Estimation of pressure gradient images from velocity encoded MR acquisitions

Abstract

Magnetic resonance imaging (MRI) offers the potential to estimate blood pressure gradients, by solving the Navier-Stokes model relating the haemodynamic pressure to the acceleration and velocity. Here we compared results directly obtained using acceleration encoded MR sequences with those calculated from MR velocity acquisitions available on clinical systems. We found that using velocity encoded data instead of acceleration encoded data significantly degrade gradient pressure images. Two experimental phantoms were used to separately evaluate the inertial and convective components of the acceleration. To improve the quality of the pressure gradient estimation, the regularization of the acceleration field, calculated from velocity encoded acquisitions was investigated. We also suggest to increase the temporal resolution while lowering spatial resolution to lessen motion artifacts.