Background & aims: Immune checkpoint inhibitors (ICIs) can induce long-term survival and even cancer cure in several cancers. Mixture cure models (MCMs) estimate the fraction of long-term survivors and cured patients but have not been applied in HCC.
Methods: We identified phase 3 randomized trials of first-line ICI in advanced HCC with mature follow-up (≥30 months) and analyzed a cohort of HCC patients treated with atezolizumab-bevacizumab. After reconstructing Kaplan-Meier curves, MCMs estimated long-term survivors' fraction (overall survival, OS) and cure fractions (progression-free survival, PFS).
Results: In HIMALAYA (median follow-up of 60 months), long-term survival was 12.8% (95%CI:8.5-18.6%) with durvalumab-tremelimumab versus 5.2% (95%CI:2.6-10.2%) with sorafenib; cure fraction was not assessable.In CheckMate9DW (median follow-up: 35.2 month), long-term survival was 8.7% (95%CI:0.2-81.3%) versus 4.1% (95%CI:0.1-66.1%) and cure fractions were 17.8% (95%CI:12.0-25.8%) versus 3.5% (95%CI:0.7-16.7%) for nivolumab-ipilimumab and sorafenib/lenvatinib respectively with long-term OS estimates remaining exploratory due to limited late numbers at risk.In RATIONALE-301, long-term survival was 25.2% (95%CI:19.2-32.2) with tislelizumab versus 15.4% (95%CI: 10.0-22.8) with sorafenib. IMbrave150 trial was not analyzed due to insufficient follow-up (15.6 months). Among a clinical cohort of 1581 patients treated by atezolizumab-bevacizumab (median follow-up: 34.7 months), long-term survival was 12.3% (95%CI:9.3-16.1%) and cure fraction 7.9% (95%CI:6.3-9.8%). In 1187 patients meeting IMbrave150 criteria, long-term survival reached 15.4% (95%CI:10.6-18.5%) and cure fraction 9.1% (95%CI:7.3-11.4%).ALBI score, and hepatitis C predicted long-term survival and albumin and hepatitis C predicted cure.
Conclusion: Across trials and real-world data, ICIs combinations achieve long-term survival in 10-15% of advanced HCC patients, with cure fractions of 7-9%.