Holo-Stroke-CTA: Stroke Hologram Teleportation for CTA Large Vessel Occlusion Assessments
Stroke: Vascular and Interventional Neurology
Abstract
Background: Augmented reality enables visualization of and interaction with both physical and virtual environments. Holograms can allow 3‐dimensional image transmission to distant sites, allowing patients to interact with providers as if in the same space. Our prior publication resulted in high satisfaction/immersion for patients interacting with Holo‐Stroke providers. Our aim here was to determine if providers assessing computed tomographic angiographies (CTAs) for large vessel occlusion would result in reliability and satisfaction.
Methods: Thirty‐six head CTAs were deidentified and scored by stroke faculty, fellows, and nurse practitioners for large vessel occlusion using digital imaging and communications in medicine (DICOM) viewer. CTAs were presented 2 months later via Holo‐Stroke. Holograms were positioned in 3‐dimensional space, viewable through the Hololens‐2, and scored by the same providers. Kappa reliability was assessed comparing scores to gold standard (radiology report). Satisfaction was assessed via Likert scale.
Results: Thirteen providers scored the CTAs. Overall Kappa reliability, compared with gold standard, was 0.78 (81%) DICOM versus 0.94 (94%) Holo‐Stroke‐CTA (P<0.0001). Overall % correct was 81% versus 94% (P<0.001). Holo‐Stroke‐CTA reliability improved for most examiners: Overall (κ = 0.78 [81%] versus 0.94 [94%]), faculty (κ = 0.85 [87%] versus 0.92 [93%]), nurse practitioners (κ = 0.81 [83%] versus 0.90 [92%]), and fellows (κ = 0.68 [72%] versus 0.97 [97%]). Overall middle cerebral artery (κ = 0.76 [86%] versus 0.93 [96%]), internal carotid artery (κ = 0.8 [88%] versus 0.9 [94%]), and basilar (κ = 0.73 [95%] versus 0.82 [96%]) scored high, with marked improvement for anterior cerebral artery (κ = 0.3 [39%] versus 0.91 [94%]), and posterior cerebral artery (κ = 0.55 [70%] versus 0.95 [98%]). Likert satisfaction “overall” was 18 DICOM, 48 Holo‐Stroke‐CTA (P = 0.002 with the percentage increasing from 39% to 96%. “Immersion” scores were 0, 10 (P = 0.001), “ease of use” 5, 9 (P = 0.002), “accuracy” 7, 9 (P = 0.002), “technology advancement” 4, 10 (P = 0.001), and “interest” 3, 10 (P = 0.002).
Conclusion: Holo‐Stroke‐CTA resulted in higher reliability and satisfaction versus standard DICOM telestroke teleradiology. Providers noted the ability to see 3‐dimensional vessels in virtual space versus scrolling through axial/sagittal/coronal images, resulting in higher accuracy. Even for trainees and difficult‐to‐assess vessels, providers were more able to identify large vessel occlusions using Holo‐Stroke‐CTA. Providers were enthusiastic for the immersive radiology assessment, with the ability to immersively resize, rotate, and investigate hologram in 3‐dimensional virtual space. Though further assessments are needed, Holo‐Stroke‐CTA can help providers more easily, and at a glance, evaluate CTA for large vessel occlusion.
Nadir Weibel
Professor of Computer Science and Engineering
