Hepatocellular Carcinoma Progression and Drug Resistance: A Multi-Omics and Machine Learning Approach to Identify Novel Therapeutic Targets and Biomarkers

• • A 12-gene signature (including TPX2, AURKA, and BIRC5) was identified using LASSO and random forest, achieving a C-index of 0.82 (95% CI: 0.78-0.86) for 5-year survival prediction, outperforming traditional staging (C-index 0.71). • • Knockdown of TPX2 in HCC cell lines (HepG2 and Huh7) reduced proliferation by 45% and migration by 60% (p < 0.01), validating its oncogenic role. • • TPX2 overexpression increased sorafenib IC50 from 2.5 μM to 8.2 μM (3.3-fold), indicating a direct role in drug resistance; combination therapy targeting TPX2 restored sensitivity by 70%. • • The nomogram integrating the gene signature and clinical factors (tumor stage, AFP level) achieved a C-index of 0.82, significantly improving risk stratification compared to AFP alone (C-index 0.65).