Intra-Operative Trajectory Realignment Algorithm for Minimizing Pedicle Screw Misplacement in Cervical Spinal Fusion

Abstract

Cervical spinal fusion is a surgical procedure that requires precise pedicle screw placement to ensure successful outcomes and avoid complications. Traditional manual methods rely on the surgeon's experience and anatomical landmark identification, which can lead to misalignment errors and complications. This study proposes an intra-operative trajectory re-alignment algorithm designed to address discrepancies between pre-operative and intra-operative images, particularly due to the inherent mobility of the cervical spine. The algorithm utilizes 3D CT scans for pre-operative imaging and 2D X-rays for intra-operative imaging. Principal Component Analysis (PCA) is applied to each 3D segmented spine to construct reference frames. These kinematic frames are manipulated to optimize the error between the simulated X-ray images and the reference 3D CT images. The CTspine1K data was used to evaluate the proposed algorithm by measuring geometric and clinical errors. The results show that the proposed algorithm reduces errors, and preventing screw perforation.