Usage of augmented reality for interventional neuraxial procedures: A phantom-based study.

BACKGROUND: Neuraxial access is necessary for an array of procedures in anaesthesia, interventional pain medicine and neurosurgery. The commonly used anatomical landmark technique is challenging and requires practical experience. OBJECTIVE: We aimed to evaluate the technical feasibility of an augmented reality-guided approach for neuraxial access and tested the hypothesis that its use would improve success as the primary outcome. As secondary outcomes, we measured accuracy and the procedural duration compared with the classical landmark approach. DESIGN: A randomised phantom-based study. SETTING: The three-dimensional image of a thoracolumbar phantom spine model with the surrounding soft tissue was created with a neurosurgical planning workstation and ideal trajectories to the epidural space on the levels T10-L1 were planned using a paramedian approach. Both the three-dimensional holographic image of the spine and the trajectories were transferred to an augmented reality-headset. Four probands (two anaesthesiologists, one neuroradiologist and one stereotactic neurosurgeon) performed 20 attempts, 10 each of either conventional landmark or augmented reality-guided epidural punctures, where anatomical level, side and sequence of modality were all randomised. OUTCOME MEASURES: Accuracy was assessed by measuring Euclidean distance and lateral deviation from the predefined target point. Success of epidural puncture on the first attempt was compared between the conventional and the augmented reality-guided approaches. RESULTS: Success was achieved in 82.5% of the attempts using augmented reality technique, compared with 40% with the conventional approach [ P  = 0.0002, odds ratio (OR) for success: 7.07]. Euclidean distance (6.1 vs. 12 mm, P  < 0.0001) and lateral deviation (3.7 vs. 9.2 mm, P  < 0.0001) were significantly smaller using augmented reality. Augmented reality-guided puncture was significantly faster than with the conventional landmark approach (52.5 vs. 67.5 s, P  = 0.0015). CONCLUSION: Augmented reality guidance significantly improved the accuracy and success in an experimental phantom model of epidural puncture. With further technical development, augmented reality guidance might prove helpful in anatomically challenging neuraxial procedures.