Magnetic resonance imaging and magnetic resonance arthrography of the shoulder: dependence on the level of training of the performing radiologist for diagnostic accuracy.

摘要

PURPOSE: Discrepancies were identified between magnetic resonance (MR) imaging and clinical findings in patients who had MR imaging examinations evaluated by community-based general radiologists. The purpose of this study was to evaluate the diagnostic performance of MR imaging examinations of the shoulder with regard to the training level of the performing radiologist. METHODS: A review of patient charts identified 238 patients (male/female, 175/63; mean age, 40.4 years) in whom 250 arthroscopies were performed and who underwent MR imaging or direct MR arthrography in either a community-based or hospital-based institution prior to surgery. All MR imaging and surgical reports were reviewed and the diagnostic performance for the detection of labral, rotator cuff, biceps, and Hill-Sachs lesions was determined. Kappa and Student's t test analyses were performed in a subset of cases in which initial community-based MR images were re-evaluated by hospital-based musculoskeletal radiologists, to determine the interobserver agreement and any differences in image interpretation. RESULTS: The diagnostic performance of community-based general radiologists was lower than that of hospital-based sub-specialized musculoskeletal radiologists. A sub-analysis of re-evaluated cases showed that musculoskeletal radiologists performed better. kappa values were 0.208, 0.396, 0.376, and 0.788 for labral, rotator cuff, biceps, and Hill-Sachs lesions (t test statistics: p = <0.001, 0.004, 0.019, and 0.235). CONCLUSIONS: Our results indicate that the diagnostic performance of MR imaging and MR arthrography of the shoulder depends on the training level of the performing radiologist, with sub-specialized musculoskeletal radiologists having a better diagnostic performance than general radiologists.