Deep Learning-reconstruction of rapid 3DEPI acquisitions for enhanced QSM in the clinical assessment of Multiple Sclerosis

Motivation: Adoption of QSM in Multiple Sclerosis(MS) is constrained by the trade-off between acquisition time and image quality of clinically-applicable protocols. Goal(s): Evaluating the potential of a deep learning-based reconstruction(DLR),denoising and super-resolution pipeline to enhance the clinical utility of 3DEPI-based QSM in MS. Approach: χ-maps derived from DLR- and conventionally-reconstructed(CR)-GRE,1 and 2-averages-3DEPI from a healthy individual were quantitatively compared. DLR- and CR-3DEPI-based χ-maps from 7 MS-patients were rated for MS-specific biomarkers identification and quality. Results: DLR-3DEPI-based QSM was comparable to state-of-the-art GRE-based QSM,with 4-fold reduction in acquisition time. Ratings showed significant improvement in MS-specific biomarkers identification and image quality for DLR-3DEPI-based vs CR-3DEPI-based QSM. Impact: Deep learning-based reconstruction, denoising and super-resolution pipeline substantially enhances the quality of QSM maps obtained from fast 3DEPI. This holds promise for advancing the broader implementation of QSM in the clinical management of Multiple Sclerosis.