Back to 2022 Posters


Virtual Reality For Image-based Quantitative Assessment Of The Aortic Root
Michael Lu, Alison Pouch, Robert C. Gorman, Ahmed Aly, Robert Gorman, Jr., Michael Goodwin, Lourdes Al Ghofaily, Michael Stauffer.
University of Pennsylvania, Philadelphia, PA, USA.

OBJECTIVE: Conventional 3D quantification of the aortic valve in pre-operative transesophageal echocardiography (TEE) involves interaction with volume-rendered images and cross-sectional planes on standard computer display. This study evaluates quantitation of aortic root dimensions in pre-operative TEE using VR to augment visualization, depth perception, and ease of measurement in 3D images towards presurgical planning. METHODS: The Oculus Quest 2 VR headset and syGlass annotation system were used to trace the aortic root at the level of the sinotubular junction (STJ) and left ventricular outflow tract (LVOT) in pre-operative TEEs of 5 trileaflet aortic valves and 5 bicuspid aortic valves. The following measurements were computed from the aortic root reconstructions: STJ area, LVOT area, and aortic root length (distance from the center of the STJ to LVOT). VR measurements were compared to those obtained by conventional segmentation in ITK-SNAP. RESULTS: All results are reported as (μ σ). LVOT area measured using VR tended to be higher than that measured with ITK-SNAP (612.2 282.9 mm2 vs. 523.0 209.8 mm2, p = 0.024). STJ area tended to be higher using VR than with ITK-SNAP (802.9 156.4 mm2 vs. 706.2 114.7 mm2, p = .0053). No significant difference was observed in measurement of aortic root length using VR and ITK-SNAP (22.0 4.6 mm vs. 24.8 3.9 mm, p = .091). The percent difference in measurements using the two methods averaged 12.6%. The average tracing time per image was 15 minutes. CONCLUSIONS: VR visualization and tracing of aortic root geometry shows potential for generating comparable measurements of aortic root dimensions from pre-operative 3D TEE. Future work will investigate reproducibility and inter-observer variability of both methods with a larger sample size.


Back to 2022 Posters