Signal intensity on MRI after conservative treatment of patients with full-thickness rotator cuff tears: Correlation with shoulder pain

High signal intensity (HSI) on T2-weighted or fat-suppressed magnetic resonance imaging (MRI) around the shoulder is often observed in patients who have a rotator cuff tear, and is generally recognized as an indicator of shoulder pain, especially pain at night. We hypothesized these HSI volumes are reduced after improvement of symptoms. We sought to compare HSI before and after conservative treatment, and to determine if HSI is associated with the patient's pain status over time. We enrolled 55 patients (average age 65 years) into the study. All the patients complained of pain at night (University of California, Los Angeles (UCLA) pain score ≤2, Japanese Orthopedic Association (JOA) pain score ≤10) at their initial visits and underwent conservative treatment. MRI was obtained at their initial visit and after improvement of pain when their initial UCLA pain score was ≥6 points or JOA pain score was ≥20 points. The mean time between the MRI examinations was 11.8 months (SD 7.7) (range 2.5–39). HSI was assessed at the location of the subacromial–subdeltoid bursa, glenohumeral joint, subcoracoid bursa, and sheath of the long head of the biceps tendon, and graded. Factors correlating with the improvement of MRI findings were evaluated retrospectively. HSI had improved in 30 shoulders in all or some of the locations (I-group). There was no change in 25 shoulders for any location or aggravation in one or more locations (N-group). Multivariate logistic regression analysis found that the time from pain improvement until the second MRI was the only independent factor related to the improvement in HSI (p = .045). Improvement of shoulder symptoms is not directly associated with MRI findings. Improvement in HSI followed symptom improvement after a delay; the changes varied for each shoulder location.