Automatic Respiratory and Bulk Patient Motion Corrected 3D Fetal MRI.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Radiology

Background:

• Fetal MRI is crucial for prenatal diagnosis but is often degraded by motion artifacts.
• Motion, primarily from maternal respiration and bulk movement, limits the diagnostic accuracy of 3D radial fetal MRI.
• Developing effective motion correction techniques is essential for improving fetal MRI quality.

Purpose of the Study:

• To develop and validate a novel framework, ACROBATIC, for correcting motion in 3D radial fetal MRI.
• To assess the efficacy of ACROBATIC in reducing motion-related artifacts and enhancing image quality.
• To compare the performance of ACROBATIC against uncorrected data and establish its potential clinical utility.

Main Methods:

• ACROBATIC utilizes self-gating signals and focused navigation to estimate respiratory motion.
• Bulk motion is quantified by reconstructing 3D volumes in sequential bins and applying rigid image registration.
• Combined motion estimates are applied to k-space for correction, with validation through simulations and in utero data acquisition.

Main Results:

• Self-gating signals demonstrated strong correlation with ground truth in simulations (R=0.97) and external sensors in utero (R=0.75).
• ACROBATIC significantly reduced cumulative position error from 13.33 to 2.20.
• Image sharpness improved substantially with ACROBATIC, corroborated by expert reviewer rankings favoring corrected images.

Conclusions:

• The ACROBATIC framework effectively corrects motion in 3D radial fetal MRI.
• Motion correction improves image quality in both simulated and in utero fetal MRI datasets.
• Further studies comparing ACROBATIC to existing 2D frameworks are recommended to determine its added diagnostic value.