Fetal 4D flow bulk motion estimation – k-space versus image-based surrogates

Motivation: As motion is inevitable during longer fetal 4D flow acquisitions, detecting and correcting sporadic bulk motion is necessary to ensure adequate image quality. Goal(s): To develop automated k-space-based motion navigation for maternal respiration and fetal bulk motion, and compare it with an image-based surrogate for bulk motion. Approach: Maternal respiration and fetal bulk motion are simultaneously extracted through principal component analysis of all k0 readouts and edge filtering for detection of bulk motion events. Results: K-space-based detection and correction of fetal bulk motion is feasible. No significant differences were found in image quality or quantitative analysis across reconstruction types. Impact: Automating detection of fetal bulk motion over longer (e.g. 4D flow) scans allows for more robust motion rejection.