Fluorescence in situ hybridization (FISH) has been reported to have much better sensitivity for the detection of bladder transitional cell carcinoma (TCC) than urine cytology. We comparatively tested cytology, FISH, and the cytokeratin-detection test of urinary bladder cancer (UBC) in routine clinical practice.

In a prospective study, FISH, the urinary bladder cancer test (UBC-enzyme-linked immunosorbent assay [ELISA]), and cytology were used in 166 patients. Of the 166 patients, 62 had primary TCC (group 1), 71 had undergone transurethral resection of primary TCC before routine secondary transurethral resection (group 2), and 33 control had not undergone TCC (group 3). All patients with false-positive test results were followed up for a mean follow-up time of 22 months.

The overall sensitivity of FISH, UBC-ELISA, and cytology was 53.2% (95% confidence interval 40% to 66%), 40.3% (95% confidence interval 28% to 53%), and 71.0% (95% confidence interval 59% to 83%), respectively (P <0.05). For grade 3 TCC, both FISH and cytology reached a sensitivity of 93.3%. In the 104 patients without TCC, the specificity of FISH, UBC-ELISA, and cytology was 74.0%, 75.0%, and 83.7%, respectively. During follow-up, 33.3% of patients with a false-positive FISH result developed recurrence, as did 23.1% with false-positive UBC results and 29.4% with false-positive cytology findings (P >0.05). Receiver operating characteristic analysis showed an area under the curve for FISH, UBC, and cytology of 0.636, 0.577, and 0.773, respectively. Only cytology and FISH were significantly predictive of a TCC finding on histologic examination (P <0.001 and P = 0.003, respectively).

In routine clinical practice, conventional cytology in experienced hands can be superior to FISH. False-positive results with all three test systems used warrant a high suspicion of subclinical precursor lesions of TCC recurrence.