Angled oblique sagittal MR imaging of rotator cuff tears: comparison with standard oblique sagittal images

Objective. To compare the accuracy for diagnosing rotator cuff tears of oblique coronal images supplemented with standard oblique sagittal images versus thinner-section angled oblique sagittal images. Design and patients. The study included 75 consecutive patients who had a shoulder MR scan followed by arthroscopy. MR images included oblique coronal, oblique sagittal (4 mm thick, 1 mm interslice gap), and angled oblique sagittal (3 mm/0.2 mm) images perpendicular to the lateral cuff. A musculoskeletal staff radiologist and fellow separately evaluated the cuff for tears on the oblique coronal images supplemented with either the oblique sagittal or the angled sagittal images. Results. For distinguishing a cuff tear from no tear, the staff radiologist had an accuracy of 0.76 (95% confidence interval: 0.67, 0.85) with the standard sagittal set, and 0.88 (0.80, 0.95) with the angled set (P=0.04). There was a nonsignificant improvement in accuracy for the fellow, calculated as 0.73 (0.63, 0.83) on the standard sagittal set and 0.76 (0.67, 0.85) on the angled set. Both readers also improved their diagnostic accuracy for partial-thickness tears with the angled set, although the improvement was statistically significant only for the staff radiologist. Conclusion. There is a slight improvement in accuracy for diagnosing rotator cuff tears, particularly partial-thickness tears, for the more experienced radiologist using thinner-section angled oblique sagittal images. These images may be useful as a supplemental sequence in patients where it is important to identify partial-thickness tears accurately. Received: 7 August 2000 Revision requested: 13 October 2000 Revision received: 27 November 2000 Accepted: 12 December 2000