To evaluate the feasibility of MR arthrography and ultrasonography in evaluating shoulder pain.

The subject group consisted of all patients who visited our institute complaining of shoulder pain or instability from June 2002 to December 2004. There were a total of 92 patients with a mean age of 48. On the basis of arthroscopic results, the sensitivity, specificity, and accuracy of ultrasonography and MR arthrography were evaluated by comparing them with each other.

In the diagnosis of supraspinatus tendon tears, ultrasonography had sensitivity and specificity of 100% and 64%, respectively, whereas MR arthrography had sensitivity and specificity of 80% and 94%, respectively. Ultrasonography also had high sensitivity and specificity in the diagnosis of subscapularis tendon tears (100% and 90%). MR arthrography was appropriate for identifying glenoid labral abnormalities (sensitivity, 95% and specificity, 61%). Similar results from ultrasonography and MR arthrography were obtained in the diagnosis of subscapular tendon tears or full-thickness tears of the rotator cuff tendons (kappa value, 0.644 and 0.911).

While evaluating rotator cuff abnormalities, ultrasonography was appropriate for screening, whereas MR arthrography was useful to confirm the results of the ultrasonography.