High diagnostic accuracy of magnetic resonance imaging in detecting subscapularis tendon tears by focusing on the most superior portion of the subscapularis tendon.

PURPOSE: This study aimed to test the hypothesis that identifying the exact location of the most superior portion of the subscapularis tendon using magnetic resonance imaging (MRI) provides high diagnostic accuracy in detecting subscapularis tendon tears. METHODS: This study included 157 patients who underwent primary arthroscopic rotator cuff repair between 2014 and 2017. All patients underwent conventional 1.5-T MRI in our hospital, within 3 months before surgery. We retrospectively compared the diagnosis of subscapularis tendon tears using MRI based on an anatomical concept focusing on the superior-most insertion point of the subscapularis tendon with intraoperative arthroscopic findings. RESULTS: Subscapularis tendon tears were detected in 80 (51.0%) of the 157 patients during arthroscopic evaluation. The overall sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the MRI examination were 90, 83, 85, 89, and 87%, respectively. With a kappa score of 0.83, the concordance rate between the two raters was almost perfect (95% confidence interval, 0.75-0.92). The sensitivities of the oblique-sagittal and axial sequences were 84 and 79%, respectively. CONCLUSIONS: Preoperative MRI evaluation focusing on the most superior portion of the subscapularis tendon demonstrated high diagnostic accuracy in detecting subscapularis tendon tears. To find the most superior portion of the subscapularis tendon tears, it was essential to check the slice at the level of the lesser tubercle tip and its adjacent slice. In addition, the combined observation of oblique-sagittal and axial sequences helped to detect subscapularis tendon tears with higher sensitivity.