Abstract
BACKGROUND: Hepatocellular carcinoma (HCC) is the fifth most common cancer. Since changes in liver metabolism contribute to liver disease development, it is necessary to build a metabolism-related prognostic model for HCC. METHODS: We constructed a metabolism-related-gene (MRG) signature comprising nine genes, which segregated HCC patients into high- and low-risk groups. RESULTS: The survival rate (overall survival: OS; relapse-free survival; and progression-free survival) of patients in the low-risk group of The Cancer Genome Atlas (TCGA) cohort was significantly higher than that of patients in the high-risk group. The OS prognostic signature was validated in the International Cancer Genome Consortium independent cohort. The corresponding receiver operating characteristic curves of the model indicated that the signature had good diagnostic efficiency, in terms of improving OS over 1, 3, and 5 years. Hierarchical analysis demonstrated that the MRG signature was significantly associated with better prognosis in male patients, patients aged ≤ 65 years, and patients carrying the wild-type TP53 or CTNNB1 genes. A nomogram was established, and good performance and clinical practicability were confirmed. Additionally, using the GSE109211 dataset from the Gene Expression Omnibus database, we were able to verify that the nine genes in this MRG signature had different responses to sorafenib, suggesting that some of these MRGs may act as therapeutic targets for HCC. CONCLUSIONS: We believe that these findings will add value in terms of the diagnosis, treatment, and prognosis of HCC.