Development of Prognostic Evaluation Model to Predict the Overall Survival and Early Recurrence of Hepatocellular Carcinoma

Background: The aberrant expressions of lncRNAs have been frequently demonstrated to be closely associated with the prognosis of patients in many cancer types including hepatocellular carcinoma (HCC). Integration of these lncRNAs might provide accurate evaluation of HCC. Therefore, this study aims to develop a novel prognostic evaluation model based on the expression of lncRNAs to predict the survival of HCC patients, postoperatively. Patients and Methods: RNA sequencing (RNA-seq) analysis was performed for 61 HCC patients (training cohort) to screen prognosis-associated lncRNAs with univariate Cox regression and Log rank test analyses. Multivariate Cox regression analysis was then applied to establish the final model, which was further verified in a validation cohort (n=191). Moreover, performance of the mode was assessed with time-dependent receiver operating characteristic curve (tdROC), Harrell's c-index, and Gonen & Heller's K. Results: After a serial statistical computation, a novel risk scoring model consisting of four lncRNAs and TNM staging was established, which could successfully divide the HCC patients into low-risk and high-risk groups with significantly different OS and RFS in both training and validation cohorts. tdROC analysis showed that this model achieved a high performance in predicting OS and 2-year RFS in both cohorts. Gene Set Enrichment Analysis revealed that HCC tumor tissues with high-risk score have stronger capacities in immune escape and resistance to treatment. Conclusion: We successfully established a novel prognostic evaluation model, which exhibited reliable capacity in predicting the OS and early recurrence of HCC patients with relatively higher accuracy.