Fatty infiltration of the supraspinatus on a single sagittal-oblique “Y” view) CT image does not accurately predict the entire muscle condition

Abstract

Introduction The condition of rotator cuff muscles is an important parameter to consider in shoulder arthroplasty, both preoperatively and postoperatively. We developed and compared 2D and 3D quantitative CT methods, and assessed whether supraspinatus muscle degeneration measured on a single sagittal-oblique (“Y” view) CT slice was equivalent to measurements obtained on several or all CT slices encompassing the muscle. Materials and methods We selected 19 patients (12 females, 68.7 ± 11.5 years) who underwent standardized CT prior to shoulder arthroplasty and were divided in two groups: 9 patients had a full-thickness complete supraspinatus tendon tear, while 10 patients had no tear. Preoperative shoulder CT images were reconstructed semi-automatedly to obtain the reference sagittal-oblique (“Y view”) CT slice. Two radiologists independently delineated the contours of the supraspinatus muscle on consecutive 1 mm thick CT slices, from the myotendinous junction laterally to the medial border of the scapula. Muscle degeneration ratios consisting of atrophy, fatty infiltration, and secondary bone formation were then automated quantified for the single reference CT slices, 3 slices (reference + 1 cm medially and 1 cm laterally), 10 slices (reference + 7 slices medially and 2 slices laterally, each at 1 cm interval), as well as for all (range 100–134) slices encompassing the supraspinatus muscle. The interobserver reproducibility of measurements was evaluated. Resultats Supraspinatus atrophy (0.54 ± 0.24) was the main determinant of muscle degeneration (0.55 ± 0.24) on the reference CT slice. Supraspinatus muscle degeneration ratios were significantly higher in patients with full-thickness complete tendon tears (0.71) than in patients with no tear (0.40) (P = 0.004). Fatty infiltration was significantly higher when measured on 10 CT slices than on the single reference slice (P = 0.048). There were no significant differences between 1, 3, 10, and all slices for atrophy, secondary bone formation, and degeneration (P > 0.383). Subgroup analysis revealed that fatty infiltration was significantly underestimated on the single reference slice in patients with full-thickness complete tears, as compared with 10 (P 0.135). There were no significant differences in patients with no tear (P > 0.737). Interobserver reproducibility of measurements was excellent (P > 0.990). Discussion Care must be taken when assessing preoperative fatty infiltration of the supraspinatus muscle on a single sagittal-oblique CT-slice. Conclusion Fatty infiltration of the supraspinatus muscle quantified on a single sagittal-oblique CT slice differs significantly from measurements obtained on 10 CT slices, particularly in patients with full-thickness complete tendon tears.