Analysis of metabolism-related gene signature for prognosis prediction of clear cell renal cell carcinoma.

e16580 Background: Clear cell renal cell carcinoma (ccRCC) is characterized by a dysregulation of changes in cellular metabolism. However, the prognostic value of metabolism-related genes in ccRCC have not been systematically profiled. In this study, a candidate prognostic gene signature of metabolism in ccRCC was explored. Methods: The clinical and gene expression profiles of ccRCC patients were downloaded from the TCGA, GEO and three clinical trials (CheckMate 009, CheckMate 010, CheckMate 025), and the metabolism-related gene set was downloaded from MSigDB. Differential expression analysis and LASSO Cox regression with binomial deviance minimization criteria were applied to identify and build a metabolism-based signature. The prognostic significance of the signature was further evaluated with the Receiver Operating Characteristic (ROC) curve analysis. Univariate and multivariate Cox regression analysis was performed to evaluate the impact of each variable on OS. Furthermore, the prediction power of the signature has been validated using different ccRCC cohort. Results: In this study, a signature of 8 metabolism-related genes (ANGPT2, ATP6V1B1, CACNA1E, CD163L1, EPN2, HOXD11, PROS1, SHOX2) was constructed as being significantly associated with overall survival (OS) among patients with ccRCC, which differentiated ccRCC patients into high- and low-risk subgroups. The Kaplan-Meier (KM) analysis showed that the survival rate of the low-risk patients was significantly higher than that of the higher-risk patients (hazard ratio (HR) in training set, 0.25 [95% CI, 0.14-0.44; P < 0.001]; testing set, 0.28 [95% CI, 0.10-0.76; P = 0.008]; validation cohort (clinical trials), 0.47 [95% CI, 0.33-0.68; P < 0 .001]; validation cohort (GSE29609), 0.25, [95% CI, 0.08-0.88; P = 0 .01]). ROC curve analysis of the prognostic signature showed that the areas under curve (AUC) for the 1-, 3-, and 5-year OS in all cohort were more than 0.70 (AUC of the signature for 3 year in the training set and validation cohort were 0.816 and 0.708, respectively, and 0.807 and 0.702, respectively, for the 5- year OS). Further more, a nomogram based on the signature was constructed and showed an accurate prediction for prognosis in ccRCC. Conclusions: Taken together, we identified the key metabolism-related genes and constructed a robust prognostic signature for the prognostic predictor of ccRCC patients, which maybe help personalized management of ccRCC patients.