Prospective analysis of sensitivity and specificity of urinary cytology and other urinary biomarkers for bladder cancer

Introduction: Although urinary cytology (C) is the most widely used noninvasive test for the detection and surveillance of bladder cancer (BC), it has poor sensitivity especially for low-grade tumors. We prospectively tested the performance of urine markers on patients with BC using C and 4 commercially available urinary marker tests: Hemoglobin Dipstick (H), BTA Stat (B), NMP22 BladderChek (N), and ImmunoCyt (I). Methods: Urinary samples from 109 consecutive patients with BC were prospectively collected. All samples were tested using conventional C and available biomarkers. Prior and subsequent surgical specimen reports were examined, and sensitivity and specificity were calculated for each. Collected variables included patient demographics, date of urinary collection, type of specimen (voided, washing, or catheterized), surgical pathology, recurrence, and follow-up. Results: Sensitivity values for each marker were as follows: C, 48% (16% for low-grade tumors and 84% for high-grade [HG] tumors); B, 61% (36% and 91%); H, 51% (38% and 66%); N, 58% (25% and 92%); and I, 62% (47% and 83%). Specificity results for each marker were as follows: C, 86%; B, 78%; H, 58%; N, 85%; and I, 79%. On multivariate analysis, higher stage (C and N) and HG disease (C, H, B, N, and I) were independent prognostic factors for improved test performance. When urinary markers were combined with C, sensitivity/specificity values for HG disease were as follows: C+H, 85%/57%; C+B, 91%/78%; C+N, 94%/84%; and C+I, 90%/78%. Conclusions: Based on these data, C seems to yield the highest specificity and N the highest sensitivity for HG tumors. The combination "C+N" seems to be the better approach to improve the sensitivity for HG tumors compared with single markers and other combinations. (C) 2014 Elsevier Inc. All rights reserved.