Abstract
The primary aim of the study was to analyze novel long non-coding RNAs (lncRNAs) in hepatocellular carcinoma (HCC) to assess their roles as potential oncogenes and tumor suppressors and to develop a survival prediction model based on their expression levels. Data from The Cancer Genome Atlas, GSE135631, and GSE214846, were utilized to evaluate changes in lncRNA expression in HCC and their associations with patient prognosis. A risk model was created based on lncRNA expression to predict patient mortality. The co-expression network was employed to identify associated pathways, and the results were subsequently validated using the RT-qPCR method. The findings indicated that 14 lncRNAs were down-regulated in HCC, and their increased expression was associated with a favorable prognosis. Additionally, eight lncRNAs were overexpressed and correlated with poorer patient outcomes. The multivariate Cox regression analysis demonstrated that overexpression of AKR1B10P1, RP11-465B22.3, WASH8P, and the downregulation of NPM1P25 could independently predict patient survival, irrespective of clinical variables. The risk score model based on these lncRNAs effectively stratified patients by their mortality risk. Furthermore, the co-expression network analysis revealed that the identified lncRNAs might be involved in various pathways, including fatty acid metabolism, mTOR signaling, glycolysis, angiogenesis, Wnt-β-catenin pathway, and DNA repair. RT-qPCR results validated the significant increase in the expression level of WASH8P in cancer specimens compared to normal tissues. Our results unveiled that changes in the expression levels of AKR1B10P1, RP11-465B22.3, WASH8P, and NPM1P25 were significantly and independently associated with prognosis. Moreover, the patient mortality risk model constructed using these four lncRNAs exhibited a robust capacity to accurately predict patients' survival rates.