|
|
Background: The aim of this study was to develop and validate a risk stratification model for the screening of patients with suspected urothelial carcinoma (UC).
Methods: We enrolled 671 consecutive patients with suspected UC and generated a risk stratification model based on urinary parameters by using an automated urinalysis analyzer (Sysmex UN-9000). All patients received urine cytology examination from January 1, 2019, to October 31, 2022.
Results: Out of the 671 patients, 191 (28.5%) were ultimately diagnosed with UC. The four features associated with the presence of malignancy on multivariable analysis can be summarized by using the mnemonic UC-PAAS: UC, protein vs. creatinine ratio (P/C), age, atypical cells (Atyp.C), and small round epithelial cell (SRC). Major criteria include Atyp.C ≥ 0.1/μL (2 points) and age ≥ 65 years (2 points); minor criteria include SRC ≥ 2.7/μL (1 point) and abnormal P/C results (1 point). The model evidenced good discrimination (area under the curve = 0.802, 95% confidence interval [0.756, 0.848]) in the training group. A UC-PAAS cutoff of more than 4 points identified a high-risk population, of whom 37 of 59 (62.7%) had UC; the negative predictive value was 0.867. The validation group yielded similar findings.
Conclusions: We present a urinalysis-based screening model, the UC-PAAS, that may serve as an accessory clinical tool for the evaluation of patients with suspected UC, because the model identifies patients at higher risk who require closer follow-up than others or additional examinations.
DOI: 10.7754/Clin.Lab.2024.240330
|
