Controlled aliasing for improved parallel imaging with a 3D spiral staircase trajectory.

Abstract

To introduce a modified 3D stack-of-spirals trajectory and efficient SENSE reconstruction for improved through-plane undersampling, while maintaining SNR efficiency and other benefits of spiral acquisitions. A novel spiral staircase trajectory is introduced. This trajectory is a modified stack of spirals, in which spiral arms are distributed between partitions along kz . The trajectory maintains the efficient separable reconstruction with a Cartesian fast Fourier transform along the kz direction, followed by a 2D slice-by-slice gridding reconstruction. An additional intermediate step introduces a phase correction to collapse the spiral arms into the prescribed slice planes. For data undersampled through plane, this produces aliasing with reduced coherence, controlled by the arm-ordering. Undersampled data can then be reconstructed with reduced g-factor using a conjugate gradient-based iterative SENSE algorithm. The trajectory significantly improves g-factor for through-plane accelerated acquisitions. Improvement manifests through both reduced overall g-factor and reduced structure in the g-factor maps. In the presented experiments, the mean g-factor decreased from 1.26 to 0.93 and the maximum g-factor decreased from 3.89 to 1.15 for R = 2 spiral staircase when compared with stack of spirals, and the mean g-factor decreased from 2.51 to 0.94 and the maximum g-factor decreased from 8.26 to 1.35 for R = 3 spiral staircase when compared with stack of spirals. The novel spiral staircase trajectory offers improved aliasing characteristics for through-plane parallel imaging acceleration in 3D spiral acquisitions.