Three-Dimensional Scapular Border Method for Glenoid Version Measurements

Abstract

Background: Variations among methods to measure glenoid version have created uncertainty regarding which method provides the most consistent measurements of morphology. Greater deformity may also make accurate depiction of the native morphology more challenging. This study examined 4 current methods (Friedman, corrected Friedman, Ganapathi-Iannotti, and Matsumura) and an experimental scapular border-derived coordinate system method, to compare measurement inconsistencies between methods and reference systems and assess the impact of glenoid deformity on measured glenoid version. Methods: Three-dimensional scapulae were created from computed tomography (CT) scans of 74 shoulders that had undergone arthroplasty (28 A2, 22 B2, 10 B3, and 14 C glenoids) and 34 shoulders that had not undergone arthroplasty. Glenoid version measurements were made in Mimics using the 4 methods. For the experimental method, scapulae were reconstructed, and 3 orthogonal global coordinate planes (GCPs) were derived from the medial and lateral borders. Version was measured as the angle between the sagittal reference plane and an anterior-posterior glenoid vector. The intraclass correlation coefficient (ICC) was calculated for the Friedman and corrected Friedman methods. Inconsistencies were assessed for all methods using the interquartile range, mean and standard deviation, and repeated-measures analysis of variance. Concordance correlation coefficients (CCCs) were calculated to assess agreement among the methods. Results: Scapular plane-based methods (experimental, Friedman, and corrected Friedman) yielded an average version between −10° and −12°, with average measurement differences among these methods of <2°. Vault methods (Ganapathi-Iannotti and Matsumura) overestimated or underestimated version by an average of 5° to 7° compared with scapular plane-based methods, and showed significant differences of >12° when compared with each other. Scapular plane-based methods maintained consistency with increasing deformity. Conclusions: The other methods of version measurement using the scapular planes as the reference were highly comparable with the corrected Friedman method. However, when the reference plane was the glenoid vault, version measurements were inconsistent with scapular plane-based methods, which is attributed to differences in the reference systems. In surgical planning, the coordinate system utilized will impact version measurements, which can result in variations in the planned surgical solutions. Additionally, as glenoid deformity increases, this variation resulting from the utilization of different coordinate systems is magnified.