Abstract
Bladder cancer (BC) is the most common malignancy involving the urinary system, and is characterized by a high recurrence rate. It is important to identify potential lncRNA signatures capable of predicting tumour recurrence risk and assessing recurrence prognosis in BC patients. We extracted data from The Cancer Genome Atlas and identified 381 differentially expressed lncRNAs, 855 mRNAs and 70 miRNAs between non-recurrent and recurrent BC tissues. Subsequently, a competing endogenous RNA (ceRNA) network composed of 29 lncRNAs, 13 miRNAs and 4 mRNAs was established. We used univariate and multivariate Cox regression to analyse the relationship between the 29 lncRNAs and recurrence-free survival (RFS) in BC patients. Six lncRNAs had significant prognostic values, and their cumulative risk score indicated that this 6-lncRNA signature independently predicted RFS in BC patients. We applied a receiver operating characteristic (ROC) analysis to assess the efficiency of our prognostic models. High-risk patients exhibited a poorer prognosis than low-risk patients did. Additionally, the 6-lncRNA signature showed a significant correlation with BC clinicopathological characteristics, which indicates that it could be used for effective risk stratification. The current study provides novel insights into the lncRNA-related ceRNA network and this 6-lncRNA signature may be an independent prognostic factor in predicting the recurrence of BC patients.