Negative predictive value and sensitivity of urine cytology prior to implementation of The Paris System for Reporting Urinary Cytology

Abstract

Urinary tract cytology (UTCy) is used for screening urothelial carcinoma (UC) and it must have a high negative predictive value (NPV) to be an effective test. To the authors' knowledge, the literature regarding the NPV of UTCy provides little information regarding the risk of malignancy, especially for patients with high-grade urothelial carcinoma (HGUC). Patients with negative UTCy specimens were identified in the pathology files at the study institution for the years 2012 through 2013. Cases were deemed true-negative cases if there was at least 1 subsequent negative specimen or negative clinical follow-up within 6months of the index case. False-negative cases were defined as HGUC or carcinoma in situ by surgical biopsy and/or any UTCy with suspicious for HGUC or HGUC follow-up. A total of 2614 UTCy specimens from 2089 patients were identified. There was a disease prevalence of 6.5%. There were 87 false-negative results for HGUC, which corresponded to an overall NPV of 96.7%. When categorized by clinical indication, hematuria resulted in the highest NPV of 99.5% followed by other indications (97.7%) and a history of UC (90.1%). When categorized by the specimen type, voided urine specimens were found to have the highest NPV of 98.7% followed by other indications (96.9%) and washing specimens (96.2%). Of the 717 patients with a history of UC, the NPV was lower for washing specimens (89.8%) than for voided urine specimens (96.2%). When including either low-grade urothelial carcinoma or HGUC as a positive follow-up, the NPV dropped to 93.3% from 96.7% (HGUC only). The sensitivity of the diagnostic category of atypical urothelial cells or higher was 93.0%. Overall, UTCy appears to have a good NPV and a high sensitivity for HGUC. The clinical indication had a greater impact on NPV compared with the specimen type.