Comparison of high-field-strength versus low-field-strength MRI of the shoulder.

OBJECTIVE Previous studies have reported similar results of shoulder MRI versus arthroscopy for high-field-strength (1.5-T) and low-field-strength (0.2-T) units. We report our experience with the accuracy of high- versus low-field-strength units versus arthroscopy for detection of supraspinatus tendon tears and labral tears in the same patients. SUBJECTS AND METHODS. Three musculoskeletal radiologists prospectively interpreted shoulder MRIs from 40 patients who had a complete shoulder MRI examination on a 0.2-T system and limited imaging on a 1.5-T unit. Proton-density axial and fat-saturated T2-weighted coronal and sagittal sequences were performed. Each radiologist interpreted the open unit images first and the high-field-strength images second. Results from 28 patients who also underwent arthroscopy were also compared with the MRI interpretations. All scans were then retrospectively reinterpreted by consensus of the three reviewers, who were unaware of the patient's name, results of previous MRI, or arthroscopy report (if any). RESULTS High-field-strength images altered reviewers' interpretations of low-field-strength scans for nine of 40 patients. In four patients, full-thickness supraspinatus tendon tears could be diagnosed definitively on the high-field-strength unit but not on the open unit. Three labral tears and two superior labral anteroposterior lesions could be depicted definitively on the high-field-strength unit but not on the open unit. All tears were confirmed at arthroscopy. CONCLUSION High-field-strength MRI units provide better spatial and contrast resolution and allow more accurate interpretations than low-field-strength units; these findings may affect clinical treatment.