Abstract
BACKGROUND: Various magnetic resonance imaging (MRI) characteristics are frequently employed to aid diagnose adhesive capsulitis of the shoulder (ACS) and offer valuable therapeutic insights. To identify and summarize the diagnostic accuracy of these features, a systematic review and meta-analysis were performed. METHODS: Four databases, including PubMed, EMBASE, Web of Science, and Cochrane Library, were searched. Overlapping descriptions used to represent the same imaging in different studies are grouped into one MRI feature. Pooled diagnostic accuracy, including sensitivity and specificity, was calculated using a bivariate random-effects model. RESULTS: The screening identified 21 studies involving 928 ACS patients and 873 non-ACS patients considered eligible for inclusion in this meta-analysis. A total of 106 overlapping descriptions were classified into 7 features, including axillary capsular thickening, axillary capsular hyperintensity, axillary capsular enhancement, fat obliteration of the rotator interval (RI), RI enhancement, RI joint capsule thickening, and coracohumeral ligament (CHL) thickening. All seven features were considered informative for the diagnosis of ACS. Axillary capsular enhancement had the highest pooled sensitivity (95%, 95% CI [91%- 98%]), the highest diagnostic odds ratios (107, 95% CI [32, 357]), and the highest area under the curve(0.96 [0.94-0.97]). All features except fat obliteration of the RI and CHL thickening showed a pooled sensitivity of > 80%. Three of seven (axillary capsular thickening, axillary capsular hyperintensity, and axillary capsular enhancement) showed a pooled specificity of > 80%. CONCLUSION: Seven informative MRI features were identified in this study, with axillary capsular enhancement and RI joint capsule thickening showing the highest diagnostic accuracy. Clinicians can refer to these MRI features to increase confidence in diagnosing ACS and rule out other confused diagnoses.