gms | German Medical Science

Objective: Augmented reality (AR) has the potential to support increasingly complex neurosurgical setups by including surgical information seamlessly into the (visual) workflow. While intraoperative utilization of AR is increasingly discussed as beneficial, exact mechanisms and impact of AR-guided interventions still remain unclear. This study examines intraoperative visualization parameters and clinical impact of AR in brain tumor surgery.

Methods: 50 cases, operated either with AR-navigated microscope (n=34) or conventional neuronavigation (n=16), were examined prospectively. Surgical resection time, duration/type/mode of AR, displayed objects (n, type), pointer-based navigation checks (n) and case-specific visual quality indicators (depth information, accuracy) have been assessed. Usability of control and overall surgical usefulness of AR were rated by neurosurgeons (n=7) on a 5-item Likert scale (1=poor, 5=very good).

Results: Mean patient age was 47,9 (±16; 11-84) with a m/f ratio of 1,27:1. Pathologies included 56% HGG, 16% LGG, 10% metastasis, 10% meningioma, 4% AVM and 4% other; 30% of cases were recurrent cases. In the AR group, AR display was used in 47% of the resection time. Predominant AR type was navigation view (78,8%), followed by target volumes (24,1%). Predominant AR mode was picture-in-picture (PiP) (80,6%), followed by 22,8% overlay display. In 32% of cases, important anatomical structures were partially or completely blocked by AR information. AR control was satisfactory in 61%. No additional objects were used in 13,9% (MRI navigation only), 30,6% used 1 and 2 objects in AR navigation respectively, 25% 3 or more objects (tumor segmentation, tractography, functional mapping data). 67,6% of surgeons found AR visualization helpful in the individual surgical case. AR depth information and accuracy have been rated acceptable (median 3±1,25 and 3±1,11 vs. median 5±0 in neuronavigation). Mean utilization of the navigation pointer was 3,1x/hour (AR) vs. 9,7x/hour (neuronavigation); thus, navigation effort was significantly reduced in AR (p<0,05).

Conclusion: Main benefit of AR visualization in brain tumor surgery is the integrated navigation display. Navigation view (PiP) provides the highest usability while rarely blocking the operative field. Integration of objects depends on pathology and tumor location; the usefulness of more details of anatomical and functional objects is limited by lower depth and accuracy rates.