Construction and evaluation of a prognostic model for metabolism-related genes in kidney renal clear cell carcinoma using TCGA database.

Objective: To investigate the expression of metabolism-related genes (MRGs) in kidney renal clear cell carcinoma (KIRC) and their association with patient prognosis, and to identify potential targets for intervention.

Methods: Bioinformatics methods were employed to mine the KIRC transcriptome data in The Cancer Genome Atlas Program (TCGA) database in order to identify MRGs that are aberrantly expressed in cancerous tissues. Subsequently, a prognostic risk score model was constructed and its predictive capacity was evaluated. Finally, the expression of prognostically relevant MRGs was validated using external datasets and KIRC clinical samples.

Results: A total of 789 differentially expressed MRGs associated with KIRC were screened, of which 465 genes were upregulated and 324 genes were downregulated, and finally 23 genes were screened to establish a risk score model. We found that the AUCs of the risk score model for predicting patients' 1-, 3- and 5-year overall survival (OS) were 0.804, 0.766 and 0.802, respectively. These findings suggest that the model has good predictive ability. A multifactorial Cox analysis revealed that 23 MRGs risk score was significantly associated with the overall survival of KIRC patients, and could therefore be used as an independent risk factor for the prognosis of KIRC patients (HR = 3.495, P < 0.001). Meanwhile, Kaplan-Meier analyses of the high-risk and low-risk groups indicated that the high-risk group exhibited a markedly inferior overall survival (OS) prognosis. The validation of clinical samples from KIRC patients and four external data sets (GSE36895, GSE40435, GSE53757 and GSE66272) demonstrated that KCNN4 and PLK1 were highly expressed in KIRC, whereas TEK, PLG, ANGPTL3, TFAP2A, ANK3, ATP1A1 and UCN exhibited low expression in KIRC.

Conclusion: Several MRGs are aberrantly expressed in KIRC, from which we screened 23 genes and constructed a MRGs prognostic risk model that can effectively predict the prognosis of KIRC patients and provide a new foundation for personalised diagnosis and treatment of KIRC.