Glenohumeral decentering in rotator cuff deficiency: relationship to rotator cuff muscle, scapula morphology, and shoulder function

Abstract

Background: The purpose of this study is to assess the influence of 3-dimensional (3D) quantitative rotator cuff muscle volume, percentage fat infiltration and scapula morphology on decentering of the humeral head (HH) from the glenoid in patients with severe rotator cuff deficiency and to assess the impact of this on preoperative American Shoulder and Elbow Surgeons (ASES) survey score and active range of motion (ROM).

Methods: This was a retrospective observational study including patients with either massive rotator cuff tears or rotator cuff tear arthropathy. ASES score and active ROM were obtained. The scapula, humerus, and rotator cuff muscles were segmented from computed tomography scans to create 3D models. Displacement analysis of the (HH) and glenoid depth (GD) from the glenoid centroid was performed with reference to the glenoid plane. Furthermore, decentered shoulders were defined as greater than 10% HH displacement from the GD in both the anteroposterior and inferosuperior planes. Backward hierarchical linear regression analysis used to identify parameters associated with HH displacement and poor shoulder function.

Results: Sixty two patients with severe rotator cuff deficiency (55 rotator cuff tear arthropathies and 7 massive rotator cuff tears) were included. In assessing displacement relative to the glenoid centroid, posterior displacement of the GD was associated with greater glenoid inclination (P = .014), while superior displacement was associated with posterior acromial height (P = .031) and steeper acromial tilt (P = .017). Posterior displacement of the HH was associated with glenoid inclination (P < .001), a greater ratio of 3-dimensional fat infiltration percentage (3DFI%) in posterior to anterior rotator cuff (P = .017), and greater subscapularis 3DFI% (P = .017). Superior displacement of the HH was associated with decreased posterior acromial coverage (P = .037) and steeper acromial tilt (P = .037). In assessing displacement of the HH from the GD, anterior displacement of the HH was associated with greater anterior:posterior rotator cuff 3DFI% (P = .045), while superior displacement was associated with increased subscapularis 3DFI% (P = 003). Decentered shoulders were associated with significantly lower ASES score (-8.4, P < .001), forward flexion (-58°, P = .011), abduction (-36°, P = .014), external rotation (-68°, P < .001), and internal rotation (-1.8 points, P = .014).

Discussion: A novel method to assess 3D glenohumeral decentering is proposed. In the setting of rotator cuff deficiency, fat infiltration of the subscapularis is associated with superior HH displacement, while greater fat infiltration of the anterior relative to the posterior rotator cuff muscles is associated with anterior HH displacement from the GD. Glenohumeral decentering is significantly associated with diminished shoulder function and active ROM in all planes.

Keywords: Rotator cuff tear; artificial intelligence; machine learning; reverse shoulder arthroplasty; rotator cuff arthropathy; total shoulder arthroplasty.