We evaluated the performance of the fluorescence in situ hybridization (FISH) assay in comparison with that of urinary cytology for the detection of bladder urothelial carcinoma in routine clinical practice. Voided urine samples from 602 patients with hematuria were analyzed. The bladder cancer group consisted of 95 patients who had biopsy-proven bladder cancer, and the control group consisted of 507 patients without bladder cancer. We found a significant difference between the overall sensitivities of FISH and cytology (60% vs. 28.4%, respectively; P < 0.0001). The overall specificity was 99% with cytology and 94.9% with FISH, although this difference was not statistically significant. The mean values for all four probes in the true-positive group were higher than those in the false-positive group. The difference in the mean values between the two groups was significant only for the CEP3 and CEP17 probes. Furthermore, the severity of the genetic alterations detected by FISH showed a positive correlation with both tumor invasiveness (stage Ta --> T1, T2) and histological grade (G1, G2 --> G3). Together, these findings suggest that FISH can be a useful diagnostic and surveillance tool for patients who are suspected of having new or recurrent bladder cancer.
Copyright 2010 Elsevier Inc. All rights reserved.