Reliability analysis of glenoid component inclination measurements on postoperative radiographs and computed tomography-based 3D models in total and reversed shoulder arthroplasty patients

Background: Only a few articles describe the reproducibility and clinical feasibility of glenoid inclination measurements on conventional radiographs, and none of them validated their method in shoulder arthroplasty cases. From a clinical point of view, the angle measured between the supraspinatus fossa and the glenoid fossa line (angle β) appears to be the most interesting angle to assess glenoid inclination. This study aimed to validate the angle β in shoulder arthroplasty patients to facilitate the assessment of glenoid component inclination.

Materials and methods: Seventeen patients who underwent total or reverse shoulder arthroplasty were evaluated. The angle β was measured by 2 independent observers on postoperative radiographs and 3-dimensional (3D) models. The interobserver variability and accuracy of angle β were analyzed by calculating the intraclass correlation coefficient (ICC) and by generating Bland-Altman plots.

Results: The angle β showed a good interobserver variability (ICC = 0.971 for radiographs, ICC = 0.980 for 3D models) and a good agreement between the radiographic and 3D measurements (ICC = 0.904 for observer 1 and ICC = 0.908 for observer 2). Bland-Altman plots demonstrated that in 95% of the measurements on radiographs, the error will be <10. In the investigated population, 85% showed an error <6.

Conclusion: This study demonstrates that angle β can be measured on radiographs to assess glenoid component inclination in total and reverse shoulder arthroplasty, but clinicians and researchers should keep in mind that measurement errors of 10° may occur in a minority of cases.