Abstract
PURPOSE: Our previous research demonstrated that under ideal conditions, rigid registration between MRI images and US images had high accuracy for real-time image guidance. The work presented in this paper focused on the application of the previously established procedures to a new context, including preoperative CT images. MATERIALS AND METHODS: We used a template to calibrate the US probe and completed the registration between preoperative CT images and US images. Marker experiments on the accuracy of real-time needle trajectories in CT images were performed using micro electromagnetic sensors. Pelvic phantom experiments were carried out to test the registration accuracy between CT and US images, in addition to registration accuracy between US images and real-time needle trajectories (real-time space model). RESULTS: The US probe calibration error in CT images was 0.879 ± 0.149 mm. The difference of registration between US images and CT images was 0.935 ± 0.166 mm in the axial plane (n = 30) and 0.916 ± 0.143 mm in the sagittal plane (n =12). The difference of registration between US images and the needle's real-time trajectories was 0.951 ± 0.202 mm. CONCLUSION: Under ideal conditions, rigid registration between CT images and US images had high accuracy for real-time image guidance.