Accelerating Multi‐Dimensional Magnetic Resonance Imaging of Blood Flow and Turbulence in the Cardiovascular System

Abstract

Disturbed hemodynamic conditions are often related to pathologies of the cardiovascular system. Phase‐contrast Magnetic Resonance Imaging (MRI) provides a non‐invasive technique for the assessment of time‐resolved blood velocity vector fields within arbitrary imaging volumes. Besides velocity vector field information, parameters related to turbulence can be calculated with advanced multi‐point velocity encoding schemes. However, long scan times are currently the main barrier for the acceptance of the method in a clinical setting. The following work presents data‐driven MRI reconstruction algorithms for undersampled measurements with the focus on accurate flow quantification and visualization.
An extension of an auto‐calibrated parallel imaging reconstruction framework for arbitrary kspace trajectories is proposed. The exploitation of temporal correlations as present in timeresolved data demonstrates further advances of scan time reduction when assessing mean velocity and turbulent kinetic energy. While most prior knowledge imposed in advanced MR image reconstruction is designed to work on magnitude images or assumes smooth background phase behavior, dedicated provisions are required for image reconstruction of phase‐contrast MRI data. To this end, it is proposed to incorporate the divergence‐free condition of blood flow into a separate magnitude and phase reconstruction framework for improving the accuracy of image reconstruction of blood velocity vector fields. To address respiratory motion artifacts, retrospective non‐rigid respiratory motion correction incorporated into an iterative parallel imaging reconstruction algorithm is proposed. Furthermore, optimized k‐t sampling patterns are derived for combined parallel imaging‐ and compressed sensing‐based scan acceleration. Finally, the dynamic parallel imaging technique is applied to study blood flow and turbulence patterns in a relevant patient population with congenital heart disease.

Date of Award	2015
Original language	English
Awarding Institution	King's College London
Supervisor	Sebastian Kozerke (Supervisor) & Tobias Schaeffter (Supervisor)