Abstract
BACKGROUND: CTNNB1 mutates in most hepatocellular carcinoma (HCC) which is the most familiar form of liver cancer with high heterogeneity. It is critical to create a specific prognostication methodology and to investigate additional treatment options for CTNNB1-mutant HCCs. METHODS: A total of 926 samples in five independent cohorts were enrolled in this study, including 127 CTNNB1-mutant samples and 75 estimated CTNNB1-mutant samples. The prognostic signature was constructed by LASSO-Cox regression and evaluated by bioinformatics analyses. The selection of possible drug targets and agents was produced based on the expression profiles and drug sensitivity data of cancer cell lines in two databases. RESULTS: A prognostic signature based on 15 genes categorized the CTNNB1-mutant HCCs into two groups with different risks. Compared to low-risk patients, high-risk patients had significantly inferior prognoses. ROC curve and multivariate analysis also indicated the superior performance of our signature on the prognosis estimation, particularly in CTNNB1-mutant HCCs. Besides, the nomogram was constructed according to the prognostic signature with excellent predictive performance confirmed by the calibration curve. Subsequently, we suggested that AT-7519 and PHA-793887 might be potential drug agents for high-risk patients. CONCLUSION: We established a 15-gene prognostic model, particularly in HCCs with CTNNB1 mutations with good predictive efficiency. Besides, we explored the potential drug targets and agents for patients with high risk. Our findings offered a fresh idea for personalized prognosis management in HCCs with CTNNB1 mutations and threw new insight for precise treatment in HCCs as well.