METHODS: Bioinformatics analysis was performed in four independent microarray data including 56 non-muscle-invasive bladder cancer (NMIBC) and 132 muscle-invasive bladder cancer (MIBC) tissues from Gene Expression Omnibus to screen out differentially expressed genes. Next, we detected CDH11 expression in BCA specimens and cell lines by qPCR and western blotting assays. Immunohistochemical analyses were performed in 209 paraffin-embedded BCA samples and 30 adjacent normal bladder tissues.
RESULTS: Bioinformatics analysis revealed that CDH11 had a higher expression level in MIBC tissues than in NMIBC, which was consistent with our clinical BCA specimens and cell lines at both mRNA and protein levels. Immunohistochemical analysis demonstrated that over-expression of CDH11 was closely related to the histological grade, pT status, tumour size and poor outcomes of BCA patients. What\'s more, CDH11 (area under curve (AUC) = 0.673 and 0.735) had a better predictive value than E-cadherin (AUC = 0.629 and 0.629) and a similar discrimination with the European Organization for Research and Treatment of Cancer (EORTC) score system (AUC = 0.719 and 0.667) in evaluating potential recurrence and progression of NMIBC. Moreover, combination of CDH11 and EORTC score system was the best predictive model in predicting recurrence of NMIBC (AUC = 0.779) among the three models.
CONCLUSIONS: CDH11 was a reliable therapeutic target in BCA and a useful index to predict the possibilities of recurrence and progression in NMIBC patients.