Abstract QS36: Spatial Fidelity Of Microvascular Perforating Vessels As Perceived By Augmented Reality Holographic Projections

Abstract

Purpose: Breast reconstruction using free-tissue transfer yields improved long-term aesthetic results as reported by patients, but requires increased resources of practitioners and hospital systems. Much of the complexity of microvascular reconstruction can be attributed to flap harvest, as dissection must account for anatomic variability, and is guided at best by ancillary imaging modalities which must be extrapolated to the patient on the table. Augmented reality affords the opportunity to superimpose relevant imaging on a surgeon’s native field of view. This potentially allows integration of imaging in real-time to better guide flap selection and facilitate dissection. We aimed to assess the spatial fidelity of key anatomic landmarks on augmented-reality projections compared to 3D printed anatomic simulacrum models formatted from CT-Angiography (CTA) imaging in patients undergoing autologous breast reconstruction. Methods: 3D-printed composite models of rectus abdominal muscles and associated Deep Inferior Epigastric Arteries (DIEP) with perforating vessels were fabricated from CTA data from de-identified patients. The corresponding CTA data was similarly formatted to be used on the Microsoft Hololens with the Microsoft Mixed Reality Toolkit 2.01 and Unity 2019.2.4 Visual Studio. Coordinates were measured using the OptiTrack Flex 13 Motion tracking system. The 3D printed models were zeroed to an arbitrary reference point maintained for each set specimen. Established anatomic points including upper and lower borders of the umbilicus, superficial location of perforating vessels, and corners of the model were initially measured. The Hololens projection was then superimposed using hand-gesture and voice commands onto the model, and set into place, ensuring immobile position despite translational movement of the headset. The model was then removed, and the corresponding anatomic reference points were recorded in triplicate from two separate vantages. Results: Five sets of 3D printed models and corresponding Hololens projections were measured. Absolute distance from left to right perforators were calculated from measurements of the model and the Hololens projection, from two vantages. Analysis of variance (ANOVA) were conducted on each group of measurements independently for each specimen. In two of the five specimens (#1 and #32), there was no significant difference in inter-perforator distance. The greatest discrepancy of inter-perforator distance between any projection and those measured on 3D models was 5.6 mm in specimen 21. Conclusions: Our data demonstrate that holographic projections can be used to accurately localize relevant three-dimensional structures extrapolated from CT imaging, on the scale of millimeters. Two of five models demonstrated no significant variability in perceived three-dimensional location of inter-perforator distance and, despite significant inter-group variability, Hololens projections were perceived within millimeters of the corresponding physical landmark. Augmented reality provides a potential paradigm shifting adjunct for intra-operative localization of perforating vessels in free-tissue flap harvest.