Comparison of ultrasonography and radiography with arthroscopy for diagnosis of dorsoproximal osteochondral fragmentation of the proximal phalanx in 56 Thoroughbred racehorses

Background: Osteochondral fragmentation of the proximal phalanx (DPP1 fragment) is a common finding in the racehorse. Whilst radiographic assessment has been reported to lack diagnostic sensitivity, ultrasonographic examination has been documented to be extremely sensitive for identification of bone fragments in the metacarpophalangeal joint.

Objectives: To determine the sensitivity and specificity of ultrasonography as an imaging modality for diagnosis of DPP1 fragmentation and to compare it with radiography in the Thoroughbred racehorse.

Study design: Retrospective cohort study.

Methods: Medical records (radiographs and ultrasonographic examinations) of 56 Thoroughbred racehorses in training (average age 26.4 months [SD ±5.9]) with a history of forelimb fetlock joint effusion and arthroscopically confirmed DPP1 fragmentation were evaluated. Inter-observer agreement was calculated between three blinded reviewers for both ultrasonographic and radiographic findings. Median values were used for statistical analyses. Ultrasonographic and radiographic identification of DPP1 fragmentation was considered correct if in agreement with arthroscopic findings. The McNemar test was used to compare correct proportions between radiography and ultrasonography. Analyses were performed to evaluate correct identification of presence of fragment(s), location of fragment(s) and presence and location of fragment(s) within the joints, with values P < .05 considered significant.

Results: Eighty-one fetlock joints from 56 Thoroughbreds were evaluated. In total, 99 individual DPP1 fragments were confirmed arthroscopically. Ultrasonography was shown to have high levels of sensitivity (94%, 95% CI 89%-99%) and specificity (97%, 95% CI 92%-100%) for identification of DPP1 fragments, with transverse ultrasonographic images being more sensitive than sagittal ultrasonography (P = .001). Ultrasonography was also shown in this population to have significantly greater sensitivity than radiography (47%, 95% CI 37%-56%) (P < .001).

Conclusions: Ultrasonography is an extremely sensitive and specific imaging modality for evaluation of DPP1 fragmentation, with transverse ultrasonographic imaging being the most sensitive technique. Radiographic evaluation was less sensitive than ultrasonography in this study.