HERON: High-Efficiency Real-Time Motion Quantification and Re-Acquisition for Fetal Diffusion MRI.

Area of Science:

• Medical Imaging
• Neuroscience
• Artificial Intelligence

Background:

• Fetal diffusion MRI (dMRI) is vital for studying brain development but is susceptible to motion artifacts from fetal and maternal movements.
• These artifacts degrade image quality, limiting diagnostic accuracy in clinical and research settings.

Purpose of the Study:

• To introduce HERON, an automated pipeline for real-time motion assessment and re-acquisition to improve fetal brain dMRI quality.
• To evaluate HERON's effectiveness in enhancing image quality, reducing motion, and enabling reliable quantitative analysis.

Main Methods:

• Development of HERON, an AI-driven pipeline for brain localization, segmentation, and motion assessment on a 0.55T scanner.
• Real-time automated planning, quality checking, and re-acquisition of motion-affected dMRI volumes.
• Post-processing correction for residual inter-volume motion.

Main Results:

• HERON significantly improved image quality and reduced both intra- and inter-volume motion in 20 tested cases.
• The pipeline demonstrated excellent agreement with human observers (specificity 97%, sensitivity 92%) in motion assessment.
• Quantitative analysis showed a reduction in mean Apparent Diffusion Coefficient and Intravoxel Incoherent Motion after correction.

Conclusions:

• The automated, AI-driven HERON pipeline effectively enhances fetal brain dMRI quality.
• This improvement facilitates more reliable quantitative analysis, supporting wider research and clinical applications.