Motivation: Knee joint MRI requires m sequences involving multiple planes and with- and without fat suppression, which is time-consuming. Goal(s): Provides detailed cartilage and ligament information in a short period of time and allows multiplanar reconstruction in any plane of any thickness from a single high-resolution isotropic imaging. Approach: CS-AI 3D mDixon TSE images were compared with conventional SENSE and CS images. Results: It indicated that 3D mDixon CS-AI can provide more accurate image information of cartilage and ligaments with high reproducibility and robustness. Impact: This technique may be helpful to further assess knee pathology.

3D high-resolution isotropic multi-contrast knee MRI using 3D mDixon TSE with deep learning constrained Compressed SENSE