Hybrid 3D augmented reality for image-guided therapy using autostereoscopic visualization

During image-guided therapy, cardiologists use 2-dimensional (2D) imaging modalities to navigate the catheters, resulting in a loss of depth perception. Augmented reality (AR) is being explored to overcome the challenges, by visualizing patient-specific 3D models or 3D shape of the catheter. However, when this 3D content is presented on a 2D display, important depth information may be lost. This paper proposes a hybrid 3D AR visualization method combining stereo 3D AR guidance with conventional 2D modalities.

A cardiovascular catheterization simulator was developed consisting of a phantom vascular model, a catheter with embedded shape sensing, and an autostereoscopic display. A user study involving interventional cardiologists ( $n=5$ ) and electrophysiologists ( $n=2$ ) was set up. The study compared the hybrid 3D AR guidance with simulated fluoroscopy and 2D AR guidance in a catheter navigation task.

Despite improvements in task time and traveled path length, the difference in performance was not significant. However, a reduction of 50% and 81% with 2D and hybrid 3D AR in the number of incorrect artery entries was found, respectively. The results of the questionnaires showed a reduced mental load and a higher confidence with the proposed hybrid 3D AR guidance. All but one participant indicated to feel comfortable looking at the hybrid 3D view.

The findings suggest that AR guidance, particularly in a hybrid 3D visualization format, enhances spatial awareness and reduces mental load for cardiologists. The autostereoscopic 3D view demonstrated superiority in estimating the pose and relationship of the catheter relative to the vascular model.