Cystoscopy revisited as the gold standard for detecting bladder cancer recurrence: diagnostic review bias in the randomized, prospective CEFUB trial

Abstract

Purpose: We evaluated the influence of knowledge of urine test outcome on the accuracy of cystoscopy (diagnostic review bias) during surveillance in patients with low grade, nonmuscle invasive urothelial carcinoma.

Materials and methods: We performed a prospective, single-blind, randomized, multicenter clinical trial of surveillance by microsatellite analysis urine test in 448 patients with nonmuscle invasive (pTa, pT1, G1, G2) urothelial carcinoma. Positive or negative urine test results were only communicated to the urologist in the intervention arm of 226 patients, in which cystoscopy was done if the test was positive, and at 3, 12 and 24 months. Urine test results were not communicated in the control arm of 222 patients who underwent standard 3-month cystoscopy. The primary outcome measure was the number of histologically proven bladder cancer recurrences.

Results: At a median 34-month followup 218 recurrences were detected in the intervention arm compared to 163 in the control arm (p <0.001). Of 131 cystoscopies done with knowledge of a positive urine test 42 recurrences were detected. Only 6 recurrences were found in the 120 cystoscopies done without information on the positive test result (chi-square p <0.001). There was no difference in recurrence detection when urine test results were negative in the intervention and control arms (18 of 260 patients or 7% and 18 of 326 or 6%, respectively, p = 0.45).

Conclusions: Diagnostic review bias should be considered in the evaluation of point of care urine tests for bladder cancer monitoring. Awareness of a positive urine test result significantly improves the urothelial carcinoma detection rate using cystoscopy.