Glenoid-based reference system to differentiate shoulder pathologies on plain radiographs

Shoulder radiographs are used for evaluation and the planning of treatment of various pathologies. Making a diagnosis of these pathologies on plain radiographs occurs by recognizing the relationship of the humeral head on the registry of the glenoid. Quantification of these changes in registry does not currently exist. We hypothesize that a geometric relationship of the humeral head and the glenoid exists that is defined on an anteroposterior Grashey view radiograph by the relationship of the best-fit circle of the humeral head relative to the best-fit circle of the glenoid such that relative measurements will define the normal shoulder and the pathologic shoulder.

One hundred fifty-six shoulders were included: 53 normal shoulders, 51 with primary glenohumeral osteoarthritis (GHOA), and 52 with cuff tear arthropathy (CTA). Humeral head best-fit circle was used to define the circle of the humeral head (cHH). A glenoid best-fit circle (cG) was defined by the following rules: (1) best fit of the glenoid articular surface and (2) was limited by the acromion such that either (a) it reaches maximal interaction with the inferior surface of the acromion or (b) the perimeter of the circle is at the lateralmost point of the acromion. The relationship between cHH and cG is defined by measurement of cHH in horizontal and vertical planes relative to the glenoid circle reference. The horizontal displacement angle (HDA) measures the horizontal position of cHH relative to cG, representing the degree of medialization toward the glenoid. The vertical displacement angle (VDA) measures the vertical position of cHH relative to cG, representing the degree of superiorization toward the acromion. Angles were compared by diagnosis and sex.

The mean HDA was 61.0° (95% confidence interval [CI] 60.3°-61.7°) in normal shoulders, 79.9° (95% CI 76.9°-82.9°) in GHOA, and 63.4° (95% CI 61.7°-65.1°) in CTA (P < .001). The mean VDA was 43.1° (95% CI 42.2°-44.0°) in normal shoulders, 40.9° (95% CI 39.9°-42.0°) in GHOA, and 59.7° (95% CI 57.6°-61.7°) in CTA (P < .001). Interobserver reliability was 0.991 (95% CI 0.94-1.0) and intraobserver reliability was 0.998 (95% CI 0.99-1.0). The geometric relationship of cHH to the glenoid circle reference was plotted for each group.

A geometric relationship exists of the humeral head in reference to the glenoid circle. Together, the HDA and the VDA distinguish between a normal shoulder and those with GHOA or CTA. This suggests that this novel methodology may provide a preoperative planning tool that is easily accessible.