Objective To construct a prognostic prediction model for hepatocellular carcinoma (HCC) based on immune and metabolism related genes, analyze the prognostic immune response of HCC patients, and screen potential drugs for HCC treatment through drug sensitivity analysis.

Methods HCC expression profiling and clinical data were obtained from The Cancer Genome Atlas (TCGA) database, and a list of immune-related genes was obtained from the Immport database; the Perl language was used to extract metabolism-related pathway gene sets from the Molecular Signatures Database(MSig DB), and co-expression related genes were found through differential analysis and co-expression analysis; the univariate Cox regression analysis, the least absolute shrinkage and selection operator (LASSO) regression analysis, and multivariate Cox regression analysis were used to screen prognosis-related genes and construct a risk prognosis model for HCC, and risk scores for all HCC samples were calculated. Using the median risk score as the critical value, the reliability of the prognostic model was evaluated through risk curves, Kaplan-Meier survival analysis, receiver operating characteristic (ROC) curves, independent prognostic analysis, and Nomograms. The correlations between risk scores and pathway enrichment analysis as well as immune cell infiltration were analyzed. Drug sensitivity analysis was used to identify potential therapeutic drugs for HCC.

Results Five immune and metabolic genes with independent prognostic value were obtained, and a prognostic model based on immune and metabolic genes was constructed. Survival analysis showed that in the total dataset, training group and validation group, the survival rate of the low-risk group was significantly higher than that ofthe high-risk group (P < 0.05). The areas under the ROC curves of the prognostic model for the training group at 1, 3 and 5 years were 0.780, 0.699 and 0.706 respectively. Cox regression analysis showed that grading and risk score could be used as independent prognostic factors for HCC (P < 0.05), with a concordance index of 0.734 (95%CI, 0.669 to 0.798), indicating good model performance. Immune cell infiltration results showed significant differences in resting NK cells, monocytes, M0 macrophages, and M1 macrophages between the high-risk and low-risk groups (P < 0.05). Drug sensitivity analysis screened 12 drugs that may have potential therapeutic effects in HCC patients (P < 0.01).

Conclusion The prognostic model of HCC based on five immune and metabolic genes has good predictive performance, and can be used as a new indicator for prognosis evaluation; the 12 drugs screened out have potential efficacy for HCC.