Abstract
Objective: To evaluate the diagnostic performance of the MUS-3600 automated urine analyzer for identifying culture-positive urine samples and to explore its potential application in laboratory-based urine culture triage.Methods: In this retrospective study, paired urinalysis and urine culture results from 2539 adults were evaluated. Culture positivity was defined as bacterial growth ≥105 colony-forming units per milliliter. Contaminated specimens were excluded from diagnostic accuracy analyses. Receiver operating characteristic analysis was performed for white blood cell (WBC) and bacterial counts. A multivariable logistic regression model incorporating WBC, bacterial count, and nitrite was developed and internally evaluated using 10-fold cross-validation. Sensitivity-prioritized thresholds and model operating points were additionally examined.Results: After excluding 160 contaminated specimens, 2379 samples were analyzed, including 402 culture-positive samples (16.9%). The area under the curves (AUCs) were 0.844 (95% CI = 0.823-0.865) for WBC and 0.802 (95% CI = 0.776-0.827) for bacterial count, with Youden-derived thresholds of 16.8/µL and 78.4/µL, respectively. Nitrite showed 26.9% sensitivity and 99.4% specificity. The multivariable model achieved an AUC of 0.880 (95% CI = 0.858-0.901) and a cross-validated AUC of 0.878. A WBC threshold of 5.6/µL provided 95.3% sensitivity and 41.3% specificity. At 95.0% model sensitivity, specificity was 42.7%, with 20/402 (5%) culture-positive samples classified as screen negative.Conclusion: Sensitivity-prioritized analyses demonstrated a trade-off between false-negative results and specificity. The evaluated approaches did not simultaneously meet the European Federation of Clinical Chemistry and Laboratory Medicine performance specifications for culture rule-out, and the sensitivity-prioritized thresholds and model operating points should therefore be considered exploratory. Prospective studies incorporating clinical information and external validation are required before their application to routine laboratory-based urine culture triage.