Automated registration of intraoperative CT image data for navigated skull base surgery

Area of Science:

• Neurosurgery
• Medical Imaging
• Surgical Navigation

Background:

• Image-guided surgery relies on accurate patient-to-image registration.
• Traditional methods often require invasive markers or are less accurate for specific skull base regions.
• Intraoperative computed tomography (CT) offers real-time imaging but requires precise registration.

Purpose of the Study:

• To investigate the accuracy of a novel, fully automated, non-invasive patient-to-image registration method using an intraoperative CT system.
• To evaluate the performance of this registration technique specifically for skull base surgery.
• To compare the accuracy between anterior and lateral skull base regions.

Main Methods:

• A phantom skull study was conducted with 33 target markers.
• CT image data were acquired using an intraoperative CT suite and transferred to a navigation system.
• Registration was performed automatically by tracking the patient and CT gantry with an infrared camera, identifying target markers.
• Accuracy was measured as the spatial difference between image space and phantom markers.

Main Results:

• The automated registration method achieved sufficient accuracy for image-guided skull base surgery, with an average target registration error below 1.2 mm.
• Unlike traditional non-invasive methods, this technique showed no significant difference in accuracy between the anterior and lateral skull base.
• The system demonstrated robustness across different skull base anatomical areas.

Conclusions:

• Fully automated registration utilizing a tracked CT gantry is a highly accurate and robust method for skull base surgery.
• This non-invasive approach eliminates the need for registration markers, simplifying the surgical workflow.
• The consistent accuracy across skull base regions makes it a valuable tool for image-guided procedures.