Abstract
Background: Hepatocellular carcinoma (HCC) refers to the malignant tumor associated with a high mortality rate. This work focused on identifying a robust tumor glycolysis-immune-related gene signature to facilitate the prognosis prediction of HCC cases. Methods: This work adopted t-SNE algorithms for predicting glycolysis status in accordance with The Cancer Genome Atlas (TCGA)-derived cohort transcriptome profiles. In addition, the Cox regression model was utilized together with LASSO to identify prognosis-related genes (PRGs). In addition, the results were externally validated with the International Cancer Genome Consortium (ICGC) cohort. Results: Accordingly, the glycolysis-immune-related gene signature, which consisted of seven genes, PSRC1, CHORDC1, KPNA2, CDCA8, G6PD, NEIL3, and EZH2, was constructed based on TCGA-HCC patients. Under a range of circumstances, low-risk patients had extended overall survival (OS) compared with high-risk patients. Additionally, the developed gene signature acted as the independent factor, which was significantly associated with clinical stage, grade, portal vein invasion, and intrahepatic vein invasion among HCC cases. In addition, as revealed by the receiver operating characteristic (ROC) curve, the model showed high efficiency. Moreover, the different glycolysis and immune statuses between the two groups were further revealed by functional analysis. Conclusion: Our as-constructed prognosis prediction model contributes to HCC risk stratification.