Novel UF-5000-based Detection, Integrating Inflammatory Parameters, for Urothelial Carcinoma of the Urinary Bladder.

Background/aim: Urothelial carcinoma (UC) is the most common epithelial bladder malignancy. Although urine cytology is widely used for screening, its sensitivity in detecting low-grade UC is limited. This study evaluated the diagnostic accuracy of the research-use parameter "Atyp.C" from the fully automated urine particle analyzer UF-5000, in combination with the neutrophil-to-lymphocyte ratio (NLR), for UC detection.

Patients and methods: Urine samples from 57 noninvasive UC, 41 invasive UC, and 61 non-UC cases (n=159) were examined at Kurume University Hospital between 2020 and 2023. Specimens with atypical cells were excluded from the study. Receiver operating characteristic curve analysis was conducted using Atyp.C data from the UF-5000 to determine the optimal cutoff value. An UC detection algorithm incorporating the NLR was examined using the AI platform DataRobot, and its diagnostic accuracy was assessed.

Results: The diagnostic accuracy for invasive UC was assessed using an Atyp.C cut-off of 0.1/μl, yielding an area under the curve (AUC) of 0.824, sensitivity 70.7%, and specificity 90.3%. Noninvasive UC showed lower accuracy (AUC=0.565; sensitivity, 22.8%; specificity, 90.3%). Incorporating NLR improved invasive UC detection (AUC=0.892; sensitivity, 75.0%; specificity, 100%). NLR was the most influential factor in UC detection.

Conclusion: The UF-5000, when combined with the NLR, may enhance UC screening and contribute to a more effective diagnostic strategy.