Single breath hold 3D isotropic whole heart coronary MRA using turbo field echo planar imaging with deep learning constrained Compressed SENSE

Motivation: Whole-heart-coronary-MRA (WHC-MRA) is typically performed during free-breathing, and respiratory motion is prospectively compensated by using a 1D-right-diaphragmatic-navigator. However, this method requires long acquisition time and causes a heavy burden on the patient. Goal(s): Our goal was to obtain high-resolution WHC-MRA with single-breath-hold while ensuring clinically acceptable signal-to-ratio. Approach: Turbo-field-echo-planar-imaging (TFEPI), a high-speed technique, was combined with deep learning constrained Compressed SENSE (CS-AI) and image quality was evaluated by visual evaluation and contrast-to-ratio. Results: TFEPI with CS-AI provided high spatial resolution isotoropic single-breath-hold WHC-MRA while maintaining clinically acceptable image quality and scan time compared to Compressed SENSE and SENSE. Impact: High-resolution single-breath-hold WHC-MRA with clinically acceptable image quality by 3D TFEPI with CS-AI may improve the throughput of cardiac MRI examinations, and in addition reduce patient burden.