Realistic simulations of aorta radius estimation

Abstract

This paper studies estimation of the dynamic aorta radius in a realistic geometry where radius variation is used as an indirect measure of central blood pressure. Four different radius estimation approaches were studied where their performance in terms of precision and sensitivity was compared. As a basis for estimation simulations, finite-difference, time-domain electromagnetic simulations of a realistic human model have been performed. Radius estimation should be based on identifying the front and rear reflections from the aorta, however the temporal sensitivity of the front reflection from the aorta is weak and non-linear; therefore robustness of radius estimates is compromised. Nonetheless, this does not preclude using the sensitivity of the rear reflections as a proxy of aorta diameter variation combined with Moens-Korteweg's relationship to perform estimations of mean pressure. Proxies of radial changes are observable and for a precision of around 0.1 mm, the results show that an emitted energy to receiver noise spectral density ratio between 110 dB and 130 dB should be sufficient, depending on the estimator.