The goal of this study was to develop and validate a non-invasive approach to estimate scapular kinematics in individual patients. We hypothesized that machine learning algorithms could be developed using motion capture data to accurately estimate dynamic scapula orientation based on measured humeral orientations and acromion process positions. The accuracy of the algorithms was evaluated against a gold standard of biplane fluoroscopy using a 2D to 3D fluoroscopy/model matching process. Individualized neural networks were developed for nine healthy adult shoulders. These models were used to predict scapulothoracic kinematics, and the predicted kinematics were compared to kinematics obtained using biplane fluoroscopy to determine the accuracy of the machine learning algorithms. Results showed correlations between predicted kinematics and validation kinematics. Estimated kinematics were within 10 of validation kinematics. We concluded that individualized machine learning algorithms show promise for providing accurate, non-invasive measurements of scapulothoracic kinematics.

This study evaluated a strategy for identifying 3D scapulothoracic orientation using bilateral X-ray scans and 3D scapula models. Both subject-specific scapula models and a scaled general model were utilized. 3D scapulothoracic orientations obtained from X-rays were compared to motion capture data. "Subjects" consisted of a skeletal model of a human torso and ten real bone scapulae. Retroreflective markers were placed on the scapulae and a three-marker triad was placed on the trunk. Marker positions were recorded using an eight camera motion capture system. A biplane X-ray system from EOS Imaging was used to collect two orthogonal 2D images of the skeleton and markers. Custom software was created for the 3D to 2D matching process. The results indicated that the matched orientations compared favorably to motion capture orientations, with RMSE errors ranging from 3.1° to 5.5° and a mean error of 3.9° The proposed strategy was shown to be accurate for both subject-specific models and a scaled general model.