Liver Fat Quantification in Infants using Free-Breathing 3D Stack-of-Radial MRI and Phase-Corrected Non-Rigid Motion Compensation

Motivation: Events during fetal life and early childhood can contribute to the development of pediatric obesity and metabolic syndrome. Currently, there is not a well-established research tool to study fat distribution in non-sedated infants. Goal(s): To develop a non-rigid motion compensation reconstruction technique for quantification of proton density fat fraction (PDFF) in infants. Approach: We incorporated deformation vector field information to a compressed sensing model for non-rigid motion. Image sharpness and PDFF measurement in the liver from different reconstruction methods were compared. Results: Non-rigid motion compensation compressed sensing reconstruction can reduce the artifacts, improve the image sharpness, leading to more reliable PDFF maps. Impact: This technique has the potential to serve as a useful research tool for investigating liver fat distribution or body composition in infants without sedation.