Anticipatory Positive Urine Tests for Bladder Cancer

Purpose The aim of this study was to establish the criteria defining an anticipatory positive test for bladder cancer. Methods We reviewed all patients at our institution who underwent urine cytology or UroVysion fluorescence in situ hybridization (FISH) and cystoscopy from 2003 to 2012. Test performance and cancer anticipation was assessed using generalized linear mixed models, mixed-effects proportional hazards models, and cumulative incidence curves using tests performed within 30 days of each other as well as within a lag time of 1 year. Results Overall, 6729 urine tests (4729 cytology and 2040 UroVysion FISH) were paired with gold-standard cystoscopies. Sensitivity and specificity were 63 and 41% for cytology, and 37 and 84% for UroVysion FISH, respectively. A 1-year lag time allowed for cancer anticipation and neither test improved. Among patients with positive cytology and initially negative cystoscopy, the hazard ratio of developing a bladder tumor at 1 year was 1.83; 76% of these patients developed a tumor within 1 year. Similarly, among patients with a positive FISH and initially negative cystoscopy, the hazard ratio of developing a bladder tumor at 1 year was 1.56; 40% of these patients developed a tumor within 1 year. Conclusions Urine-based tests for bladder cancer are frequently falsely positive. With further follow-up time, some of these false positive tests are vindicated as true (anticipatory) positive tests, although many will remain false positives. We developed statistical criteria to determine if a test anticipates future cancers or not.