Development of a novel spherical navigator-based motion measurement technique in magnetic resonance imaging

Buschbeck, Richard Peter Martin; Fitter, Jörg Ludwig (Thesis advisor); Shah, Nadim Joni (Thesis advisor)

Aachen (2019, 2020)
Dissertation / PhD Thesis

Dissertation, RWTH Aachen University, 2019

Abstract

Patient motion is a frequent problem in magnetic resonance imaging (MRI) causing significant degradations of the image quality. While numerous motion correction techniques are available, most of them require time-resolved information on the movements occurring during the MRI scan. One way to measure 3D rigid-body patient motion are so-called spherical navigator (SNAV) scans. SNAVs have an excellent clinical feasibility, but have not gained widespread adoption so far, because available techniques suffer from several technical and practical limitations. This PhD thesis investigates a novel SNAV technique aiming to overcome several of these challenges. The proposed SNAV concept uses a spherical Lissajous navigator (LNAV) at a k-space radius of 0.1/cm. This value is >70 % smaller than the ones used in the previous state of the art and has been considered infeasible due to restrictions in the SNAV trajectories and processing algorithms. Unlike commonly used helical spiral SNAVs, LNAVs require only a single RF excitation, can be acquired very fast at small radii and have moderate gradient requirements. LNAV-based rotation estimations are calculated by means of a spherical harmonics-based algorithm from computer vision, capable of dealing with the small amount of magnitude features associated with small SNAV radii. The translations are calculated as described in the literature. Phantom and in vivo experiments on a commercial 3 T MRI scanner as well as simulations were conducted to investigate the performance of the novel technique. The results suggest that, while having motion measurement accuracies in the sub-degree/sub-millimetre range comparable to previous SNAV methods and maintaining the good clinical feasibility, the new technique has significant advantages over to the previous state of the art. The most important benefits are the reduced LNAV acquisition time (<5 ms), a markedly higher signal-to-noise ratio, gradient demands well within the scanner limits and significantly reduced computational complexity allowing for processing times of <18 ms on a conventional laptop. The latter makes the proposed technique potentially suited for applications in prospective motion correction. This work is a proof of concept demonstrating the feasibility of measuring patient motion with the novel LNAV concept. The results may serve as a basis for further developments and can potentially increase the performance of SNAV-based MRI motion correction in the future.

Institutions

• Department of Physics [130000]
• Chair of Experimental Physics I A and I. Institute of Physics [131110]