(P129) EXPLORATORY IN-CLINIC STUDY EVALUATING FEASIBILITY OF JVP HEIGHT ASSESSMENT BY EARLY PROTOTYPE OF MACHINE VISION ALGORITHM

Background: Congestive Heart Failure (CHF) is one of the world’s most prevalent medical conditions, with an estimated 64M cases globally. CHF causes a significant burden on the healthcare system, with approximately 25% of cases readmitted within one month. Jugular venous pressure (JVP) is a way to noninvasively measure Central Venous Pressure (CVP) and is used to assess volume status and worsening of CHF. It is the measurement of the highest oscillation point of biphasic internal jugular vein pulsations and is inherently difficult to assess. The accepted practice for the measurement of JVP includes Visual Gestalt and the Two-Ruler Technique, both of which have poorly reproducible results and inherent inaccuracies. New technological advancements for measuring JVP have been developed, including Video Magnification, Subpixel algorithms, and NIRS, however, these technologies are currently inaccessible and lack robustness in the clinical setting. The JVPHome medical device is an accessible early prototype using machine vision being investigated for accuracy in the non-invasive measurement of jugular venous pressure.

METHODS AND RESULTS: This pilot study included patients diagnosed with congestive heart failure managed by the Heart Failure Clinic at the Health Sciences Center in St. John’s, NL. Patients initially had their JVP height measured using visual gestalt, the current standard of care which was documented by the responsible healthcare provider. Additionally, patients had their JVP assessed using gold-standard ultrasound detection with the two-ruler method, and then remote video assessment using the JVPHome. Using the JVPHome machine vision algorithm, the JVP impulse was able to be successfully identified in 21 of 24 study participants following assessment by one non-blinded physician. The JVP height above the sternal angle as measured by the JVPHome is plotted relative to those obtained through ultrasound and the 2-ruler method in Figure 2. The Pearson correlation coefficient (r-squared) was determined to be 0.9259. The JVP height above the sternal angle as measured through visual gestalt is plotted relative to those obtained through ultrasound and the 2-ruler method in Figure 3. The Pearson correlation coefficient (r-squared) was determined to be 0.3454.

Conclusion: Results of this initial feasibility study suggest that JVP heights measured remotely by the JVPHome compare favourably relative to standard bedside assessment when correlated with the gold standard of ultrasound. Overall, the JVPHome device presents itself as a novel clinical assessment tool for measuring JVP in a remote setting.