Abstract
BACKGROUND: Pyruvate metabolism presents a novel, therapeutically targetable metabolic vulnerability in hepatocellular carcinoma (HCC). In this study, we sought to identify HCC molecular subtypes and develop prognostic signatures based on pyruvate metabolism-related genes (PMRGs) to inform personalized therapeutic approaches. METHODS: Transcriptional profiles and clinical data of HCC patients were obtained from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) datasets. Consensus clustering was employed for molecular classification, while a least absolute shrinkage and selection operator (LASSO) Cox regression model was constructed for risk score calculation. The relationship between the risk score and HCC prognosis, immune landscape, gene expression, and drug sensitivity was analyzed. RESULTS: Twenty PMRGs were identified as significantly associated with HCC prognosis. Consensus clustering of these genes revealed two distinct molecular subtypes that stratified patients into groups with favorable and unfavorable outcomes. A novel six-gene signature, comprising ACACA, ACAT1, CYP1, DLAT, LDHA, and ME1, was developed for HCC prognostication. The receiver operating characteristic (ROC) curve demonstrated robust survival prediction in all cohorts, allowing the stratification of patients into high- and low-risk groups with markedly different overall survival (OS). The signature-derived nomogram displayed appreciable clinical net benefit. Enrichment analysis revealed activation of PMRGs and enrichment of diverse metabolic processes and signaling pathways in the high-risk group. Moreover, the prognostic signature showed significant correlations with immune landscapes and therapeutic responses, enabling prediction of immunotherapy responsiveness. CONCLUSIONS: Collectively, a unique PMRG-based signature effectively predicts prognosis in HCC patients and provides valuable insights into chemotherapy and immunotherapy strategies for these individuals.