Prediction of Hepatocellular Carcinoma Prognosis and Immunotherapy Response Using Mitochondrial Dysregulation Features

Hepatocellular carcinoma (HCC) is a major contributor to cancer-related deaths globally. Although there have been improvements in identifying treating the disease, patient outcomes are still unfavourable because of the significant variation in HCC. Mitochondrial-related genes (MRGs) are crucial in tumour metabolism, cell death and immune response, emerging as potential therapeutic targets. We analysed 2030 MRGs using TCGA, GEO and HCCDB18 databases. Differentially expressed genes were identified using edgeR and limma, and enrichment analysis was performed via the clusterProfiler package. A prognostic model was built using machine learning algorithms and evaluated using LOOCV. Immune infiltration was assessed with CIBERSORT, EPIC, MCPCounter and TIMER algorithms, and drug sensitivity was analysed using the CTRP and PRISM datasets. MRG expression levels are significantly associated with worse outcomes in HCC patients outperformed conventional clinical indicators in immune response revealed that individuals at high risk exhibited weaker immune responses, characterised by reduced immune scores, and elevated levels of CD8+ T cells and macrophages. Notably, high-risk patients also displayed heightened susceptibility to chemotherapy agents such as paclitaxel and irinotecan. Abnormal MRG expression serves as a significant biomarker for HCC prognosis. The developed model accurately predicts disease progression and can guide personalised treatment, especially for immune and chemotherapeutic therapies. Further validation with broader clinical samples is needed.