Objective  To construct a prognostic risk model for clinical treatment of hepatocellular carcinoma (HCC) based on public databases.

  Methods  The mRNA expression data and clinical information of HCC and adjacent normal tissues were downloaded from The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC). Differentially expressed genes (DEGs) related to overall survival (OS) were screened in the TCGA cohort, 2 or 3 mRNAs were selected to form a combination, and Cox risk proportional regression model was constructed for all combinations. The optimal gene combination was determined by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve, and external validation based on ICGC cohort was carried out. The patients were divided into high-risk group and low-risk group according to the median risk score of TCGA cohort, gene set enrichment analysis (GSEA) was performed, and the relative half-inhibitory concentrations (IC₅₀) of sorafenib, mitomycin, etoposide, adriamycin, paclitaxel and cisplatin in HCC patients were predicted by pRRophetic R software package.

  Results  For this prognostic risk model, the AUC of the ROC curve for predicting 1-, 3- and 5-year OS in the TCGA cohort were 0.786, 0.713 and 0.699, respectively, and the AUC of the ROC curve for predicting 1-, 3- and 4-year OS in the ICGC cohort were 0.719, 0.709 and 0.766, respectively. GSEA revealed that cell cycle related pathways were activated and bile acid metabolism was inhibited in the high-risk group. The IC₅₀ of sorafenib in the low-risk group was significantly lower than that in the high-risk group, while the IC₅₀ of cell cycle related chemotherapy drugs in the low-risk group was significantly higher than that in the high-risk group (P < 0.05).

  Conclusion  This study establishes and verifies the prognostic risk model for HCC, and provides a reference for the formulation of individualized diagnosis and treatment plan for HCC patients.