Low-Discrepancy Trajectories for 3D Cardiac Magnetic Resonance Imaging

• Applicant: Dr. Tobias Speidel
• Subject Area: Medical Physics, Biomedical Technology
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 535445977

Magnetic resonance imaging (MRI) has already successfully established itself as the gold standard in clinical routine for the functional analysis of the heart and cardiovascular system. The diagnostic procedure is mainly based on the use of two-dimensional MRI sequences in combination with breathhold commands, as well as an ECG-synchronized data acquisition. Although three-dimensional (3D) methods offer intrinsic advantages, such as higher signal-to-noise ratios and the possibility of reconstructing arbitrarily positioned anatomical planes, they are often not applicable due to long acquisition times (breathhold). Recently, great progress has been made in accelerating 3D MRI acquisitions, particularly based on the combination of efficient MRI sequences and iterative image reconstruction techniques. Established approaches for an efficient acquisition of the frequency domain in MRI have already facilitated distinct reductions in scan duration compared to traditional acquisition geometries (cartesian, radial, 3D Cones). However, all known and published acquisition schemes for the frequency domain show limitations, which make an efficient combination with iterative reconstruction techniques difficult and therefore prevent further reductions in scan duration. The aim of this project is to investigate the, by the applicants already developed acquisition scheme, based on low-discrepancy trajectories with iterative and non-linear reconstruction techniques as well as with self-gating approaches, for three-dimensional cardiac and respiratory motion synchronized and resolved imaging of the heart (3D+t) and the great vessels. The following acquisition scheme (k-space trajectory) forms a generic basis for the accelerated 3D acquisition and reconstruction of the measured data in the spatial frequency domain. In order to achieve the desired and application-specific contrasts (e.g. T2 contrast in coronary vessels) and encodings (e.g. additional bipolar gradients in flow imaging), the acquisition scheme is combined with the desired contrast mechanisms (RF and gradient pulses) to form a final MRI sequence.

• DFG Programme: Research Grants
• Co-Investigator: Professor Dr. Volker Rasche