Motion-Robust Reconstruction of Multishot Diffusion-Weighted Images without Phase Estimation through Locally Low-Rank Regularization
Diffusion-weighted MRI is a widely used technique in neuroscience research, as well as for many clinical applications. Compared with single-shot acquisition, multi-shot acquisition help achieve higher resolution and SNR with reduced distortion. But the motion-induced shot-to-shot phase variations lead to ghosting artifacts. In this work, we propose a relaxed model for multi-shot diffusion-weighted MRI reconstruction (shot-LLR), in which spatial-shot matrices are constructed and low-rank regularizations are applied on them. Shot-LLR shows decreased ghost artifacts versus state-of-the-art reconstruction methods (POCS-MUSE and POCS-ICE), and compared with the single-shot image, the 8-shot acquisition and shot-LLR reconstruction together provide higher in-plane resolution (1 mm) and sharper boundaries for clinical evaluation.
Hu Y, Levine EG, Tian Q, Moran CJ, Wang X, Taviani V, Vasanawala SS, McNab JA, Daniel BL, Hargreaves BA. Motion-robust reconstruction of multishot diffusion-weighted images without phase estimation through locally low-rank regularization. Magn Reson Med. 2019 Feb;81(2):1181-90.
Evan Levine and Valentina Taviani are alumni of the BMR group