An Automatic Reverse Shoulder Arthroplasty Planning Method Based on Point Cloud Processing

The effectiveness and accuracy of the planning of reverse shoulder arthroplasty plays a key role in the success of the surgery. In this paper, an automated reverse shoulder arthroplasty planning method based on point cloud processing is proposed. CT data of 55 patients were retrospectively collected, and 3D reconstruction was performed using 3D Slicer software to obtain the skeletal model of the shoulder joint area, which was converted into point cloud data. Accurate segmentation of the scapula and articular glenoid was realized by the PointNet++ network, which provided the basis for surgical planning. In the study, PCA and RANSAC algorithms were used to extract the humeral stem axis and the feature points of the scapular glenoid, respectively, and then automatically plan the surgical paths. The results of the study showed that the average differences between this automatic planning method and the actual surgical results in surgical path planning on the humeral and scapular were 2.247 mm and 2.604 mm, respectively, and the angular differences were 1.932° and 3.686°, respectively, which were within acceptable ranges, proving the effectiveness and accuracy of the proposed method.