Programmed cell death-related genes define distinct molecular subtypes and risk profiles in hepatocellular carcinoma.

Hepatocellular carcinoma (HCC) is a biologically and clinically heterogeneous malignancy, whose initiation and progression are increasingly recognized to be driven by the aberrant regulation of programmed cell death (PCD) pathways. To elucidate this association, we systematically integrated gene signatures from 21 distinct PCD types to characterize their expression patterns in HCC and construct a prognostic model for survival and therapeutic response prediction. Based on the TCGA-LIHC, GSE14520, and GSE116174 datasets, 85 candidate genes were identified through differential expression analysis and random survival forest algorithms. A 10-gene PCD-based risk score model was developed using machine learning including key genes such as KIF20A (associated with ferroptosis) and SLC2A1 (associated with anoikis), which demonstrated robust prognostic performance across three independent cohorts by stratifying patients into high- and low-risk groups. The risk score significantly correlated with immune infiltration, immune evasion potential, and predicted sensitivity to multiple anticancer agents. Consensus clustering based on model gene expression revealed two molecular subtypes with distinct survival outcomes and immune characteristics. A nomogram integrating the risk score exhibited favorable calibration and clinical applicability. Collectively, these findings propose a novel PCD-based molecular framework for prognosis assessment and personalized therapy in HCC.