A pulsatile 3D flow relevant to thoracic hemodynamics: CFD - 4D MRI comparison

Application Challenge AC7-04 © copyright ERCOFTAC 2021

Description

Introduction

The objective of the current Application Challenge is to provide a well-controlled environment and procedure to enable comparison between CFD and 4D Flow MRI. The experiment is designed to remove most classical sources of uncertainties inherent to the in vivo MRI such as moving deformable walls or undefined blood properties. The experimental MRI setup is shown on figure 1, where the phantom presents topological complexities analogous to the cardiovascular system (Fig. 2). Large Eddy Simulation (LES) has been conducted on the same geometry. Both the experiment and the simulation have been performed under a sinusoidal pulsatile flow, with flow conditions in the laminar-turbulent transition regime similar to what one can found in the cardiovascular system.

Figure 1: Schematic experimental setup

The methods described in this Application Challenge are mainly adopted from Puiseux et al. (2019) [2].

Relevance to Industrial Sector

4D Flow MRI is a specific type of MRI sequences, which allows to image the velocity field of moving fluids such as blood. Therefore it represents an interesting opportunity for clinicians to detect and follow the evolution of cardiovascular diseases. However, this technique suffers from some limitations, which question its accuracy and efficiency to measure relevant hemodynamics quantities. Although CFD accuracy depends on model assumptions, it also enables to bypass some of the limitations inherent to the MRI process in the way that it provides higher spatio-temporal resolutions and gives access to instantaneous and derived quantities. Thereby the idea of this Application Challenge is to use 4D Flow MRI and CFD to develop a reproducible method to enable the comparison of the two modalities.

Design or Assessment Parameters

The velocity field along the three Cartesian directions within the phantom were measured in the 4D Flow experiment, as well as the velocity field on a transverse plane located at the inlet of the phantom with a 2D cine PC-MRI. This latter technique is similar to the 4D Flow, but since it is performed on a slice only, it permits a higher spatio-temporal resolution. The velocity field was numerically predicted within the phantom as well, where the 2D MRI data were used to apply boundary conditions at the inlet.

Flow Domain Geometry

The flow phantom was constructed to generate a complex and realistic flow, such as that observed in the cardiovascular system. The aim was not to reproduce a patient-specific geometry, but to gather several geometrical features yielding complex flow patterns analogous to in vivo flow patterns, while keeping a relatively compact and well-controlled flow phantom. A 26 mm inner diameter pipe bend was designed with a 50 mm radius of curvature to mimic aortic arch blood flows. A bifurcation was set in analogy with collateral arteries. Daughter vessel sizes were designed to replicate typical flow split that can be found in vivo between supraceliac and infrarenal arteries. Finally, a protuberance was attached at the intersection between the collateral and main branch to mimic blood flows patterns in aortic aneurysm (Fig. 2). Typical flow structures expected are e.g. a flow recirculation in the aneurysm, regurgitation in the bifurcation, or strong longitudinal vorticity and secondary flow in the bended arch.

An engineering drawing of the phantom is given on Figure 3. The phantom is made of Nylon PA 12 with a 60 μm geometric tolerance in one block and immersed within a silicon bath both for post-processing eddy currents effects on MRI images and for securing the phantom.