Surveillance of patients with bladder carcinoma using fluorescent in-situ hybridization on bladder washings

OBJECTIVES To compare the sensitivity and specificity of the UroVysion((TM)) (Abbott Laboratories Inc., Downers Grove, IL, USA) fluorescent in-situ hybridization (FISH) assay to that of urinary cytology obtained from bladder irrigation during cystoscopic surveillance in patients with bladder carcinoma. PATIENTS AND METHODS The medical records were retrospectively reviewed for 41 consecutive patients screened at the authors' institution between August 2000 and December 2006 for recurrence of pathologically confirmed bladder cancer. All 162 cytology examinations and 141 FISH assay results obtained from bladder washing were included. Recurrence was determined by cystoscopy, bladder biopsy and upper-tract imaging. Sensitivity, specificity, positive predictive and negative predictive values were assessed using a chi-square distribution with one degree of freedom. RESULTS There were 24 men and 17 women (male to female ratio 0.59), the mean (range) age was 56 (33-73) years and the mean follow-up 30 (2-57) months. At the initial diagnosis, 35 of the 41 patients (85%) had superficial tumours (stage <= T1), while six (15%) had muscle-invasive tumours (stage >= T2). Twenty-six (63%) had low-grade and 15 (37%) had high-grade tumours. In 16 of 141 (11%) of the FISH assays and 16 of 162 (10%) of the cytological samples that were collected from bladder irrigations, there were too few cells for an adequate analysis. The FISH assay correctly correlated with subsequent cystoscopy, bladder biopsy or upper-tract imaging in 110/125 (88%) cases but not in 15/125 (12%). Cytology correctly correlated with the subsequent evaluation in 112/146 (77%) cases but did not in 34/146 (23%). When the FISH was compared with cytology in this setting, the sensitivity was 77% (30/39) vs 74% (37/50; P > 0.1), the specificity was 93% (80/86) vs 78% (75/96; P < 0.01), the positive predictive value was 83% (30/36) vs 64% (37/58; P < 0.05), and the negative predictive value was 90% (80/89) vs 85% (75/88; P > 0.1), respectively. CONCLUSION The UroVysion FISH assay obtained from bladder washings during cystoscopic surveillance of patients with a history of bladder cancer provides a similar specificity but greater sensitivity than that of cytology for detecting bladder cancer recurrences. Given the better specificity and similar sensitivity of UroVysion compared with urine cytology obtained from bladder washings, a reasonable approach might be to use the UroVysion assay as the primary marker for recurrence, with urine cytology used as a complementary examination.