A novel prognostic model based on MVIRGs identifies ANGPT2 as a key target driving the malignant progression of HCC

BACKGROUND: Hepatocellular carcinoma (HCC) is a highly lethal malignancy with high recurrence and limited therapies. Microvascular invasion (MVI) is a critical prognostic factor, but its molecular mechanisms and precise predictive tools are lacking. This study aimed to identify key MVI-associated genes (MVIRGs), develop a prognostic model, and elucidate the role of ANGPT2 in HCC progression and therapeutic responses. METHODS: Machine learning on multi-cohort data (TCGA, ICGC) identified MVIRGs and constructed a prognostic model. ANGPT2, the core oncogenic MVIRG, was validated clinically (IHC) and functionally (in vitro and in vivo). We analyzed its association with the tumor immune microenvironment (TIME), immune checkpoint inhibitors (ICIs), and drug sensitivity. RESULTS: A powerful six-MVIRG prognostic model was developed, with ANGPT2 identified as the sole oncogenic gene. Clinically, ANGPT2 was significantly overexpressed, correlating strongly with aggressive features like MVI, vascular invasion, and advanced stages, serving as a robust poor prognostic biomarker for OS and PFS. Functionally, ANGPT2 knockdown significantly inhibited HCC cell growth and migration in vitro, and inhibited tumor growth in vivo. Bioinformatics revealed ANGPT2 fosters an immunosuppressive TIME, upregulating immune checkpoints (e.g., PD-L1) to mediate resistance to ICIs. High ANGPT2 expression showed a differential drug response: resistance to targeted therapies (e.g., EGFR-TKIs) but paradoxically increased sensitivity to chemotherapies (e.g., docetaxel, paclitaxel). CONCLUSION: This study constructed a powerful MVIRG model and identified ANGPT2 as a core oncogenic molecule. ANGPT2 drives HCC progression, mediates immune resistance, and differentially regulates drug sensitivity, making it a key biomarker for prognosis and treatment guidance.