Highly accelerated cardiac cine acquisition with deep learning reconstruction and protocol optimization at 0.5T

Motivation: To Study Cardiac Cine FIESTA with high acceleration factors and Deep learning reconstruction at 0.5T Goal(s): Functional assessment of ventricles using highly accelerated cine fiesta acquired in 3 Heart beats at 0.5T field strength. Approach: DLCine uses a variable density sampling scheme and deep learning reconstruction algorithm to accelerate FIESTA Cine scans. This data was compared to the conventional cine fiesta accelerated by parallel imaging. Results: We observed DL cine method has 2.4 times higher SNR compared to the conventional acceleration method such as parallel imaging and scan time is reduced by 70% Impact: Acquisition of cardiac short axis scan in 3 heart beats using DL Cine for ventricular functional assessment at 0.5T