Arthroscopic and magnetic resonance arthrogram features of Kim's lesion in posterior shoulder instability.

PURPOSE The purpose of this study was to describe the anatomic features of the glenoid and labrum for shoulders with and without Kim's lesions, as well as define the sensitivity, specificity, and reliability of magnetic resonance arthrogram for the detection of these lesions and determine the interobserver reliability of Kim's classification for posterior labral tears. METHODS A retrospective, blinded analysis of 41 shoulders was performed. Twenty-one shoulders composed the Kim's lesion group. Three observers measured glenoid and chondrolabral retroversion and anterior and posterior labral height. The shoulders were classified as 1 of 3 types according to Kim's classification. RESULTS There were no differences in glenoid version, chondrolabral version, glenoid depth, and labral height between the groups. The sensitivity, specificity, positive predictive value, and negative predictive value for diagnosing Kim's lesions were 85.7%, 75%, 78.3%, and 83.3%, respectively. The κ values for interobserver reliability for detecting and classifying Kim's lesions were 0.739 (substantial) and 0.329 (fair), respectively. CONCLUSIONS Patients with incomplete posteroinferior labral tears have similar retroversion (bony and soft tissue) and labral height to patients without posterior shoulder instability. The sensitivity, specificity, and reliability of magnetic resonance arthrogram for detecting these lesions are substantial. The interobserver reliability of Kim's classification is fair. LEVEL OF EVIDENCE Level IV, retrospective case-control study.