Purpose: To evaluate the predictive value of preoperative magnetic resonance imaging (MRI) with respect to rotator cuff ruptures.
Material and Methods: Thirty‐one patients with rheumatic disease underwent preoperative MRI before shoulder arthroplasty. The scans were reviewed independently by two experienced radiologists. Three surgeons performed all the replacements (hemiarthroplasties), and the condition of the rotator cuff was assessed. Complete and massive tears of the rotator cuff were recorded and compared at surgery and on MRI.
Results: With MRI, 21 shoulders (68%) were classified as having complete or massive tears of the rotator cuff and at surgery 14 shoulders (45%). Cohen's kappa coefficient was 0.44 (95% CI: 0.16 to 0.72) and accuracy 0.71 (95% CI: 0.52 to 0.86).
Conclusion: In severely destroyed rheumatoid shoulder, the findings of soft tissues were incoherent both with MRI and at surgery. The integrity of tendons could not readily be elucidated with MRI because of an inflammatory process and scarred tissues; in surgery, too, changes were frequently difficult to categorize. Preoperative MRI of severely destroyed rheumatoid shoulder before arthroplasty turned out to be of only minor importance.
