Improving group velocity based estimates of arterial stiffness

Abstract

Arterial stiffness is a predictor of cardiovascular diseases, the leading causes of death worldwide. Ultrasound-based methods that measure arterial wave propagation have promise for evaluating local stiffness in vivo. However, geometric properties of arteries cause dispersion, invalidating typical assumptions underlying the relationship between shear modulus G and group velocity cg, which clinical implementations of ultrasound shear wave elastography do not consider. Here, we examine the dependence of these estimates on geometry and evaluate alternative approaches. Wave motion in the proximal wall of an artery after application of focused acoustic radiation force is simulated with a semi-analytical finite element (SAFE) model, and cg is estimated using a time-to-peak algorithm to determine G with several methods. First, the value of G a clinical scanner would report is calculated assuming a bulk medium. Second, a pulse wave velocity (PWV)-based G estimate is calculated by taking cg = PWV and applying the Moens–Korteweg equation. Third, we develop an interpolation-based method to provide a corrected G estimate using data generated by the SAFE model. Simulation results show severe geometry-dependent bias with the bulk method, which is partially ameliorated with the PWV method and substantially improved with the interpolation approach. Results are validated using arterial phantoms.