Interobserver agreement during clinical magnetic resonance imaging of the equine foot.
Authors: Byrne Christian A, Voute Lance C, Marshall John F
Journal: Equine veterinary journal
Summary
Standardised interpretation of equine foot MRI remains challenging despite observer expertise, as Byrne and colleagues demonstrate through analysis of ten experienced radiologists' independent assessments of 15 clinical foot studies across three MRI systems. Using structured pathology grading (Grades 1–4) across seven key anatomical structures, the researchers found poor-to-fair agreement for most regions, with the notable exception of navicular bone pathology (κ = 0.52, moderate agreement), whilst relative agreement for grading severity ranged from 0.19 for the distal interphalangeal joint to 0.70 for navicular structures. Observers reported moderate-to-high confidence in their assessments overall, yet this confidence did not translate consistently into agreement, suggesting inherent variability in how MRI changes are interpreted rather than uncertainties expressed by the readers themselves. The findings carry immediate practical implications: subtle-to-moderate pathology grades are particularly prone to observer variation, meaning clinicians should seek specialist review of borderline MRI findings and exercise caution when making treatment decisions based on severity assessments alone, especially for non-navicular pathology. Standardised interpretation protocols and potentially additional training on low-field MRI systems—where most observers gained experience—may help reduce this variability and improve the reliability of MRI as a diagnostic and prognostic tool in equine lameness investigation.
Read the full abstract on PubMed
Practical Takeaways
- •MRI interpretation of equine foot pathology varies significantly even among experienced observers, so consider seeking a second opinion or specialist review for borderline cases affecting treatment decisions
- •Greatest diagnostic certainty exists at the extremes (clearly normal or clearly severe pathology); mild-to-moderate changes have lower reproducibility and may warrant cautious clinical correlation
- •Navicular bone pathology shows the most reliable interpretation (KCC = 0.70), while distal interphalangeal joint assessment is least reliable (KCC = 0.19) — adjust your confidence in these respective diagnoses accordingly
Key Findings
- •Interobserver agreement for presence/absence of foot pathology ranged from poor-to-fair across most structures, with only navicular bone showing moderate agreement (k = 0.52)
- •Relative agreement for pathology severity grading varied substantially by anatomical structure, from KCC = 0.19 for distal interphalangeal joint to KCC = 0.70 for navicular bone
- •Agreement was generally greatest at the extremes of pathology (no pathology or severe pathology) with more disagreement in mild-to-moderate ranges
- •Observer confidence in pathology assessment was generally moderate to high despite notable variation in perceived severity of pathology between experienced observers