Magnetic resonance arthrography of the shoulder: accuracy of gadolinium versus saline for rotator cuff and labral pathology

The purpose of this study was to evaluate the necessity of intra-articular gadolinium versus saline alone in magnetic resonance arthrography (MRA) of the shoulder. Our database was reviewed for 100 consecutive shoulder MRA examinations performed between January 2007 and December 2007. Patient information was blinded and images were retrospectively reviewed by at least two radiologists with dedicated musculoskeletal training. T2-weighted (T2W) images were initially analyzed in isolation to simulate MRA with saline alone. After a delay, the full study was analyzed including T1-weighted (T1W) and T2W images. If there was a significant discordance between the two analyses with regard to rotator cuff or labral pathology, the study was again reviewed by all evaluators in consensus to determine if the T1W images offered additional diagnostic information and increased diagnostic confidence. Of the 100 MRA examinations, there were 15 discordant cases. Two cases were discordant with regard to rotator cuff pathology and 13 were discordant on the basis of labral pathology. When the discordant cases were reviewed in consensus, the T2W images appeared to display rotator cuff and labral pathology as definitively as the T1W images. Interobserver and intraobserver variability was favored to have played a role in causing the discordances. MRA of the shoulder performed with joint distention provided by saline alone appears to offer equivalent diagnostic information to MRA performed with gadolinium enhancement. This protocol modification improves efficiency by eliminating several image series and provides a small cost savings by eliminating gadolinium. (orig.)