearson correlation and mixed-effects meta-regression analyses were performed. Strength of correlation was determined using the criteria from the Institute for Quality and Efficiency in Health Care (IQWIG). Surrogate threshold effect (STE) was determined for each comparison when possible. Subgroup analysis was performed. A p < 0.05 was considered to be statistically significant.
Results: In total, 21 1st-line and 12 2nd-line RCTs were included. Of 1st-line RCTs, 48% evaluated ICIs either alone or in combination. There was a weak correlation between HR-PFS and HR-OS (n = 18, r = 0.64, 95% confidence interval (CI): 0.25 – 0.85, p = 0.004) and no correlation between ∆ORR and HR-OS (n=21, r = 0.42, 95% CI: : -0.01 – 0.72, p = 0.06). The STE for HR-PFS was 0.68. Subgroup analyses revealed a moderate correlation between HR-PFS and HR-OS in 1st-line RCTs enrolling fewer patients with non-viral etiology (n=9, r = 0.75, 95% CI: 0.17 – 0.94, p = 0.02). There was, however, a strong correlation between HR-PFS and HR-OS in 2nd-line RCTs (n=10, r = 0.90, 95% CI: 0.61 – 0.98, p < 0.001). The STE for HR-PFS in 2nd-line RCTs was 0.87.
Conclusion: The correlation between HR-PFS and HR-OS is weak in 1st-line RCTs in advanced HCC where OS remains the most appropriate endpoint. There is a strong association between HR-PFS and HR-OS for 2nd-line RCTs, suggesting that PFS as a suitable primary endpoint in that setting.