Rapid Prototyping of Two-Dimensional Non-Cartesian K-Space Trajectories (ROCKET) Using Pulseq and Graphical Programming Interface.

Abstract

Magnetic resonance imaging is a well-established method for diagnostics and/or prognostics of various pathological conditions. Cartesian k-space trajectory-based acquisition is the popular choice in clinical magnetic resonance imaging, owing to its simple acquisition, reconstruction schemes, and well-understood artifacts. However, non-Cartesian trajectories are relatively more time efficient, with involved methods for image reconstruction. In this review, we survey non-Cartesian trajectories from the standpoint of rapid prototyping and/or implementation. We provide examples of two-dimensional (2D) and 3D non-Cartesian k-space trajectories with analytical equations, merits, limitations, and applications. We also demonstrate implementation of three variants of the 2D radial and spiral trajectories (standard, golden angle, and tiny golden angle), using open-source software. For rapid prototyping, pulse sequences were designed with the help of Pulseq. In-vitro phantom and in-vivo brain data were acquired with three variants of radial and spiral trajectories. The obtained raw data were reconstructed using a graphical programming interface. The signal-to-noise ratios of each of these reconstructions were quantified and assessed.