Comparison of radiography and CT for the evaluation of third carpal bone slab fractures in Thoroughbred racehorses.

Slab fractures of the third carpal bone (C3) are a common cause of lameness in Thoroughbred racehorses. Information on fracture morphology is commonly obtained from radiographs or CT. This retrospective, methods comparison aimed to explore the agreement between radiography and CT for imaging C3 slab fractures and discuss the contribution of the latter to clinical case management. Thoroughbred racehorses with a slab or incomplete slab fracture of C3 identified on radiographs that subsequently underwent CT examination were included. Fracture characteristics (location, plane, classification, displacement, comminution) and fracture length as a percentage of the proximodistal length of the bone, termed the proximodistal fracture percentage (PFP) were recorded independently from both modalities and then compared. Across all fractures (n = 82) radiographs and CT showed slight agreement on the presence of comminution (Cohen's Kappa (κ) 0.108, P 0.031) and moderate agreement on fracture displacement (K 0.683, P < 0.001). Computed tomography identified comminution in 49 (59.8%) and displacement in nine (11.0%) fractures that were not detected by radiographs. Half of the fractures were only seen on flexed dorsoproximal-dorsodistal oblique (DPr-DDiO) radiographs and therefore were of unknown length without additional CT imaging. Incomplete fractures that could be measured on radiographs (n = 12) had a median (IQR) PFP of 40% (30%-52%) on radiographs and 53% (38%-59%) on CT, a statistically significant difference (P = 0.026). Radiography and CT showed the poorest agreement when determining the presence of comminution. Additionally, radiography often underestimated the incidence of displacement, and fracture length, and resulted in more fractures being classified as incomplete when compared to CT.