Abstract
Previous studies have shown that the miR-17-92 cluster is involved in the occurrence and development of bladder cancer. However, the role of serum miR-17-92 cluster in the diagnosis of bladder cancer has not been studied. In the present study, we evaluated the expression of miR-17-92 cluster members in bladder cancer tissues by analyzing 428 cases from TCGA database. Next, we collected the sera of 74 bladder cancer patients and 90 controls, and used qRT-PCR to detect the relative expression of the cluster. The results showed that the expression of the cluster members in the sera of patients were significantly higher than that of the controls, and they were positively correlated with the clinical stage and pathological grade of the patients. We evaluated their ability to diagnose bladder cancer using ROC, of which miR-92a-3p (AUC = 0.902), miR-17-5p (AUC = 0.845) and miR-20a-5p (AUC = 0.806) were the most prominent. Finally, we established a diagnostic model by logistic regression (AUC = 0.969). We further validated the results of the study using another dataset from the GEO database. Moreover, we evaluated the prognostic value of the cluster. The results revealed that miR-20a-5p was correlated with recurrence of bladder cancer. In summary, the present study validated the overexpression of serum miR-17-92 cluster in bladder cancer. The model composed of the three cluster members were confirmed to be a promising noninvasive biomarker for bladder cancer diagnosis.