Abstract
Background/Objectives: Urinary biomarkers may assist hematuria risk stratification, but diagnostic performance must be assessed against an appropriate reference standard. We evaluated urinary keratin 17 immunocytochemistry (URO17®) during hematuria evaluation and compared it with urine cytology. Methods: This retrospective, single-center study included adults evaluated between August 2019 and March 2026. The analytic cohort comprised 746 specimens from 612 patients. Formal diagnostic accuracy analyses used one result-independently selected specimen per patient in the reference-standard cohort. URO17® positivity was defined as at least 10 keratin 17-positive cells. Sensitivity and specificity were compared using paired McNemar tests; atypical cytology was classified as negative in the primary comparison. Results: The reference-standard analysis included 313 patients, including 20 with histopathologically confirmed urothelial carcinoma. URO17® sensitivity was 90.0% (95% confidence interval, 69.9–97.2), specificity 90.1% (86.1–93.0), and negative predictive value 99.2% (97.3–99.8). Cytology sensitivity was 20.0% and specificity 98.6%. URO17® was 70.0 percentage points more sensitive and 8.5 percentage points less specific (both p < 0.001). Quantitative cell counts yielded an area under the receiver operating characteristic curve of 0.938. Patients without reference-standard verification were analyzed descriptively. Conclusions: URO17® demonstrated higher sensitivity but lower specificity than conventional cytology when atypical cytology was classified as negative. The high negative predictive value reflects, in part, the 6.4% cancer prevalence in the verified cohort. Selective verification and the small number of cancer-positive patients limit interpretation. Prospective multicenter validation is needed to define its adjunctive clinical role.