Placing an external ventricular drain (EVD) at bedside using augmented reality (AR) guidance is more precise than freehand placement and is associated with fewer reinterventions and complications, according to a clinical pilot study of a novel system.

Frederick Van Gestel, MD, a neurosurgery resident at Universitair Ziekenhuis Brussel and Ph.D. researcher at Vrije Universiteit Brussel in Brussels, Belgium, and colleagues report first-in-human results in Neurosurgery.

Conventional neuronavigation systems are restricted to the operating room because of their size and lack of mobility, the authors note: “Moreover, these systems necessitate unintuitive mental transformations of 2D information on remote displays into 3D action on the patient, dividing attention between both.”

In contrast, “AR provides overlays of 3D information onto the real world and has the potential to augment the surgeon’s view with relevant data.”

Headset-based AR navigation system designed specifically for EVD placement

The new system pairs a commercial AR headset, in this case the Microsoft HoloLens II, with surgical navigation software that was custom-built by the researchers. Their solution integrates all components necessary for neuronavigation (i.e. high-accuracy tracking, image display and processing unit) in a standalone headset, without requiring external cameras or computers.

The first step in its use is preoperative imaging and the creation of a 3D model of the patient’s anatomy, which is labeled with coordinates relevant for AR guidance, including the foramen of Monro as a target point. An infrared-labeled reference frame is attached to the patient’s head with a clamping headband, allowing the AR application to track head position and orientation.

The 3D model is exported to the AR headset into the surgical navigation application, and image-to-patient registration is achieved using an infrared-tracked stylus to match the virtual representation of the patient’s anatomy to the physical one. During EVD trajectory planning, the system determines the path to the foramen of Monro with the shortest distance between the skin surface and the ipsilateral frontal horn of the lateral ventricles to ensure the shortest path through brain parenchyma.

The surgeon can adapt this trajectory on the fly by moving the stylus freely around the expected entry point, taking into account underlying anatomy or pathology, before confirming the patient-specific approach.

The trajectory and all necessary anatomical information are displayed as a 3D object on the patient and can be inspected from all angles, providing a sense of depth. A tracker is attached to the surgical drill, and the system indicates the translational and angular error of the drill bit’s tip and drilling direction based on the planned trajectory.

Once these errors are reduced below predefined thresholds of 2 mm and 2°, the trajectory line turns green to indicate proper drill alignment. During drilling, this color-coded guidance continues to give feedback about the alignment of the drill’s trajectory.

Pilot study: Complete absence of poor or failed placements

Following successful phantom studies, Dr. Van Gestel and three other neurosurgeons used the AR navigation system for intraoperative guidance during 11 EVD placement cases in a critical care setting, for adults presenting at their university hospital and requiring bedside EVD placement as their sole primary intervention. They also collected data on 11 patients who had an EVD placed at their bedside using the freehand technique during the same timeframe, performed by one of the surgeons participating in the study.

AR guidance resulted in better placement outcomes and functional placement in all cases versus seven in the freehand group (statistically significant difference), successful in nine versus five, optimal in eight (including two slit ventricle cases) versus three (statistically significant difference), and failed in zero versus one. No AR-guided placement required revision, but four freehand placements did, another statistically significant difference.

“The reinterventions, along with the implied multitude of attempts, constituted the primary cause of all procedure-related complications in the freehand group,” the authors report, “once more emphasizing the hazard related to multiple stick-and-poke attempts and the importance of first-attempt success.”

“Although for some cases freehand placement might have been equally successful, the distinct improvement over the control group indicates that the freehand technique may not consistently yield straightforward results,” Dr. Van Gestel and his colleagues emphasize. “The performance in slit ventricle cases further substantiates these presumptions because accomplishing successful ventricular puncture can be notably challenging, especially within a single attempt.”

Building on their initial success, the researchers are planning a multicenter randomized controlled trial, which is scheduled for launch at the end of 2025.