BACKGROUND: Early detection of bladder cancer and its recurrence is the key to long term survival in patients. However, because of nonspecific symptoms, it is difficult to verify such assumption. The UroVysion(R) fluorescence in situ hybridization (FISH) assay (Vysis, Downers Grove, IL, USA) is a new method to detect early bladder cancer. It can detect specific genetic alterations (aneuploidy of chromosome 3, 7, 17, and loss of 9p21) observed in bladder cancer. To investigate clinical usefulness, we evaluated UroVysion(R) on bladder cancer patients with various status. METHODS: A total of 56 patients were evaluated. Single voided urine specimens were obtained, and four types of probes (CEP3, CEP7, CEP17, and LSI9p21) were applied to each specimen according to the conventional FISH method. The results were interpreted according to the manufacturer's criteria. RESULTS: Of the 56 cases evaluated, 33 were being followed after treatment, 8 were newly diagnosed with bladder cancer, 5 had recurred, and 10 had hematuria with no evidence of cancer. The overall sensitivity of FISH, BTA stat test, and urine cytology was 71.4%, 57.1%, and 57.1%, and the specificity was 94.3%, 97.3%, and 89.7%, respectively. Especially, the sensitivity of FISH was 75% (6/8) in the newly diagnosed patients and 60% (3/5) in recurrent cases. A false positivity in FISH was shown in patients with urinary stone. CONCLUSIONS: FISH assay is the most sensitive method in the detection of bladder cancer, and also shows a good specificity. If other tests are done simultaneously, FISH assay will be very useful in detecting bladder cancer. However, whether FISH can replace cystoscopy is yet to be assessed.