The diagnostic value of magnetic resonance imaging for different types of subscapularis lesions

Purpose To investigate the accuracy of preoperative 1.5 T magnetic resonance imaging (MRI) interpreted with standardized procedure for diagnosing subscapularis (SSC) tears identified by arthroscopy. The diagnostic accuracy of MRI for different types of SSC lesions was also evaluated. Methods Two-hundred and seventy-two consecutive patients with rotator cuff tears identified by arthroscopy were included into this study. All patients had preoperative MRI scans with 1.5 T strength. Totally, seven different signs of SSC tears were evaluated on MRI scans. The diagnostic value of MRI was evaluated, respectively, according to two classifications of SSC lesions. The definitive diagnosis for SSC lesions was based on the arthroscopic findings. Results Among the 272 patients in this study, 107 (39 %) had SSC tears confirmed by arthroscopy. The surgeons correctly diagnosed 88 of 107 patients with SSC tears. The overall sensitivity was 82.2 %. The sensitivity of MRI for types I, II, III, IV and V was 70, 82.4, 96, 100 and 100 %, respectively, based on the classification by Lafosse. The sensitivity of MRI for partial-thickness and partial-width, full-thickness and partial-width, full-thickness and full-width tear was 75.3, 96.2, 100 %, respectively, according to the classification by Kim. Conclusions The sensitivity of MRI in predicting SSC tears preoperatively was improved with the current procedure. Understanding of certain MRI characteristics of SSC tears could increase the accuracy for diagnosing SSC lesions. 1.5 T MRI of the SSC was not reliable for predicting partial thickness and especially less than 1/3 width tears. Level of evidence III.