Objectives: To assess the diagnostic accuracy of radiographs in identifying humeral intracondylar fissures in dogs.
Materials and methods: Clinical data from dogs undergoing CT and radiographic elbow investigations at a single institution were reviewed. Based on CT diagnosis, radiographs were classified as having fissures and not having fissures. The radiographic images were evaluated by three blinded observers for the presence/absence of a fissure, type (partial or complete), and for secondary signs that could suggest the presence of a fissure in case this was not visible.
Results: A total of 81 radiographs (76 elbows) were reviewed. Based on CT findings, 37 elbows had fissures and 44 did not. Radiographic interpretation demonstrated a sensitivity of 47.7% (confidence interval 95%: 0.40 to 0.59) and specificity of 85.6% (confidence interval 95%: 0.72 to 0.93) in correctly identifying humeral intracondylar fissures. The positive predictive value was 75.5% (confidence interval 95%: 0.64 to 0.84) and the negative predictive value was 66.1% (confidence interval 95%: 0.63 to 0.68). Agreement with CT images was <0.4, and interobserver and intraobserver agreements were <0.8. The radiographic orientation and type of fissure did not significantly affect interpretation outcomes.
Clinical significance: Radiographic imaging alone is insufficient for reliably diagnosing humeral intracondylar fissures in dogs.
