Leveraging MRI characterization of longitudinal tears of the deep digital flexor tendon in horses using machine learning

Abstract

While MRI is the modality of choice for the diagnosis of longitudinal tears (LTs) of the deep digital flexor tendon (DDFT) of horses, differentiating between various grades of tears based on imaging characteristics is challenging due to overlapping imaging features. In this retrospective, exploratory, diagnostic accuracy study, a machine learning (ML) scheme was applied to link quantitative features and qualitative descriptors to leverage MRI characteristics of different grades of tearing of the DDFT of horses. A qualitative MRI characteristic scheme, combining tendon morphologic features, altered signal intensity, and synovial sheath distention, was used for LT classification with an excellent diagnostic accuracy of the high‐grade tears but more limited accuracy for the detection of low‐grade tears. A quantitative ML approach was followed to measure the contribution of 30 quantitative phenotypic features for characterizing and classifying tendinous tears. Among the 30 imaging features, boundary curvature represented by the standard deviation and maximum had the most significant discriminatory power (P < 0.05) between normal and abnormal tendons and could be used as an aid for classifying the different grades of LTs of DDFTs. Imaging analysis‐based 3D interactive surface plot supports qualitative characterization of different grades of LTs of the DDFT through clearer visualization of the tendon in three dimensions and simple integration of two perspectives features (i.e., margin/distribution and intensity/distribution). A systematic approach combining quantitative features with qualitative analyses using ML was diagnostically beneficial in MRI characterization and in discriminating between different grades of LTs of the DDFT of horses.

Artificial intelligence in veterinary diagnostic imaging: A literature review

“…A similar investigation into hip dysplasia was performed by Gomes et al 34 El-Khamary et al 36 were the first to apply ML to the equine distal dataset. 37 Larger datasets are essential to advancing the use of AI in veterinary imaging, and training on human images when appropriate disease correlates are available may be one method to accomplish this.…”

“…A positive correlation between the maximum and standard deviation of the boundary and the grade of the longitudinal tear was noted. The authors included surface graphs illustrating the recognition of the tendon borders, and demonstrating that irregularity of the tendon/synovium interface was the anatomic correlation corresponding to the boundary features identified by the algorithm 36 . This study demonstrated the potential for ML algorithms and quantitative analysis to aid in the interpretation of musculoskeletal injury in MRI.…”

Presence and size of synovial masses within the navicular bursa correlate well between magnetic resonance imaging and bursoscopy and have a guarded prognosis

“…A tendon lesion was recognised as a region of focal, well-defined hyperintensity on T1w GRE sequences protruding dorsally into the proximal recess of the navicular bursa, often associated with concurrent T2w hypointense or intermediate signal, causing alteration of the dorsal contour of the tendon, as described previously. 2,6,7,15 A grading system for evaluation of DDFT lesions was developed (Table 1, Figures 1 and 2). This was based on T1w 3D (HR) transverse sequences, for their ability to detect smaller lesions not always visible on T2w sequence.…”

“…14 In a different study comparing high-field MRI and tenoscopic examination of the digital flexor tendon sheath, MRI had the highest sensitivity (100%) for the detection of large longitudinal tears of the DDFT; however, the sensitivity for the diagnosis of small tears was only average (66.6%). 15 Similarly, when assessing the presence of adhesions in the navicular bursa on high-field MRI and navicular bursoscopy, the positive predictive value of MRI was 100% for type 3 (large) adhesions, but only 50% for type 1 (small) adhesions. 16 To date, no study has assessed the significance of DDFT lesions and synovial masses identified on MRI within the navicular bursa, bursoscopic findings and long-term athletic performance.…”