OBJECTIVE. The coracohumeral interval previously has beet) described as predictive of subcoracoid impingement on dynamic screening. The purpose of this study was to determine whether a coracohumeral interval acquired front routinely performed MRI can reliably diagnose subcoracoid impingement. MATERIALS AND METHODS. Preoperative MRI examinations of 19 patients (16 males, three females) with subsequent surgical confirmation ot'subcoracoid impingement were reviewed retrospectively and compared with MRI studies of 41 control subjects (22 males, 19 females). Axial and oblique sagittal coracohumeral interval measurements were taken. The morphology of the coracoid process and lesser tuberosity was assessed, Postanalysis application of the data to two smaller groups of patients was performed. The first group consisted of nine subjects (three males, six females) for whom subcoracoid impingement was diagnosed prospectively on the basis of abnormalities found by MRI. The second group consisted of seven patients (two males, five females) who were referred for MRI evaluation because of clinically suspected subcoracoid impingement. RESULTS. The average coracohumeral interval for females was 3 mm smaller than that for males. Using sex-adjusted data, we found a statistically significant difference between individuals with or without subcoracoid impingement in the axial coracohumeral interval (p = 0.01). This value, however, was poorly predictive (area under the receiver operating characteristic curve, 0.73). An 11.5-mm axial coracohumeral interval had 84% sensitivity but only 44% specificity. A 10.5-mm axial coracohumeral interval had 79% sensitivity and 59% specificity. The shoulder morphologic features assessed and intraarticular contrast use were not statistically significantly related to the coracohumeral interval, In postanalysis application of data. in the group of nine subjects without clinical diagnosis ot'subcoracoid impingement, all prospective MRI subcoracoid impingement diagnoses were falsely positive. However, if subcoracoid impingement was the referring diagnosis. prospective MRI evaluation more often was correct (n = 7 [three true-negatives, two true-positives, two false-negatives]), CONCLUSION. A sex-adjusted coracohumeral interval of 10.5-11.5 mm, although statistically significantly related to subcoracoid impingement, is poorly predictive of (his diagnosis when acquired via routinely performed MRI. Subcoracoid impingement is primarily a clinical diagnosis that may be supported, but not established, by this means.
