Evaluation of the Fluorescence In Situ Hybridization Test to Predict Recurrence and/or Progression of Disease after bacillus Calmette-Guerin for Primary High Grade Nonmuscle Invasive Bladder Cancer: Results from a Prospective Multicenter Trial

Purpose: Single center studies have shown that positive UroVysion (R) fluorescence in situ hybridization results were associated with recurrence of nonmuscle invasive bladder cancer treated with intravesical bacillus Calmette-Guerin. Our goal was to validate these findings. Materials and Methods: We performed a prospective, multicenter diagnostic trial to determine whether the fluorescence in situ hybridization test could predict recurrence or progression in patients with primary high grade nonmuscle invasive bladder cancer who were scheduled to receive bacillus Calmette-Guerin. Fluorescence in situ hybridization testing was performed prior to the first bacillus Calmette-Guerin instillation, prior to the sixth instillation and at 3-month cystoscopy. The performance of fluorescence in situ hybridization was evaluated. Results: A total of 150 patients were enrolled in analysis, including 68 with Ta disease, 41 with T1 disease, 26 with carcinoma in situ alone and 15 with papillary carcinoma plus carcinoma in situ. At 9 months of followup there were 46 events, including 37 recurrences and 9 progressions. For events with positive fluorescence in situ hybridization findings the HR was 2.59 (95% CI 1.42-4.73) for the baseline test, 1.94 (95% CI 1.04-3.59) for the 6-week test and 3.22 (95% CI 1.65-6.27) at 3 months. Patients with positive results at baseline, 6 weeks and 3 months had events 55% of the time and patients with negative results at each time point had no event 76% of the time. Conclusions: The study validated that a positive UroVysion fluorescence in situ hybridization test was associated with a 3.3-fold increased risk of recurrence. The test may be useful to risk stratify patients entering clinical trials in whom induction therapy fails. However, using the test to change management decisions is limited due to the discordance between results and outcomes as well as the variance of tests results with time.