Development of Innovative Neurosurgical Operation Support Method Using Mixed-Reality Computer Graphics.

Area of Science:

• Neurosurgery
• Medical Imaging
• Computer Graphics

Background:

• Accurate spatial correspondence between preoperative medical images and intraoperative findings is crucial for neurosurgical safety.
• Brain shift during craniotomy reduces the accuracy of traditional navigation systems.
• A novel registration method is needed to address these limitations.

Purpose of the Study:

• To develop and evaluate a spatially accurate registration method using fused 3D computer graphics and intraoperative brain surface photographs.
• To improve the precision of aligning virtual surgical plans with real-time surgical conditions.

Main Methods:

• Developed a mixed-reality computer graphics technique fusing 3D models with intraoperative brain surface photographs.
• Employed nonrigid registration with landmarks and thin-plate spline methods.
• Measured registration accuracy using target registration error (TRE) on 16 glioma patients.

Main Results:

• Achieved a highly accurate TRE of 0.72 ± 0.04 mm across 160 measurement points.
• Average alignment time was approximately 10 minutes using an average of 24.6 landmarks.
• Demonstrated the feasibility of the mixed-reality approach in a clinical setting.

Conclusions:

• Mixed-reality computer graphics provide precise spatial alignment between real and virtual surgical spaces.
• This technology has the potential to enhance neurosurgical safety through complementary visualization.
• Further integration of mixed-reality can improve intraoperative decision-making and surgical outcomes.