Significant reduction of indeterminate (atypical) diagnosis after implementation of The Paris System for Reporting Urinary Cytology: A single-institution study of more than 27,000 cases.

Abstract

Urinary cytology is a noninvasive and cost-effective diagnostic and surveillance test in the clinical management of urothelial carcinoma (UC). The Paris System for Reporting Urinary Cytology (TPS), published in 2016, introduced definite diagnostic criteria aimed at improving performance in detecting high-grade UC (HGUC) and decreasing the indeterminate (atypical) diagnosis. The authors retrospectively reviewed and compared urinary cytology diagnoses reported between January 2013 and December 2014 (pre-TPS, 7658 cases) and between May 2016 and April 2018 (post-TPS, 20,026 cases) to assess the influence of TPS in their practice. The time in between was used as a learning period. Follow-up information and correlation with the UroVysion fluorescence in situ hybridization test were obtained when available. Urinary cytology diagnoses pre-TPS included negative for UC (NUC) (n = 5293; 69.2%), atypical urothelial cells (AUC) (n = 2227; 29%), and suspicious/positive for HGUC (SHGUC/HGUC) (n = 138; 1.8%). Diagnoses post-TPS included negative for HGUC (NHGUC) (n = 18,507; 92.4%), AUC (n = 1237; 6.2%), and SHGUC/HGUC (n = 282; 1.4%). Comparing the pre-TPS and post-TPS periods, AUC diagnoses decreased from 29% to 6.2% (P < .00001), and the specificity and positive predictive value of AUC to detect HGUC significantly improved from 49% to 86% (P < .00001) and from 9% to 39% (P = .002), respectively. The correlation of an AUC diagnosis with a positive UroVysion test improved from 17% to 38% (P < .00001), whereas overall use of the UroVysion test was decreased. Implementation of TPS resulted in a significant reduction in AUC diagnoses that had a superior correlation with a subsequent biopsy and a UroVysion test, resulting in potential reductions in test use and medical cost.