Identification of serum proteomic signatures for predicting clinical outcomes of immunotherapy in advanced hepatocellular carcinoma

초록

Background and aims: Although combination immunotherapies, such as atezolizumab/bevacizumab and durvalumab/tremelimumab, have revolutionized the first-line treatment of advanced hepatocellular carcinoma (HCC), the lack of validated predictive biomarkers remains a clinical hurdle. This study aimed to identify circulating serum protein signatures that predict therapeutic response to atezolizumab/bevacizumab. Method: Pre-treatment serum samples were obtained from 44 patients with advanced HCC across four tertiary centers. Patients were stratified into responders (CR/PR/SD, n = 32) and non-responders (PD, n = 12). We employed mass spectrometry-based proteomics to characterize the differential protein expression landscape between the two groups. Results: Proteomic analysis revealed distinct molecular signatures associated with treatment resistance. Non-responders exhibited significantly elevated levels of pro-inflammatory markers, specifically C-reactive protein (CRP) and S100-A9. Conversely, several proteins involved in metabolic homeostasis and transport—including TF, TTR, APOC3, CLU, RBP4, ITIH2, IGFALS, and SERPINC1—were downregulated in non-responders. Notably, network analysis identified APOC3 and CLU as highly interconnected central hubs, suggesting their pivotal roles in the systemic modulation of treatment response. Conclusion: Circulating serum proteomic signatures, characterized by a balance between inflammatory markers and metabolic hub proteins, effectively predict the efficacy of atezolizumab/bevacizumab in advanced HCC. These findings provide a robust framework for non-invasive treatment stratification and personalized therapeutic strategies.