Determination of pre-arthropathy scapular anatomy with a statistical shape model: part I-rotator cuff tear arthropathy.

ABSTRACT

HYPOTHESIS AND BACKGROUND: Rotator cuff tear arthropathy (RCTA) is a pathology characterized by a massive rotator cuff tear combined with acromiohumeral and/or glenohumeral arthritis. The severity of RCTA can be staged according to the Hamada classification. Why RCTA develops in some patients is unknown. Furthermore, in RCTA patients, distinctly different articular damage patterns can develop on the glenoid side as categorized by the Sirveaux classification (glenoid erosion). The goal of this study was to determine whether an association exists between scapular anatomy and RCTA and different severity stages of RCTA, as well as the associated glenoid erosion types. METHODS: A statistical shape model of the scapula was constructed from a data set of 110 computed tomography scans using principal component analysis. Sixty-six patients with degenerative rotator cuff pathology formed the control group. The computed tomography scan images of 89 patients with RCTA were included and grouped according to the Hamada and Sirveaux classifications. A complete 3-dimensional scapular bone model was created, and statistical shape model reconstruction was performed. Next, automated 3-dimensional measurements of glenoid version and inclination, scapular offset, the critical shoulder angle (CSA), the posterior acromial slope (PAS), and the lateral acromial angle (LAA) were performed. All measurements were then compared between controls and RCTA patients. RESULTS: The control group had a median of 7° of retroversion (variance, 16°), 8° of superior inclination (variance, 19°), and 106 mm of scapular offset (variance, 58 mm). The median CSA, PAS, and LAA were 30° (variance, 14°), 65° (variance, 60°), and 90° (variance, 17°), respectively. In terms of inclination, version, scapular offset, and the PAS, we found no statistically significant differences between the RCTA and control groups. For RCTA patients, the median CSA and median LAA were 32° (P ≤ .01) and 86° (P ≤ .01), respectively. For all investigated parameters, we did not find any significant difference between the different stages of RCTA. Patients with type E3 erosion had a different pre-arthropathy anatomy with increased retroversion (12°, P = .006), an increased CSA (40°, P ≤ .001), and a reduced LAA (79°, P ≤ .001). DISCUSSION: Our results seem to indicate that a 4° more inferiorly tilted and 2° more laterally extended acromion is associated with RCTA. RCTA patients in whom type E3 erosion develops have a distinct pre-arthropathy scapular anatomy with a more laterally extended and more inferiorly tilted acromion and a more retroverted glenoid in comparison with RCTA patients with no erosion. The pre-arthropathy scapular anatomy does not seem to differ between patients with different stages of RCTA.