Observer variation in the evaluation and classification of severe central tarsal bone fractures in racing Greyhounds

Abstract

To determine observer agreement on radiographic evaluation of central tarsal bone (CTB) fractures and compare this with evaluation of the same fractures using computed tomography (CT). Radiographs and CT scans were obtained of the right tarsi from limbs of Greyhounds euthanatized after sustaining severe CTB fracture during racing. Four observers described and classified each fracture. Inter- and intra-observer agreements were calculated. Inter-observer agreement was higher for assessment of fractures using CT. Several fractures assessed by radiography were misclassified as a less severe type. Intra-observer agreement for assessment and classification of CTB fractures via radiography versus CT was variable. Overall agreement among all four observers was higher for CT than radiography. Additionally, when identifying fractures of the adjacent tarsal bones, observer agreement was higher for CT than radiography. Computed tomography improved observer ability to correctly evaluate CTB fracture and detect the degree of displacement and extent of any comminution. Identification of fractures of adjacent tarsal bones was also improved when tarsi were assessed using CT. These data suggest that treatment decisions based solely on radiographic assessment of CTB fractures may not produce the expected outcome.