How reliable are imaging protocols in the diagnosis of subscapularis tears?

The purpose of the study is to evaluate the accuracy of detecting subscapularis tendon tears on different imaging modalities in comparison with surgical findings. In addition, the accuracy of long head of biceps tendon pathology in assisting the diagnosis of a subscapularis tear was evaluated. Retrospectively, 336 patients who underwent surgery in the UZ Brussel for rotator cuff pathology and had pre-operative imaging at the hospital were included. Pathology of the subscapularis tendon and the long head of biceps tendon on imaging modalities was compared to arthroscopic and/or open surgery findings. 111 of the 336 patients (33.0%) had a subscapularis tear diagnosed during surgery. None of the imaging modalities reaches the cut-off weighted kappa value (k) for substantial agreement of 0.61. Magnetic re- sonance imaging and magnetic resonance arthro- graphy have the highest k of 0.288, indicating minimal agreement with arthroscopy. Computed tomography arthrography (k = 0.167) and ultrasound (k = 0.173) shows both no agreement. Biceps instability was significantly correlated with a subscapularis tear, but the negative predictive value was always higher than the positive predictive value on ultrasound, magnetic resonance arthrography and computed tomography arthrography. The negative predictive value for detection of full thickness tears is as high as 96.2% on magnetic resonance arthrography. Accurate imaging diagnosis in daily practice of subscapularis tendon tears remains a challenge with the best results for magnetic resonance arthrography. The value of biceps instability lies in its negative predictive value rather than its positive predictive value.