Is prostatic adenocarcinoma detectable by urine cytology-A multicenter retrospective review.

Abstract

Urine cytology is robust for the diagnosis of urothelial lesions, but data on the detection rates of prostatic adenocarcinoma in urine cytology is limited. In this study, a multicenter review was performed to define the clinical role of urine cytology in diagnosis of prostatic adenocarcinoma. Cytologic diagnoses of lower tract urine cytology specimens with histology-proven prostatic adenocarcinoma from three institutions, from a period of over two decades, were reviewed. Clinicopathological parameters-tumor grade, stage, histologic features, and preanalytical factors-prostate-specific antigen (PSA) level and lesion size, were retrieved and compared with cytologic diagnoses. In total, 2115 urine cytology specimens from 1119 patients were retrieved. The atypia (or above/C3+) and suspicious (or above/C4+) rates were 19.48% and 3.36%. Bilobar and extracapsular involvement, lymphovascular invasion, Gleason score, and International Society of Urological Pathology grade were associated with a positive urine diagnosis (p < 0.05). The atypia (C3+) and suspicious (C4+) rates of urine cytology in patients with a PSA level of ≤4.0 ng/mL was paradoxically higher (p < 0.01), but PSA levels correlated positively with urine diagnosis at higher cutoffs (>10, >20, >50, >100 ng/mL). All these factors remained significant on multivariate analysis (p < 0.05), including a negative correlation with low-PSA (≤4.0 ng/mL, p = 0.001) and positive correlation with high-PSA (>20 ng/mL, p = 0.020). Lesion size and multifocality were not associated with urine cytology diagnosis (p > 0.05). Urine cytology showed low sensitivity in detection of prostatic adenocarcinoma. Detection rates were largely positively correlated with PSA levels but not for lesion size nor multifocality, limiting its clinical utility.