Skip to Main content Skip to Navigation

Numerical simulations for phase-contrast magnetic resonance imaging

Abstract : Hemodynamics (blood flow dynamics) is now recognized as a key marker in the onset and evolution of many cardiovascular disorders such as aneurysms, stenoses, or blood clot formation. As it provides a comprehensive access to blood flows in-vivo, time-resolved 3D phase-contrast magnetic resonance imaging (or 4D Flow MRI) has gained an increasing interest over the last years and stands out as a highly relevant tool for diagnosis, patient follow-up and research in cardiovascular diseases. On top of providing a non-invasive access to the 3D velocity field in-vivo, this technique allows retrospective quantification of velocity-derived hemodynamic biomarkers such as relative pressure or shear stress, which are pertinent for medical diagnosis but difficult to measure in practice. However, several acquisition parameters (spatio-temporal resolution, encoding velocity, imaging artifacts) might limit the expected accuracy of the measurements and potentially lead to erroneous diagnosis. Moreover, the intrinsic complexities of the MRI acquisition process make it generally difficult to localize the sources of measurement errors.This thesis aims at developing a methodology for the assessment of 4D Flow MRI measurements in complex flow configuration. A well-controlled experiment gathering an idealized in-vitro flow phantom generating flow structures typical of that observed in the cardiovascular system is designed. The flow is simultaneously predicted by means of a high-order Computational Fluid Dynamics (CFD) solver and measured with 4D flow MRI. By evaluating the differences between the two modalities, it is first shown that the numerical solution can be considered very close to the ground truth velocity field. The analysis also reveals the typical errors present in 4D flow MRI images, whether relevant to the velocity field itself or to classical derived quantities (relative pressure, wall shear stress). Finally, a 4D Flow MRI simulation framework is developed and coupled with CFD to reconstruct the synthetic MR images of the reference flow that correspond to the acquisition protocol, but exempted from experimental measurement errors. Thanks to this new capability, the sources of the potential errors in 4D Flow MRI (hardware, software, sequence) can be identified.
Document type :
Complete list of metadata

Cited literature [187 references]  Display  Hide  Download
Contributor : ABES STAR :  Contact
Submitted on : Wednesday, September 9, 2020 - 10:43:12 AM
Last modification on : Friday, August 5, 2022 - 10:51:46 AM
Long-term archiving on: : Thursday, December 3, 2020 - 12:45:06 AM


Version validated by the jury (STAR)


  • HAL Id : tel-02934286, version 1


Thomas Puiseux. Numerical simulations for phase-contrast magnetic resonance imaging. General Mathematics [math.GM]. Université Montpellier, 2019. English. ⟨NNT : 2019MONTS109⟩. ⟨tel-02934286⟩



Record views


Files downloads