Development and validation of a novel prognostic model for hepatocellular carcinoma

Objective To establish a prognostic model for hepatocellular carcinoma patients with good predictive performance and low cost. Methods The transcriptome data and clinical information were downloaded from the Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC). The differential genes associated with overall sunival (OS) were screened. TCGA cohort was modeled using permutation and the predictive performance was evaluated by the AUC of ROC curve based on TCGA cohort and ICGC cohort. Results A total of 692 prognosis-related differential genes were screened out, with the best gene combinations of DPYSL4, DTYMK, and HOXD9. Then, a three-gene model was constructed where patients were classified into two risk groups. Patients in the high-risk group showed significantly reduced OS compared with those in the low-risk group (P < 0.001 in both TCGA and ICGC cohorts). The risk score was an independent predictor for OS in multivariate Cox regression analysis (HR>1 and P<0.001). Functional analysis revealed that ribosome-related, cell cycle, and metabolism-related pathways were enriched. Conclusions This study establishes a novel hepatocellular carcinoma prognostic model that can be used to predict the prognosis of hepatocellular patients.