Robust localized stiffness assessment by combining flow and wall motion using a 1-D wave propagation model

Arterial stiffness has been reported to be a predictor of all-cause cardiovascular mortality. Ultrasound pulse wave imaging (PWI) uses motion tracking to image the vessel wall displacement resulting from the propagation of the arterial pressure and blood flow velocity at high frame-rate (3.3kHz). Normally, data from only one frame per scan line is used to estimate the pulse wave velocity (PWV, Method A), while every frame carries information about the vessel properties which could be utilized. The first aim of this paper is to investigate the benefit of flow by utilizing the numerical 1- D wave propagation model (Method B). Second, the feasibility and benefit of utilizing a physics-informed neural (Method C) over the numerical approach is investigated. All approaches showed good agreement in simulation. Interestingly, when assessing the approach in a small subset of healthy and diseased subjects, the approach that utilized the most data, i.e. Method C, was able to best differentiate the two groups (p=0.07). Potentially due to the increased robustness caused by utilizing more of the available measurements.