RESUMO
PURPOSE: To assess the accuracy and reliability of routine preoperative magnetic resonance imaging (MRI) in the detection of the comma sign compared with the gold standard of arthroscopic findings. METHODS AND MATERIALS: Preoperative MRI exams in consecutive patients undergoing arthroscopic subscapularis tendon repair, over a 5-year time frame, were retrospectively reviewed for full-thickness tears of the subscapularis and supraspinatus tendons, fatty atrophy of the subscapularis and supraspinatus muscles, and status of the long head of the biceps tendon. Each case was also evaluated for presence or absence of a comma sign on MRI. Surgical findings served as the diagnostic standard of reference in determination of a comma sign. RESULTS: The study cohort included 45 male and 10 female patients (mean age, 56; range, 32-80 years). A comma sign was present at arthroscopy in 19 patients (34.5%). Interclass and intrarater correlation showed 100% agreement in preoperative assessment of a comma sign on MRI. MRI showed an overall accuracy of 83.6% in diagnosis of a comma sign (sensitivity, 63.2%; specificity, 94.4%; positive predictive value, 85.7%; negative predictive value, 82.9%; positive likelihood ratio, 11.37; negative likelihood ratio, 0.39). No statistically significant association was observed between an arthroscopic comma sign and patient demographics or MRI findings of full-thickness rotator cuff tears, muscle fatty atrophy, or long head of the biceps tendon pathology. CONCLUSIONS: MR imaging illustrates excellent reliability and good specificity and accuracy in detection of the arthroscopic comma sign in the setting of subscapularis tendon tearing. Detection of a comma sign on MRI may be important preoperative planning information in the arthroscopic management of patients with subscapularis tendon tears. LEVEL OF EVIDENCE: Level IV, retrospective diagnostic study.
