Institutional review board approval and informed consent were obtained. The purpose of the study was to prospectively perform magnetic resonance (MR) arthrography of the glenohumeral joint by using modified posterior approach without ultrasonographic or fluoroscopic guidance. A solution containing 0.1 mL of gadolinium chelate, 15 mL of saline, and 5 mL of 2% lidocaine was subsequently injected into the glenohumeral joint in 147 patients (81 men, 66 women; age range, 20-79 years). A 21-gauge needle was advanced along a trajectory connecting a skin mark 3-4 cm below and 2 cm medially to the posterolateral margin of the acromion and the coracoid process, as assessed with palpation, proceeding in posteroanterior direction. The joint was successfully entered at first attempt in 125 (85%) patients, at second attempt in 19 (13%), and at third attempt in three (2%). Contrast material-enhanced images were evaluated for presence, site, and maximal extent of contrast material extravasation; route of diffusion of the extravasation; compromised or noncompromised diagnostic quality; and presence of gas bubbles. Extravasation occurred in seven patients: at the interval between the teres minor muscle and infraspinatus muscle in five and within the infraspinatus muscle belly in two; extravasation had diffused along the teres minor muscle and infraspinatus muscle in five (71%) and along the teres minor muscle in two (29%). The mean extension of extravasation was 15 mm. Image quality was not compromised, and no gas bubbles were detected. The procedure was successful in all patients, with no complications.
(c) RSNA, 2007.