Joint Reconstruction of Multiple b-Values and Multiple Directions for Accelerating Diffusion Kurtosis Imaging

While multishot interleaved echo-planar imaging (iEPI) enables higher resolution diffusion kurtosis imaging (DKI) compared to single-shot EPI, its clinical application is hindered by the lengthy acquisition time. This study proposes a novel model-based reconstruction approach to accelerate iEPI DKI acquisition.

The proposed model-based framework directly estimates DKI tensors from k-space data through joint reconstruction of all k-space data across multiple b values and diffusion directions. It incorporates the intrinsic DKI signal model as a prior and integrates two key components: (1) total variation (TV) regularization to suppress noise in DKI tensor maps, and (2) a physically relevant (PhyR) constraint to ensure biologically plausible parameter estimates, termed mDKI-TV-PhyR. The performance of mDKI-TV-PhyR is compared with its two variants (mDKI and mDKI-TV) and conventional reconstruction-fitting pipelines using both simulated and in vivo data with 4-fold in-plane undersampling.

Compared to the conventional methods, the proposed mDKI-TV-PhyR method achieves lower RMSE for all DKI parameters. Bland-Altman analysis shows the smallest bias for FA and MK, as well as the narrowest limits of agreement for FA and MD. Compared to mDKI-TV, mDKI-TV-PhyR produces MK maps without "black holes," exhibits improved stability across all DKI parameters, and achieves lower FA bias.

The proposed method shows substantial promise for clinical applications where both temporal efficiency and spatial resolution are paramount.