gms | German Medical Science

Objective: Augmented reality (AR) visualization is a promising tool to increase imaging data use efficacy during surgical procedures. Despite the general clinical availability of AR visualization in microscopic devices, published studies are limited in number, sample and detailed practical recommendations. With release of the current navigation software generation offering improved visualization options for intraoperative overlay, there is an increased demand to identify suitable pathologies, visualization and display modes, and to provide optimal workflow integration for routine utilization.

Methods: A prospective study protocol for comparative evaluation of AR and non-AR cases has been designed. All navigated lesions with compatible imaging data are included and randomly assigned to one group. In AR cases, microscope and navigation streams are recorded. Type of view, quality, accuracy, complexity and duration of image injection in the heads-up display (HUD) are matched with the corresponding surgical task and surgical outcome. In non-AR cases, navigation interaction and surgical outcome are documented.

Results: During the first cases (n=7, all brain tumors WHO grade I to IV), AR was utilized in 15% of total resection time. Frequently used HUD display modes were navigation view (70%) and target volume (40%). All surgeons used tumor segmentation in AR visualization with additional anatomical (35%) and functional (50%) structures. In target volume mode, the average overlay accuracy of tumor margins has been rated 65% (0-7mm offset) by the surgeon during early resection. Workflow facilitations, such as pointer-free navigation and fade-in display of surgical information, were accompanied by partial blocking of the surgical field and impaired depth assessment. During the later resection phase, surgeons switched back to conventional navigation (50%) or used it in addition to AR overlay (100%). In non-AR resection, pointer-based navigation checks were associated with frequent workflow interruptions (4-7 per case).

Conclusion: Aside from improved workflow management, preliminary study results underscore the relevance of a continuous case evaluation. While the technical workflow is largely compliant with daily surgical routine, visualization quality still impacts surgical cognitive load and surgical performance. Future subgroup analyses and case-control matches will be used for further quantification of results and correlation with surgical outcome.