Feature-Based Registration Framework for Pedicle Screw Trajectory Registration Between Multimodal Images

Pedicle screw placement for vertebral fixation is a complicated surgery orthopaedic surgeons. The main challenge to estimate the accurate trajectory’s position minimize post-operative complications related pedicle placement. Different types of 3D 2D registration techniques have been employed avoid misplacement during surgery. However, these cannot be applied directly MR X-ray due differences in image intensity and tissue non-correspondence. To overcome limitations, feature-based technique was developed map trajectory intra-operative on pre-operative image. framework validated by generating projection images that perfectly matched simulated images, then back-projecting using estimated transformation parameters. accuracy registered evaluated measuring displacement directional errors between planned trajectory. proposed method successfully position. A number experiments are performed dataset assess effectiveness method. Euclidean distance entry end points error from were below 1mm AP, Lateral, combination both planes. mean length difference less than 1mm.