Lesions of the rotator cuff footprint: diagnostic performance of MR arthrography compared with arthroscopy

Background To evaluate the diagnostic performance of magnetic resonance arthrography (MR-A) of the shoulder in the diagnosis of rotator cuff tears involving the humeral insertion of the supraspinatus and infraspinatus tendon (footprint), using arthroscopy as the reference standard. Materials and methods The study population included 90 consecutive patients with history and clinical diagnosis of instability of the shoulder, rotator cuff tear or posterosuperior glenoid impingement. A total of 108 MR arthrograms were performed, since 18 patients had undergone a bilateral procedure. Arthroscopy, which was performed within 45 days after MR-A, was used as the reference standard. Sensitivity, specificity, accuracy, positive and negative predictive values were then calculated. Results Magnetic resonance arthrography showed a sensitivity of 92 % and a specificity of 78 % for the overall detection of tears involving the rotator cuff footprint. The diagnostic accuracy was 90 %, and the positive and negative predictive values were 95 and 64 %, respectively. Ten lesions were non-classifiable on surgery, of which eight were non-classifiable on MR-A also. Conclusions Magnetic resonance arthrography is extremely accurate for the detection and classification of rotator cuff footprint tears. Most of these lesions are articular-sided (partial articular-sided supraspinatus tendon avulsion lesions) with predominance in younger patients and concealed type of tear (concealed interstitial delamination lesions).