The Effects of Instrumentation on Urine Cytology and CK-20 Analysis for the Detection of Bladder Cancer

Abstract

Objective To evaluate the effects of cystoscopy on urine cytology and additional cytokeratin-20 (CK-20) staining in patients presenting with gross hematuria. Patients and Methods For 83 patients presenting with gross hematuria, spontaneous and instrumented paired urine samples were analyzed. Three patients were excluded. Spontaneous samples were collected within 1 hour before cystoscopy, and the instrumented samples were tapped through the cystoscope. Subsequently, patients underwent cystoscopic evaluation and imaging of the urinary tract. If tumor suspicious lesions were found on cystoscopy or imaging, subjects underwent transurethral resection or ureterorenoscopy. Two blinded uropathological reviewers (DB, KK) evaluated 160 urine samples. Reference standards were results of cystoscopy, imaging, or histopathology. Results Thirty-seven patients (46.3%) underwent transurethral resection or ureterorenoscopy procedures. In 30 patients (37.5%) tumor presence was confirmed by histopathology. The specificity of urine analysis was significantly higher for spontaneous samples than instrumented samples for both cytology alone (94% vs 72%, P =.01) and for cytology combined with CK-20 analysis (98% vs 84%, P =.02). The difference in sensitivity between spontaneous and instrumented samples was not significant for both cytology alone (40% vs 53%) and combined with CK-20 analysis (67% vs 67%). The addition of CK-20 analysis to cytology significantly increases test sensitivity in spontaneous urine cytology (67% vs 40%, P =.03). Conclusion Instrumentation significantly decreases specificity of urine cytology. This may lead to unnecessary diagnostic procedures. Additional CK-20 staining in spontaneous urine cytology significantly increases sensitivity but did not improve the already high specificity. We suggest performing urine cytology and CK-20 analysis on spontaneously voided urine.