Adjuvant donafenib plus sintilimab therapy for hepatocellular carcinoma at high risk of recurrence after resection: a retrospective cohort study

Background Hepatocellular carcinoma (HCC) carries a high recurrence risk after curative resection, yet effective adjuvant strategies remain lacking. Although donafenib plus anti-PD-1 monoclonal antibodies has shown efficacy in advanced HCC, the effectiveness in patients at high recurrence risk after curative resection remains unclear. Methods This retrospective cohort study included 143 patients with high-risk HCC after curative resection between June 2019 and April 2024. Patients received adjuvant donafenib plus sintilimab (n=76) or active surveillance (n=67). IPTW, stabilized IPTW (sIPTW), and propensity score matching (PSM) were applied to reduce baseline imbalance. The primary endpoint was recurrence-free survival (RFS), with overall survival (OS) and treatment-emergent adverse events (TEAEs) as secondary endpoints. Survival outcomes were assessed using Kaplan–Meier analysis and Cox regression models. Results After IPTW adjustment, adjuvant donafenib plus sintilimab was associated with improved RFS compared with surveillance (median RFS, 22.4 vs. 17.4 months; P = 0.036). The 12- and 24-month RFS rates were 78.2% and 44.4% versus 64.4% and 33.2%, respectively. Multivariable Cox analysis confirmed an independent association between adjuvant therapy and reduced recurrence risk (HR, 0.58; 95% CI, 0.35–0.96; P = 0.036), with consistent results in sIPTW and PSM analyses. OS showed a favorable but nonsignificant trend. Subgroup analyses demonstrated a consistent trend toward RFS benefit across most clinically relevant strata. Among treated patients, TEAEs occurred in 72.4%, mainly grade 1–2 events; grade ≥3 events occurred in 7.9%, with no treatment-related deaths. Conclusions Adjuvant donafenib plus sintilimab was associated with improved RFS in patients with high-risk HCC, with an acceptable safety profile, supporting its potential as a risk-adapted adjuvant strategy that warrants further validation in prospective studies.