External Ventricular Drain Placement Using Active Augmented Reality Guidance: A Proof of Concept of a Clinically Integrable System

Abstract

BACKGROUND AND OBJECTIVES: Suboptimal placement occurs in 26% of external ventricular drain (EVD) procedures performed using traditional freehand methods. We developed a low-cost augmented reality stereotactic navigation system aimed at improving accuracy and safety of the procedure, which is readily compatible with existing Picture Archiving and Communication Systems and automated image segmentation algorithms. METHODS: The system integrates cloud storage, image segmentation, trajectory planning, point-based image-to-patient registration, and real-time 3-dimensional guidance superimposed over the surgical field. As a proof of concept, 15 neurosurgeons, neurosurgical residents, and physician assistants used anatomical landmark-based registration to conduct 29 EVD placements on anatomical phantoms with small ventricles within a simulated surgical environment. From postoperative computed tomography, placement accuracy was assessed using the Kakarla grading scale, along with the distance to target and angular deviation. RESULTS: Twenty EVDs (69.0%; 95% CI, 52.1%-85.8%) were graded as optimal Kakarla 1 placements, 4 (13.8%; 95% CI, 1.2%-26.3%) as suboptimal Kakarla 2 placements, and 5 (17.2%; 95% CI, 3.5%-31.0%) as suboptimal Kakarla 3 placements. The mean distance to target was 9.49 mm (SD, 4.64 mm), and the mean angular deviation was 9.20° (SD, 6.35°). The mean workflow time was 22 minutes 45 seconds (SD, 11 minutes 38 seconds), and the system demonstrated a fiducial registration error of 4.00 mm (SD, 1.16 mm). Challenges related to human-computer interaction were identified, suggesting further refinement is needed to optimize usability. CONCLUSION: While the accuracy, user interface, and procedural time of the system require further refinement for clinical implementation, this proof of concept demonstrates the clinical and technical feasibility of an end-to-end 3-dimensional augmented reality system with the potential to enhance the safety and accuracy of EVD placements.