Muscle Edema of Retraction and Pseudo–Fatty Infiltration After Traumatic Rotator Cuff Tears: An Experimental Model in Sheep

BACKGROUND: Traumatic rotator cuff tears can result in retraction of the tendon and may be associated with muscle edema, which may be confused with fatty infiltration as seen on magnetic resonance imaging (MRI). PURPOSE: To describe the characteristics of a type of edema associated with acute retraction of the rotator cuff tendon (termed “edema of retraction”) and to highlight the risk of mistaking it with pseudo–fatty infiltration of the rotator cuff muscle. STUDY DESIGN: Descriptive laboratory study. METHODS: A total of 12 alpine sheep were used for analysis. On the right shoulder, osteotomy of the greater tuberosity was performed to release the infraspinatus tendon; the contralateral limb acted as the control. MRI was performed immediately after surgery (time zero) and at 2 and 4 weeks postoperatively. T1-weighted, T2-weighted, and Dixon pure-fat sequences were reviewed for hyperintense signals. RESULTS: Edema of retraction resulted in hyperintense signals around or within the retracted rotator cuff muscle on both T1- and T2-weighted imaging, but there was an absence of hyperintense signals on Dixon pure-fat imaging. This represented pseudo–fatty infiltration. Edema of retraction created a characteristic “ground glass” appearance of the muscle on T1-weighted sequences and was often found in either the perimuscular or intramuscular location of the rotator cuff muscle. Compared to time zero values, a decrease in the percentage of fatty infiltration was observed at 4 weeks postoperatively (16.5% ± 4.0% vs 13.8% ± 2.9%, respectively; P < .005). CONCLUSION: The location of edema of retraction was often peri- or intramuscular. Edema of retraction presented as a characteristic “ground glass” appearance of the muscle on T1-weighted sequences and led to a decrease in the fat percentage because of a dilution effect. CLINICAL RELEVANCE: Physicians should be aware that this edema can result in a form of pseudo–fatty infiltration, as it is associated with hyperintense signals on both T1- and T2-weighted sequences, and it can be mistaken for fatty infiltration.