Comparison between exclusively long-axis and multiple-axis sonographic protocols for screening of rotator cuff...

The purpose of this study was to determine whether exclusively long-axis sonography differs from a multiple-axis scanning protocol as a screening tool for rotator cuff lesions in symptomatic shoulders when compared with magnetic resonance imaging (MRI).

A total of 509 consecutive patients (mean age, 52.8 years) referred for MRI were also routinely evaluated by sonography. We initially performed exclusively long-axis sonography and graded the rotator cuff as normal or abnormal. Patients subsequently underwent a full sonographic protocol using multiple-axis views. Magnetic resonance imaging findings were compared with sonographic findings for both techniques.

The overall accuracy of sonography was greater than 90%. We found divergent results from different sonographic techniques in 34 patients. Of these, 8 were multiple-axis false-negative; 14 were exclusively long-axis false-negative; 6 were exclusively long-axis false-positive; and 6 were multiple-axis false-positive. All cases with divergent false-negative findings on multiple-axis sonography showed tendinosis on MRI. Causes for false-negative findings on exclusively long-axis sonography included tendinosis and partial-thickness tears of the supraspinatus. No statistically significant difference was seen between both sonographic techniques compared with MRI in terms of accuracy, sensitivity, specificity, positive predictive value, and negative predictive value (P > .05).

Sonography is reliable for detecting rotator cuff abnormalities. Exclusively long-axis sonography seems appropriate as a screening tool for rotator cuff lesions in symptomatic shoulders.