Enabling Highly Accelerated Multi-shot IVIM-DWI with a Data-sharing-guided Framework: Joint Image Reconstruction and Biomarker Estimation.

Authors

Chenglang Yuan, Shihui Chen, Liyuan Liang, Xiaorui Xu, Hailin Xiong, Tianbaige Liu, Yi Li, Qiting Wu, Wing Yat Cheung, Edward S Hui, Qi Dou, Hing-Chiu Chang

Published in

NeuroImage. Pages 122233. Sep 11, 2026. Epub Sep 11, 2026.

Abstract

To develop a novel deep learning-based data-sharing-guided framework for jointly generating high-quality intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) images and IVIM-biomarker maps from highly accelerated under-sampled multi-shot IVIM-DWI data.
To enable significant acceleration of data acquisition while providing effective initial inputs for enhanced deep learning-based feature representation, a MUSE-based data-sharing acquisition strategy was developed for multi-shot IVIM-DWI. Combined with k-space data projection and the physical IVIM-DWI signal model, a novel end-to-end framework was proposed for joint image reconstruction and biomarker estimation, ensuring signal consistency between acquired data and the reconstructed images and IVIM-biomarker maps. In vivo experiments were conducted on nine subjects using two 1.5T MRI scanners to evaluate the feasibility, generalizability, and clinical applicability of the proposed framework with various metrics.
The proposed framework effectively minimized residual artifacts, improved geometric fidelity with a mean NRMSE of 0.017, PSNR of 36.71, and SSIM of 0.973, and significantly reduced acquisition time (from 13 minutes and 4 seconds to 3 minutes and 28 seconds), yielding superior IVIM-DWI images compared to conventional SENSE and data-sharing-based MUSE reconstruction (all p-values < 0.01). Furthermore, the proposed framework demonstrated robustness across different MRI scanners and varying scan parameters (e.g., slice thickness, FOV, and in-plane resolution), resulting in improved reconstruction quality and generalization. IVIM-biomarkers derived from undersampled 4-shot IVIM-DWI data using the proposed framework accurately characterized tissue diffusion and perfusion, showing good agreement with those obtained from fully-sampled data.
We demonstrated the effectiveness of the proposed data-sharing-guided framework for highly accelerated multi-shot IVIM-DWI, which has the potential to improve quantitative IVIM-DWI assessment in cerebrovascular and neurological diseases.

PMID:
42727741
Bibliographic data and abstract were imported from PubMed on 12 Sep 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.