Radiomics and blood biomarkers for predicting efficacy of sintilimab plus lenvatinib in advanced hepatocellular carcinoma

Background Immune checkpoint inhibitor–based combination therapy has emerged as an important treatment option for advanced hepatocellular carcinoma (HCC), yet therapeutic response remains highly heterogeneous. Biomarkers that jointly reflect tumor-intrinsic heterogeneity and host immune status are needed to improve response stratification and better understand variability in immunotherapy outcomes. Methods We evaluated the predictive value of Computed Tomography (CT) -derived radiomic heterogeneity and systemic inflammatory biomarkers in patients with advanced HCC treated with sintilimab plus lenvatinib. A total of 62 patients were included and divided into a training cohort (n = 36) and an independent real-world validation cohort (n = 26). Radiomic features were extracted from multiphase contrast-enhanced CT and summarized as a radiomics score. Hematological indices reflecting systemic immune-inflammatory status were assessed in parallel. An integrated model combining imaging-derived heterogeneity and immune-inflammatory markers was constructed for response stratification. Results The radiomics score discriminated responders from non-responders with area under the curve (AUC) values of 0.840 in the training cohort and 0.759 in the validation cohort. Among systemic biomarkers, the systemic immune-inflammation index (SII) was independently associated with treatment response. The integrated model combining radiomic heterogeneity and SII demonstrated improved discriminatory performance (AUC 0.938 and 0.819 in training and validation cohorts, respectively). Stratification based on the combined model was associated with differences in event-free survival, supporting the biological relevance of multimodal immune–tumor characterization. Conclusions Integrating imaging-derived tumor heterogeneity with systemic immune-inflammatory status may provide a biologically informed and non-invasive strategy for exploratory response stratification in advanced HCC undergoing immune checkpoint inhibitor–based therapy. Larger multicenter studies incorporating prospective validation and immune profiling are warranted to confirm clinical applicability.