Magnetic resonance rotator cuff fat fraction and its relationship with tendon tear severity and subject characteristics.

Abstract

Compared with the Goutallier classification, chemical shift-based magnetic resonance (MR) fat quantification has superior reliability and accuracy in evaluation of muscle fatty infiltration. We used this method to assess the relationship between rotator cuff (RC) muscle fat fractions, tendon disease severity, and subject characteristics. In total, 182 subjects with shoulder symptoms underwent shoulder MR imaging including additional sequences for fat quantitation. Then, fat fraction maps were manually segmented, and custom software was used to compute the fat fraction. Goutallier scores were also obtained. The relationship between fat fraction and tendon tear severity and subject characteristics was assessed with descriptive statistics, analysis of variance, Student t test of different subgroups, and simple and multiple linear regression analysis. Statistically higher supraspinatus fat fractions were observed in subgroups with tendon tears >3 cm, retraction >1 cm, age >50 years, body mass index (BMI) >30, higher Goutallier score, female gender, and longer symptom duration. A significant linear relationship was seen between RC fat fraction and tendon disease severity, age, and BMI but not symptom duration. Multiple regression models with fat fraction and tendon disease, age, BMI, and gender were significant for all 4 muscles (P < .001). The slope of fatty infiltration increase with age was reduced after adjustment for tendon disease, BMI, and gender. RC fat fraction assessed by chemical shift MR demonstrated a significant linear relationship with tendon tear severity, age, BMI, and gender but not with symptom duration.