(CaRad Study) Radial vs Cartesian 4D Flow MRI: An In Vitro and In Vivo Comparison in the Portal Circulation

4D Flow Magnetic Resonance Imaging (MRI) is a powerful technique that allows the non-invasive measurement of velocity in all 3 spatial directions, resolved in time. A promising evolving application is the characterization of hepatic hemodynamics, e.g., in patients with cirrhosis and gastroesophageal varices. There are only a few groups specializing in this area. The two main sequences that are used by these groups differ in their spatial encoding methods for filling the k-space: line by line Cartesian versus radial sampling. Vendor-provided 4D flow sequences typically use Cartesian volumetric data collection with acceleration methods such as parallel imaging, compressed sensing, or kt-acceleration. PC-VIPR (phase contrast vastly undersampled isotropic voxel radial projection imaging) leverages radial undersampling to accelerate data acquisition. It thereby allows for larger volume coverage and higher spatial resolution compared to Cartesian sampling. To date, no direct comparison between these two spatial sampling approaches exists for the abdomen. However, reproducibility between sequences is essential for broad clinical implementation of 4D flow MRI at different sites, clinical follow up, and generation of reference values.
The goal of this study is to compare PC-VIPR and the Cartesian GE product sequence in the portal circulation both in vitro and in vivo. We will determine the reproducibility between the two sequences, and evaluate the accuracy, bias, precision, and image quality of both methods. As a secondary goal we will also compare abbreviated Cartesian and PC-VIPR sequences aimed at reduced scan time through time-averaging, a strategy that exploits the non-pulsatile flow in the portal circulation.
This project consists of two phases: Firstly, we will acquire both sequences in an anatomical accurate 3D printed portal vein flow phantom and compare the results of the 4D flow measurements between sequences, as well as with particle imaging velocimetry (PIV) as an external reference standard. Secondly, we will examine 24 volunteers that can be classified in three groups: 12 healthy volunteers that receive Ferumoxytol as contrast agent for optimal acquisition conditions; 6 healthy subjects that do not receive a contrast agent to test the sequences under challenging acquisition conditions; and 6 patients with portal hypertension who will receive Ferumoxytol to test the sequences in a challenging cohort. The Cartesian 4D flow sequence and PC-VIPR will be acquired for ~12 minutes each, followed by an abbreviated version of both sequences (~3-6min). This protocol will be repeated to evaluate test-retest repeatability. In each data set, we plan to measure flow and velocity at standardized contours in the portal venous system, systemic venous, and arterial abdominal vasculature. Secondary flow patterns and image quality will be semi-quantitatively evaluated on a Likert scale.