Abstract 18245: ViTrack™ : A Computer Vision Technology for Continuous Blood Pressure and Advanced Cardio Hemodynamics
Mohan Thanikachalam, Greg Boverman, Gokul Prasath Rajamanickam, Stamatios Aleiferis, Brendan Huang, José Wong- Physiology (medical)
- Cardiology and Cardiovascular Medicine
Introduction: There is an unmet need for a wearable device for accurate and continuous non-invasive blood pressure (BP) and other cardio-hemodynamic parameters in real-world situations.
Methods: To address the need, we have developed ViTrack™, a wrist wearable optomechanical sensor capable of 3D contact sensing at the microscale. ViTrack, based on a fundamentally new method, utilizes pressure-dependent spatiotemporal skin displacement to directly measure beat-to-beat SBP and DBP. ViTrack also measures heart rate and respiratory parameters, and advanced hemodynamic parameters. In real-world settings, patient activities such as body or arm movements introduce motion artifacts that interfere with BP measurement. To address this, ViTrack’s 3D imaging system has been enhanced with advanced computer vision and AI/ML techniques to ensure precise and consistent measurements, regardless of patient movement or wrist position relative to the heart (hydrostatic pressure change).
Methods: We have built a minimum viable product. Animal studies (n = 12) showed that the device correlates well (R2 value = 0.958) with invasive arterial pressure (IAP). In outpatient settings, in more than 120 subjects, both male and female, with a wide range of BP (86-180 SBP; 60-110 DBP), when compared to the auscultatory method, ViTrack measures BP in accordance with FDA standards (Association for the Advancement of Medical Instrumentation [AAMI] standard: mean offset < 5mmHg; standard deviation <8 mmHg). Subsequently, in unconstrained patients (n=14), we compared the ViTrack continuously over 2 hours against IAP in real-life ICU conditions. The arm with ViTrack was unconstrained, while the opposite arm was constrained because of the IAP line. With the implementation of the computer vision and AI/ML algorithms, when beat-to-beat measures were averaged across 20 seconds and compared to IAP, ViTrack demonstrated excellent accuracy for both DBP (mean error = -1.37 mmHg [SD = 3.8 mmHg]) and SBP (mean error = -2.69 mmHg [SD = 5.99 mmHg]).
Conclusions: Studies show that ViTrack, a wearable device, can accurately and continuously measure BP in real-world settings. The technology could disrupt the way we measure and manage BP. A multi-center trial is further validating the technology.