Structured low-rank approach for high temporal resolution of TWIST imaging

Abstract

Magnetic resonance imaging (MRI) is a widely used modality and dynamic contrast enhanced (DCE) imaging is one of the important methods to diagnose a tumor. In DCE MRI, temporal resolution and spatial resolution are important factors to determine image quality. In particular, time-resolved angiography with interleaved stochastic trajectories (TWIST) contributes wonderful improvement in temporal and spatial resolution. However, the temporal resolution of TWIST is not a true one. This is because the periphery of k-space data from several time frames should be shared to make one image. A conventional method to reconstruct the k-space data is generalized autocalibrating partially parallel acquisitions(GRAPPA) which is one of the popular parallel imaging(PI) algorithms. We proposed the method to improve the temporal resolution of TWIST with comparable spatial resolution. For our purpose, we employ a recently proposed annihilating filter-based low rank Hankel matrix approach (ALOHA). In vivo results showed significantly improved temporal resolution than the standard TWIST reconstruction.