Screening hypoxia-related genes as prognostic biomarkers and modeling the individualized prognostic predictor in hepatocellular carcinoma

Background: Hypoxia closely relates to malignant progression and appears to be prognostic for outcome in hepatocellular carcinoma (HCC). Our research aims to mine the Hypoxic related genes (HRGs) on the role of clinical prognosis in HCC. Moreover, we construct and define a model of prognostic predictor (PP) to estimate and improve prognosis of HCC patients.Results 37 differentially expressed HRGs were obtained. It contained 28 up-regulated and 9 down-regulated genes. After the univariate Cox regression model analysis, we obtained 27 prognosis-related HRGs. Of these, 25 genes were risk factors for cancer and 2 genes were protective factors. The PP was composed of the 10 key genes (HDLBP, SAP30, PFKP, DPYSL4, SLC2A1, PGK1, ERO1A, LDHA, ENO2, TPI1), and significantly divided patients of HCC into high- and low-risk groups according to overall survival (OS) ( P <0.001). We got the Area Under Curve(AUC) value of risk score calculated by PP was 0.777, which much bigger than other clinical parameters. Besides, PP was verified as an independent prognosis-related parameter (in univariate analyse, HR=1.484, 95% CI=1.342–1.642, P<0.001; in multivariate analyse, HR=1.414, 95% CI=1.258–1.588, P <0.001). Finally, the application of PP in clinic was concluded that the higher the patient's risk score, the higher the corresponding tumor stage and T stage, and the patient's prognosis was poor.Conclusions This study provides hypoxic related molecular targets for the therapeutic intervention. In addition, an individualized prognostic predictor was constructed to predict prognosis for HCC patients .