RESUMO
OBJECTIVE: The development of the Paris System (TPS) has provided a standard and reproducible system for reporting urine cytopathology. Our goal was to study the impact of TPS on the diagnostic accuracy of urine cytology since we began using it in 2016. METHODS: We performed a retrospective study of all urine cytology specimens received at our institution from January 2015 through July 2017. Cases were included in the study if they had corresponding surgical pathology follow up. In total, 3829 cases were identified over this time period, with 381 cases meeting inclusion criteria, comprising 87 cases from 2015, 166 from 2016 and 128 from 2017. Using the histopathology diagnosis as the gold standard, sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV) and diagnostic accuracy (DA) for the detection of carcinoma were calculated. RESULTS: Before TPS, urine cytology had a sensitivity of 100.0%, specificity of 12.5%, PPV of 83.5%, NPV of 100.0% and DA of 83.9%. After TPS, for 2016, urine cytology had sensitivity of 87.1%, specificity of 95.9%, PPV of 96.4%, NPV of 85.4% and DA of 91.0%. For 2017 after TPS, the sensitivity was 81.7%, specificity was 100.0%, PPV was 100.0%, NPV was 81.4% and DA was 89.8%. CONCLUSION: For the detection of urinary tract malignancy, after switching to TPS, we observed a marked increase in urine cytology specificity and PPV, both of which continued to gradually increase from 2016 to 2017. The DA also improved with TPS.