Introduction: Cytology and cystoscopy have limited sensitivity and specificity for the diagnosis of urothelial carcinoma (UC), due to inflammatory atypia, sampling errors, and other clinicopathologic factors that may obscure test results. Therefore, there remains a clinical need to identify biomarkers to improve the diagnostic accuracy to detect UC. Keratin 17 (K17) is an embryologic cytokeratin that functions as an oncoprotein to promote the degradation of tumor suppressors that drive tumorigenesis. Published work from our group established that K17 is highly expressed in UC and confirmed specificity for UC by immunohistochemical localization of K17 in tissue biopsies of both low-grade and high-grade UCs (Babu et al., Mod Pathol, 2018).

Objective: The objective of the current study was to develop an immunocytochemical (ICC) assay to determine K17 is a sensitive and specific biomarker to enhance diagnostic accuracy for UC in urine cytology.

Methods: 80 ThinPrep CytoLyt-fixed urine specimens, collected at Stony Brook Medicine in 2018, including 39 with a clinicopathologic diagnosis of UC based on current cytologic diagnosis, and/or a history of biopsy confirmed UC within one year (+/-) of urine cytology specimen collection. K17 ICC was performed by indirect immunoperoxidase methods and K17 test results were scored based on the detection of immunoreactive urothelial cells and not on an assessment of cytologic atypia. The sensitivity and specificity of urine K17 ICC for detection of urothelial neoplasia was calculated by comparison of ICC test results with the cytologic diagnosis and/or the histologic diagnosis of positive cystoscopic biopsy specimens (papillary urothelial neoplasm of low malignant potential [PUNLMP, n=1] or greater). Samples that had no history of abnormal urine cytologic or biopsy diagnosis were categorized as negative for urothelial neoplasia.

Results: Relative to the final clinicopathologic diagnosis of UC (n=38) or PUNLMP (n=1), K17 ICC had a sensitivity of 100% (95% CI: 91-100%) and specificity of 90% (95% CI: 77-96%). The positive predictive value was 90% (95% Cl: 78-96%) and the negative predictive value was 100% (95% Cl: 90-100%). Comparing ICC test results to cytologic and biopsy diagnoses, K17 ICC was positive in 9/23 (39%) cases with negative urine cytology. Of these 9 cases, 4 cases had biopsy confirmed UC. K17 ICC was positive in 4/11 (36%) of cases with inflammatory/reactive changes. Of these 4 cases, 3 had biopsy-confirmed UC. K17 ICC was positive in 14/30 (46%) of cases with mild atypia. Of these 14 cases, 12 had biopsy-confirmed UC and 1 had PUNLMP. K17 ICC was positive in 16/16 (100%) of cases with moderate atypia (n=8), severe atypia (n=3), or UC (n=5), all with biopsy-confirmed UC.

Conclusions: K17 ICC is a novel and highly sensitive and specific diagnostic test for underlying biopsy-confirmed UC among samples with inflammatory/reactive changes, cytologic atypia, or positive urine cytology. Thus, the K17 test could serve as an adjunct to guide the clinical management of UC cases.

Citation Format: Sruthi Babu, Lucia Roa-Peña, Ina Chan, Nam W. Kim, Sholeh Jahanfard, Luisa F. Escobar-Hoyos, Kenneth R. Shroyer. Validation of a novel cytologic biomarker for urothelial carcinoma [abstract]. In: Proceedings of the AACR Special Conference on Bladder Cancer: Transforming the Field; 2019 May 18-21; Denver, CO. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(15_Suppl):Abstract nr B01.