Surface-based registration accuracy of CT-based image-guided spine surgery.

Area of Science:

• Neurosurgery
• Medical Imaging
• Orthopedic Surgery

Background:

• Accurate image-guided surgery is crucial for spinal procedures.
• Surface-based registration is a key component of CT-based navigation systems.
• Optimizing intraoperative data sampling is essential for registration accuracy.

Purpose of the Study:

• To evaluate the impact of intraoperative data sampling area and point number on CT-based spinal navigation registration accuracy.
• To determine the optimal sampling strategy for surface-based registration in spinal surgery.
• To assess registration accuracy using a cadaveric lumbar spine phantom and an optical 3D localizer.

Main Methods:

• A cadaveric lumbar spine phantom was used with ceramic balls for CT imaging.
• Twenty surface points were digitized from five defined zones on the posterior vertebral body.
• Registration accuracy was evaluated based on positional and rotational errors, analyzing the lamina as a standard zone.

Main Results:

• Using 20 surface points on the lamina yielded positional error of 0.96 mm (RMS) and rotational error of 0.91 degrees (RMS).
• The lamina was identified as the most surgically accessible and easily digitized zone.
• Combining lamina points with points from other zones (e.g., spinous process) improved accuracy in cases with spinal deformities.

Conclusions:

• Twenty surface points on the lamina generally ensure sufficient accuracy for computer-aided spine surgery.
• In complex cases like severe spondylosis, combining sampling points from the lamina and other anatomical areas enhances registration precision.
• Optimizing surface sampling strategies is vital for reliable CT-based spinal navigation.