Novel radiographic grading system for radial facet of third carpal bone
Alvise Pasinato, Michael C. Shepherd, Jack S. Day, Vanessa G. Peter, Sarah E. Powell, Thomas O'Keeffe, Kristien L. Verheyen, Pieter H. L. RamzanAbstract
Background
Carpal lameness is a common cause of wastage in Thoroughbred racehorses, with the middle carpal joint most frequently affected. Evaluation of the dorsoproximal–dorsodistal oblique (DPr–DDiO) radiographic projection of the third carpal bone (C3) is fundamental in clinical and pre‐purchase assessment of Thoroughbred racehorses. A standardised grading system is needed to improve consistency of diagnosis and reporting between clinicians.
Objectives
To develop and assess inter‐rater reliability of a novel grading system for evaluating radiological features of the third carpal bone on the C3 skyline projection.
Study Design
Retrospective multi‐rater agreement study.
Methods
The novel grading system assessed bone radiopacity changes (Grades 0–4) and lucency characteristics (Grades A–C). Bone density ranged from Grade 0 (even trabecular distribution) to Grade 3 (severe radiopacity), with Grade 4 indicating overt evidence of fracture. Lucencies were classified as A (small, dorsal), B (linear or curvilinear, central), or C (extensive). Six observers independently graded 250 radiographs from a first‐opinion Thoroughbred racehorse caseload. Inter‐rater agreement was determined using Kappa statistics.
Results
Category‐specific bone density grading revealed moderate agreement for Grades 0, 1, and 3 ( κ = 0.45, 0.41, and 0.53), fair agreement for Grade 2 ( κ = 0.36), and moderate to substantial agreement for Grade 4 ( κ = 0.68). Inter‐rater agreement on lucency classification ranged from fair (Grade B, κ = 0.28; Grade A, κ = 0.39) to moderate (Grade C, κ = 0.52). Unweighted kappa for bone density grading was 0.44 and 0.47 for presence of lucency grading. Weighted kappa for bone density grading was 0.59, indicating moderate agreement.
Main Limitations
All observers were highly experienced clinicians and intra‐observer reliability was not assessed.
Conclusions
Inter‐rater agreement using a novel grading system for the evaluation of the C3 ‘skyline’ radiographic projection was generally moderate, and substantial for fracture. Adoption of this grading system in clinical practice will assist standardisation of diagnostic and prognostic evaluation of this important injury site.