Interobserver agreement of standing magnetic resonance imaging for assessment of deep digital flexor tendon pathology

Reasons for performing study:

Standing magnetic resonance imaging (sMRI) is commonly used to evaluate the distal aspect of the deep digital flexor tendon (DDFT) of lame horses. There is paucity of information comparing interobserver agreement of this imaging modality.

Objectives:

To determine the interobserver agreement for assessment of the DDFT on sMRI using a newly developed grading system. It was hypothesised that different lesion types would be more frequently observed at different anatomical locations of the DDFT, and that interobserver agreement would be better for the presence of a lesion compared to lesion grading.

Study design:

A retrospective study assessing interobserver agreement. Methods: sMRI studies of 40 equine front feet were retrospectively graded by 2 observer teams using a newly developed grading system based on previous studies. Individual DDFT parameters and an overall DDFT grade were scored. A Pearson’s Chi-square test quantified the association between anatomical location and lesion type and interobserver agreement was quantified by calculating weighted-kappa statistic.

Results:

Observer team B found a statistically significant association between anatomical location and lesion type but observer team A did not. Interobserver agreement was fair to moderate for the majority of individual DDFT parameters. For most parameters, better interobserver agreement was found for presence of a lesion than for lesion grading. Interobserver agreement was fair for the overall DDFT grade.

Conclusions:

Interobserver agreement was acceptable for overall DDFT grade and was better for the presence of an overall grade. Interobserver agreement was also acceptable for most DDFT parameters and was better for assessing whether a lesion was present compared to the grading of lesions. This study highlights the value of a standardised grading system for assessing the DDFT on sMRI, and provides a method to improve the repeatibilty and reliability of future studies using sMRI.