Clinical Scorecard: Enhanced Localization of Flexible Electrode Grids in Epilepsy Surgery Through Augmented Reality Technology

At a Glance

Key Highlights

• Precise localization of deformable ioECoG grids is critical for successful epilepsy surgery outcomes.
• Current localization methods rely on visual inspection or photograph-based techniques with limitations in accuracy, time, and expertise required.
• Augmented reality head-mounted displays combined with AI techniques offer a novel approach for real-time, accurate, and efficient ioECoG grid localization.

Guideline-Based Recommendations

Diagnosis

• Use intraoperative electrocorticography (ioECoG) grids placed on the cortex to localize epileptic tissue.
• Employ presurgical 3D MRI for anatomical reference in grid localization.

Management

• Consider epilepsy surgery for patients with focal refractory epilepsy unresponsive to medication.
• Utilize augmented reality (AR) head-mounted displays integrated with AI for enhanced localization of ioECoG grids during surgery.

Monitoring & Follow-up

• Monitor ioECoG grid placement accuracy intraoperatively to ensure complete resection of epileptic tissue.
• Compare localization results against established tracking systems (e.g., commercial NDI tracking) as ground truth.

Risks

• Inaccurate localization of ioECoG grids may lead to incomplete resection and suboptimal surgical outcomes.
• Prolonged registration times and need for specialized expertise may delay surgery and increase risk.

Patient & Prescribing Data

Patients with focal epilepsy refractory to anti-seizure medications undergoing epilepsy surgery

Accurate localization of epileptic tissue via ioECoG grids improves surgical success; AR and AI integration may enhance precision and workflow.

Clinical Best Practices

• Integrate AR HMDs with AI-based object detection, super-resolution, and 2D pose estimation to improve ioECoG grid localization accuracy.
• Correct camera image distortion and apply image rectification prior to grid detection to enhance triangulation accuracy.
• Use synthetic data to train AI models to overcome medical data scarcity challenges.
• Aim for localization accuracy within clinically acceptable thresholds (mean error <5 mm).
• Prefer hands-free AR HMDs over bulky or workflow-disruptive devices for intraoperative use.

References

• Pieters et al. on ioECoG grid localization methods
• Augmented reality in surgical navigation
• FDA-approved AR HMD solutions for neurosurgery