Free‐Breathing Fat Quantification Using a Phase Error‐Corrected Cartesian Acquisition With Spiral Profile Ordering

PURPOSE: To develop a phase-corrected time-interleaved multi-echo gradient echo Cartesian acquisition with spiral profile ordering (TIMGRECASPR) for abdominal large-FOV proton density fat fraction (PDFF) mapping at 3 T, demonstrating its sampling flexibility and inherent self-gating capabilities at high isotropic resolutions. METHODS: A methodological framework was developed to correct eddy current- and concomitant gradient-related phase errors in a TIMGRECASPR acquisition. Sequence parameters were optimized with an XCAT simulation. Validation was conducted in phantoms and in vivo thigh and abdominal imaging. In five healthy volunteers, liver PDFF and T2*, alongside pancreatic PDFF, were compared to a clinical Cartesian breath-hold vendor technique. Liver PDFF homogeneity was evaluated visually and across five different axial subvolumes. Mean absolute PDFF bias and within-subject standard deviation (wSD) across repeated scans were quantified. Pancreatic delineation was quantified by line profiles across the tissue boundary. RESULTS: Short shot lengths were required for a sufficient temporal resolution of the respiratory motion self-navigation. Without phase corrections, data showed severe peripheral PDFF bias, especially at higher spatial resolutions. In a water phantom, PDFF bias ranged from -5% to 5% over 400 mm in readout direction, successfully mitigated by the proposed phase corrections. Liver PDFF maps exhibited improved homogeneity and accurate quantification results, verified by the breath-hold reference scan (0.6% PDFF bias, wSD = 0.11%). Pancreatic boundaries were clearly visualized and provided accurate pancreatic PDFF maps (1.1% PDFF bias, wSD = 0.37%). CONCLUSION: This work demonstrates the feasibility of a free-breathing monopolar time-interleaved gradient echo sequence based on a Cartesian spiral acquisition at 3 T when accounting for phase errors.