Frozen shoulder: MR arthrographic findings

To evaluate the magnetic resonance (MR) arthrographic findings in patients with frozen shoulder.

Preoperative MR arthrograms of 22 patients (six women, 16 men; mean age, 54.7 years) with frozen shoulder treated with arthroscopic capsulotomy were compared with arthrograms of 22 age- and sex-matched control subjects without frozen shoulder. The thickness of the coracohumeral ligament (CHL) and the joint capsule, as well as the volume of the axillary recess, were measured (Mann-Whitney test). Abnormalities in the CHL, subcoracoid fat, superior glenohumeral ligament, superior border of the subscapularis tendon, long biceps tendon, and subscapularis recess were analyzed in consensus by two blinded radiologists (chi(2) test).

Patients with frozen shoulder had a significantly thickened CHL (4.1 mm vs 2.7 mm in controls) and a thickened joint capsule in the rotator cuff interval (7.1 mm vs 4.5 mm; P < .001 for both comparisons, Mann-Whitney test) but not in the axillary recess. The volume of the axillary recess was significantly smaller in patients with frozen shoulder than in control subjects (P = .03, Mann-Whitney test). Thickening of the CHL to 4 mm or more had a specificity of 95% and a sensitivity of 59% for diagnosis of frozen shoulder. Thickening of the capsule in the rotator cuff interval to 7 mm or more had a specificity of 86% and a sensitivity of 64%. Synovitis-like abnormalities at the superior border of the subscapularis tendon were significantly more common in patients with frozen shoulder than in control subjects (P = .014, chi(2) test). Complete obliteration of the fat triangle between the CHL and the coracoid process (subcoracoid triangle sign) was specific (100%) but not sensitive (32%).

Thickening of the CHL and the joint capsule in the rotator cuff interval, as well as the subcoracoid triangle sign, are characteristic MR arthrographic findings in frozen shoulder.