A novel clinical multidimensional transcriptome signature predicts prognosis in bladder cancer

Abstract

A number of studies has shown that long non-coding RNAs (lncRNAs), microRNAs (miRNAs) and protein coding genes (PCGs) are involved in various pathophysiological processes and can be used as prognostic biomarkers in cancer patients. The purpose of this study was to find a multidimensional transcriptome signature to predict clinical outcomes in bladder cancer. Using Cox's proportional hazards regression analysis and the random survival forest algorithm, we mined the expression profile data of 239 bladder cancer patients derived from The Cancer Genome Atlas (TCGA) public database. A signature comprised of two PCGs (ACADS and C1QTNF9B), two lncRNAs (RP11-60L3.1 and CTD-3195I5.3) and two microRNAs (has-miR-3913-1 and has-miR-891a) with highest accuracy prediction (AUC=0.79 in the training dataset and 0.64 in the test dataset) was selected. The signature had an ability to stratifypatientsintohigh-andlow-riskgroupswithsignificantly different survival rates (median 16.9 vs. 54.9 months, log-rank test P<0.001) in the training dataset, and its performance was validated for risk stratification in the test dataset (median 18.2 vs. 58.9 months, log-rank test P=0.002). Multivariable Cox regression analysis revealed that the signature was an independent prognostic factor for patients with bladder urothelial carcinoma (BLCA). A comparison of tumour node metastasis (TNM) stage and the signature indicated that the signature had better survival prediction power (AUCsignature=0.79/0.64 vs. AUCTNM=0.67/0.60, P<0.05). Functional analyses indicated that these prognostic genes from the signature may be involved in tumourigenesis-related biological processes and pathways. In conclusion, the multidimensional PCG-lncRNA-microRNA signature can be a novel prognostic marker to predict the survival of bladder cancer patients.