Pre- and Intra-operative Dynamic Registration for Total Knee Arthroplasty Based on CT Image Annotation

Abstract

Robot-assisted total knee arthroplasty (TKA) is a promising technique with improved precision and stability when compared with the conventional TKAs performed by surgeons. This paper presents a pre- and intra-operative dynamic registration method based on Computed Tomography (CT) image annotation for an automated robotic surgery system for TKA. In order to register the pre-operatively planned travel path of the oscillating saw to the intraoperative position and orientation of the patient’s lower extremity, tracking tools were fixed to the femur and tibia before surgery for CT scanning. A CT annotation method is proposed to mark the points on the tracking tools’ spherical surfaces. Subsequently, the centers of the spheres are fitted and used to establish the respective coordinate systems. The transformation relationship between the image coordinate system and the pre-operative coordinate system can be obtained. The result shows that registration with a maximum error of 0.0719mm is achieved, which is precise enough to facilitate the intra-operative bone cuts and navigation.