Clinical predictors of response to atezolizumab/bevacizumab in Child-Pugh B patients with hepatocellular carcinoma.
BACKGROUND & AIMS: We aimed to identify, among patients with advanced hepatocellular carcinoma (HCC) and Child-Pugh B cirrhosis, typically excluded from clinical trials, a subgroup that may benefit from first-line atezolizumab/bevacizumab (A/B). METHODS: We conducted a retrospective international multicenter study including patients with unresectable HCC treated with first-line A/B between 2020 and 2024 across 12 centers. A cohort of Child-Pugh B patients treated with sorafenib served as a control. Baseline clinical, biological, and tumor features were correlated with radiological response, progression-free survival (PFS), and overall survival (OS). RESULTS: Among 1,499 patients, 246 (16.4%) had Child-Pugh B cirrhosis. Within Child-Pugh B, 72% were B7, 21.5% B8, and 6.5% B9; 73% had albumin-bilirubin (ALBI) grade 2 and 27% grade 3. Median OS and PFS were significantly shorter in Child-Pugh B (8.1 and 5.2 months, respectively) vs. Child-Pugh A (16.8 and 8.6 months, respectively; both p <0.001). Two-year OS was 20% for Child-Pugh B vs. 38% for Child-Pugh A. Child-Pugh B patients treated with A/B had longer OS than those treated with sorafenib (p = 0.002). A score combining ALBI grade 1/2 and metastatic status identified prognostic subgroups (10.3 vs. 7.9 vs. 4.2 months; p <0.0001). Improvement to Child-Pugh A occurred in 31% and was associated with recent treatment of underlying liver disease. Radiological response (hazard ratio = 0.58, p = 0.021) and liver function improvement (hazard ratio = 0.59, p = 0.006) correlated with reduced mortality. CONCLUSIONS: Although Child-Pugh B patients have poorer survival, a subgroup, those with ALBI grade 1/2 and no extrahepatic metastasis, can derive meaningful benefit from A/B therapy. Improving underlying liver disease may contribute to better outcomes. IMPACT AND IMPLICATIONS: Child-Pugh B patients with advanced HCC are systematically underrepresented in clinical trials, creating a critical evidence gap for a population frequently encountered in real-world practice. This large multicenter study shows that a subset of these patients, those with ALBI grade 1/2 and without extrahepatic metastases, can have clinically significant benefit from first-line A/B, providing a practical prognostic tool to guide patient selection. Moreover, the association between treatment of the underlying liver disease and Child-Pugh class improvement suggests that optimizing hepatic function alongside systemic therapy may represent an actionable strategy to improve outcomes, warranting prospective validation.