To investigate the role of deubiquitination enzyme USP24 and its co-expression of tumor metabolic genes in clinical diagnosis and prognosis prediction of hepatocellular carcinoma.

The expression of USP24 in hepatocellular carcinoma was studied using the TCGA-LIHC cohort. Survival analysis was used to evaluate the role of USP24 in the prognosis of hepatocellular carcinoma. GSEA software was used to predict the possible tumor metabolic pathways regulated by USP24, and the major contributing genes of tumor metabolic pathways were screened by co-expression analysis. Differential expression analysis and univariate COX analysis were used to identify the metabolic prognostic genes associated with USP24, and the prognostic model was constructed using LASSO-COX algorithm. The diagnosis model of hepatocellular carcinoma was constructed by BP neural network, support vector machine, random forest and XGBoost algorithm. R4.2.2 software was used for statistical analysis, Wilcoxon test was used for comparison between the two groups, and Kaplan-Meier method was used for survival analysis. P<0.05 was considered statistically significant.

USP24 is significantly overexpressed in hepatocellular carcinoma, and the survival rate of patients with high expression of USP24 is lower. GSEA analysis showed that USP24 was associated with 11 tumor metabolic pathways. Coexpression analysis, differential expression analysis and univariate COX analysis selected 19 USP24-related metabolic prognostic genes, and 8 genes were selected by LASSO-COX algorithm to construct prognostic model. Survival analysis results suggested that the high-risk group had a worse prognosis than the low-risk group. The AUC of 1-year, 2-year and 3-year ROC curves of the training group were 0.793, 0.706 and 0.696, respectively, while the AUC of 1-year, 2-year and 3-year ROC curves of the verification group were 0.701, 0.684 and 0.728, respectively. The XGBoost model had better diagnostic ability, and the AUC of ROC curve in the verification group was 0.809.

USP24 is associated with the prognosis of hepatocellular carcinoma, which may regulate the progression of hepatocellular carcinoma through tumor cell metabolism. The diagnostic model and prognostic model constructed based on 8 tumor metabolic genes may be effective in diagnosing hepatocellular carcinoma and predicting the prognosis of patients, respectively.