Abstract
BACKGROUND: Kidney renal clear cell carcinoma (KIRC) is a potentially fatal urogenital disease. It is a major cause of renal cell carcinoma and is often associated with late diagnosis and poor treatment outcomes. More evidence is emerging that genetic models can be used to predict the prognosis of KIRC. This study aimed to develop a model for predicting the overall survival of KIRC patients. RESULTS: We identified 333 differentially expressed genes (DEGs) between KIRC and normal tissues from the Gene Expression Omnibus (GEO) database. We randomly divided 591 cases from The Cancer Genome Atlas (TCGA) into training and internal testing sets. In the training set, we used univariate Cox regression analysis to retrieve the survival-related DEGs and futher used multivariate Cox regression with the LASSO penalty to identify potential prognostic genes. A seven-gene signature was identified that included APOLD1, C9orf66, G6PC, PPP1R1A, CNN1G, TIMP1, and TUBB2B. The seven-gene signature was evaluated in the training set, internal testing set, and external validation using data from the ICGC database. The Kaplan-Meier analysis showed that the high risk group had a significantly shorter overall survival time than the low risk group in the training, testing, and ICGC datasets. ROC analysis showed that the model had a high performance with an AUC of 0.738 in the training set, 0.706 in the internal testing set, and 0.656 in the ICGC external validation set. CONCLUSION: Our findings show that a seven-gene signature can serve as an independent biomarker for predicting prognosis in KIRC patients.