Validation of the Collaborative Ocular Tuberculosis Study (COTS) Calculator for the Diagnosis of Ocular Tuberculosis in Thai Patients

PURPOSE: To evaluate the diagnostic accuracy of the Collaborative Ocular Tuberculosis Study (COTS) calculator in predicting ocular tuberculosis (OTB) and compare its performance with ophthalmologists' clinical judgment among Thai patients. METHODS: This retrospective study included patients with suspected OTB. Demographic, clinical, laboratory, and imaging data were scored using the COTS calculator (scores 1-5). Clinician judgment combined with treatment success served as the reference standard. Diagnostic accuracy was assessed at two thresholds: COTS ≥ 4 and COTS = 5. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and area under the ROC curve (AUC) were calculated. COTS = 5 showed the highest specificity (76.5%;95%CI,52.7-90.4) but low sensitivity (35.6%;95%CI,23.2-50.2). Clinician judgment had the highest sensitivity (100.0%;95%CI,92.1-100.0) with low specificity(29.4%;95%CI,13.3-53.1%). COTS ≥ 4 provided the best balance (sensitivity 68.9%;95%CI,54.3-80.5, specificity 70.6%,95%CI,46.9-86.7), with PPV/NPV of 86.1%/46.2% (vs. 80.0%/31.0% for COTS = 5). Clinician judgment showed the highest screening performance (PPV 78.9%,NPV 100.0%). The overall accuracy for all locations, using COTS ≥ 4, was 69.35%, while the specific accuracies for choroiditis, panuveitis, and retinal vasculitis were 80.00%, 73.08%, and 63.16% respectively. While ophthalmologists' clinical judgment provided superior sensitivity and overall accuracy. These findings support the COTS calculator as a useful adjunct, showing high accuracy specifically for choroiditis and panuveitis in diagnosing OTB.