Comparison of conventional and liquid-based cytology using The Paris System for Reporting Urinary Cytology

Background This retrospective study was conducted to compare the conventional cytospin method and ThinPrep liquid-based urinary cytology in diagnosing bladder cancer using The Paris System (TPS) of classification. Methods We retrieved files for 2020, at the Cytopathology Department of Laiko Hospital, of urinary cases diagnosed according to TPS. Cytospin and ThinPrep slides were separately reviewed and new diagnoses were rendered, then compared with the original diagnosis and histology when available. Risk of high-grade malignancy (ROHM) for each TPS category was assessed, along with accuracy parameters of each method and their combination. Results The study material comprised 100 cases of void urinary cytology classified as 20 high-grade urothelial carcinoma (HGUC = TPS5) cases, 20 of suspicion for HGUC (SHGUC = TPS4), 25 of atypical urothelial cells (AUC = TPS3), and 35 of negative for HGUC (NHGUC = TPS2). A single inadequate (TPS1) case and 4 of low-grade urothelial neoplasm (TPS6) were excluded as small in number. The ROHM was 95% for HGUC, 55% for SHGUC, 28% for AUC and 5.7% for NHGUC. Agreement with the original diagnosis was 86% for cytospin and 82% for ThinPrep. No significant differences were observed among the two techniques or their combination regarding sensitivity and specificity, with a mild advantage for cytospin. Interobserver reproducibility and repeatability were high. Conclusion No significant differences were found concerning sensitivity and specificity between cytospin and ThinPrep when applying TPS criteria. TPS is a reliable classification scheme for either conventional/cytospin or liquid-based cytology, or their combination.