Identification of genes related to macrophage polarization and mitochondrial dysfunction in hepatocellular carcinoma and prognostic modeling via LASSO-Cox regression

AIM: Hepatocellular carcinoma (HCC) is the most common primary liver cancer in adults, with increasing incidence. It is helpful to establish a prognostic risk prediction model related to macrophage polarization and mitochondrial dysfunction associated with HCC. METHODS: Cox and Least Absolute Shrinkage and Selection Operator (LASSO) regression, Gene Set Variation Analysis (GSVA) and Gene Set Enrichment Analysis (GSEA), protein-protein interaction (PPI) network. RESULTS: A total of 7 model genes (EZH2, G6PD, HMGA2, MAGEB2, PYCR1, SLC7A11, and SPP1) were identified. And the LASSO-Cox model showed preliminary prognostic predictive performance in the TCGA-LIHC cohort (0.9 > AUC > 0.7), with decision curve analysis (DCA) showing clinical potential, particularly in the third year. GSEA revealed that genes associated with Liver hepatocellular carcinoma (LIHC) exhibited enrichment in functions and pathways, notably including FCGR3A Mediated IL10 Synthesis, etc. GSVA indicated multiple pathways were significant in both Low-Risk and High-Risk groups, such as the biocatamcm pathway (p < 0.05). The PPI Network shows connections among G6PD, SLC7A11, HMGA2, and EZH2, with GeneMANIA predicting their interactions with similar function genes. CONCLUSION: Our study screened candidate prognostic genes associated with macrophage polarization and mitochondrial dysfunction based on bioinformatics analysis and established a prognostic model.