RESUMO
OBJECTIVE: To assess diagnostic value and clinical utility of multidetector computed tomographic positive contrast arthrography (CTA) for meniscal lesions in dogs. STUDY DESIGN: Prospective case series. STUDY POPULATION: Client-owned dogs (n = 55) with cranial cruciate ligament injuries. METHODS: Sedated dogs underwent CTA using a 16-slice scanner, and subsequently received mini-medial arthrotomy for meniscal assessment. Scans were anonymized, randomized, and reviewed twice for meniscal lesions by three independent observers with varying experience. Results were compared with surgical findings. Reproducibility and repeatability were assessed with kappa statistics, intraobserver changes in diagnosis by McNemar's test, and interobserver differences using Cochran's Q test. Test performance was calculated using sensitivity, specificity, proportion correctly identified, and positive and negative predictive values and likelihood ratios. RESULTS: Analysis was based on 52 scans from 44 dogs. Sensitivity for identifying meniscal lesions was 0.62-1.00 and specificity was 0.70-0.96. Intraobserver agreement was 0.50-0.78, and interobserver agreement was 0.47-0.83. There was a significant change between readings one and two for the least experienced observers (p < .05). The sum of sensitivity and specificity exceeded 1.5 for both readings and all observers. CONCLUSION: Diagnostic performance was suitable for identifying meniscal lesions. An effect of experience and learning was seen in this study.