Determination of magnetic resonance imaging criteria for diagnosis of adhesive capsulitis

The objective of this study was to identify optimal magnetic resonance imaging (MRI) parameters and their cutoff values for diagnosing adhesive capsulitis (AC). One hundred shoulder MRI images with stage 2 AC (AC group) and 100 MRI images without AC (control group) were randomly reviewed by two experienced shoulder specialists. They were asked whether MRI findings were compatible with AC and measurement of MRI parameters. Sensitivity, specificity, and accuracy were calculated. Correlation between MRI parameters and the range of motions was also analyzed. The mean capsular thickness in the axillary recess (AR) (5.9 mm in the AC group vs. 3.6 mm in the control group) on coronal oblique T2-weighted images and the rotator interval (RI) (7.2 mm vs. 4.8 mm, respectively) on oblique sagittal proton-density images were significantly greater in the AC group than in the control group, whereas the width of RI showed no significant difference between two groups. The highest diagnostic cutoff values were 4.5 mm for the AR and 6 mm for the RI, with sensitivity (91% and 88%, respectively), specificity (90% and 90%), and accuracy (90% and 89%). Capsular thickness in the AR and RI was significantly correlated with external rotation (P = 0.047) and internal rotation (P = 0.023). On conventional MRI, capsular thickness greater than 4.5 mm in the AR or 6 mm in the RI can be an optimal criterion for diagnosing AC. Capsular thickness in the AR and RI was correlated with the range of rotational motion.