Comparison of Manual Two-dimensional and Automated Three-dimensional Methods of Assessing Shoulder Joint Morphology through Computed Tomography Images.

To evaluate the interobserver agreement in the measurement of anatomical parameters of the shoulder using manual methods of two-dimensional (2D) computed tomography (CT) unformatted in the plane of the scapula and to compare them with the automated measurement obtained through the Blueprint (Wright Medical, Memphis, TN, United States) software, which uses reconstructed three-dimensional (3D) images.  The present is a cross-sectional study in which 2D CT images of 38 patients with different diagnoses were used. The anatomical parameters were measured by the manual methods described by Friedman et al., the glenoid vault method, the Maurer et al. method, and shoulder subluxation according to Walch et al., by five independent qualified surgeons and compared with the parameters obtained through the Blueprint automated software.  Significant differences were found between the manual measurement version obtained through the Friedman et al. method and the automated version. The mean values found for inclination did not show statistically significant differences among the methods. The mean value found for subluxation showed significant differences between the average observed in the analyses performed by the automated method and those performed by the surgeons.  The manual measurements of glenoid version and inclination performed by experienced surgeons are effective, and the vault method is superior to the Friedman et al. method in the analysis of severe glenoid deformities.