Comparison study of indirect MR arthrography and direct MR arthrography of the shoulder

Objective  To compare the diagnostic value of indirect magnetic resonance arthrography (I-MRA) with that of direct MR arthrography (D-MRA) for labral tears, rotator cuff tears, and long head of biceps tendon (LHBT) tears using a 3-T MR unit. Materials and methods  Institutional review board approval was given; written informed consent was obtained from all patients. From November 2005 to June 2006, 19 patients (eight men and 11 women; mean age, 51 years) who had undergone both I-MRA and D-MRA underwent arthroscopic surgery. Both methods were performed in fat-saturated axial, coronal oblique, and sagittal oblique T1-weighted sequences, as well as axial and coronal oblique T2-weighted sequences. Two radiologists independently and retrospectively evaluated two sets of MRA for the diagnosis of superior and anterior labral tears, subscapularis tendon (SSC), and supraspinatus–infraspinatus tendon (SSP–ISP) tears, and LHBT tears. With the arthroscopic finding as a gold standard, we analyzed statistical differences of sensitivities and specificities between two sets of MRA and inter-observer agreement was evaluated using the kappa value. Results  The sensitivity and specificity of I-MRA and D-MRA for reader 1 were 79/80% and 71/80%, respectively, for superior labral tears; 100/100% and 100/100%, respectively, for anterior labral tears; 64/75% and 64/100%, respectively, for SSC tears; 100/86% and 100/100%, respectively, for SSP–ISP tears; and 67/100% and 78/100%, respectively, for LHBT tears. Those of I-MRA and D-MRA for reader 2 were 86/80% and 71/100%, respectively, for superior labral tears; 100/83% and 100/100%, respectively, for anterior labral tears; 64/88% and 82/100%, respectively, for SSC tears; 92/86% and 100/100%, respectively, for SSP–ISP tears; and 78/90% and 89/100%, respectively, for LHBT tears. No significant differences were found between the methods. Inter-observer agreements were higher than moderate (κ > 0.41) with both methods. Conclusions  Based on a relatively small number of patients, no significant difference was detected between I-MRI and D-MRI with regard rotator cuff, labral, and LHBT tears.