Geometric Reliability of Super-Resolution Reconstructed Images from Clinical Fetal MRI in the Second Trimester

Area of Science:

• Medical Imaging
• Neuroscience
• Radiology

Background:

• Fetal Magnetic Resonance Imaging (MRI) is crucial for noninvasive central nervous system (CNS) development assessment.
• Current clinical practice relies on manual biometric measurements from 2D fetal brain MRI sequences.
• Super-Resolution (SR) reconstruction enables 3D analysis from 2D images, offering potential improvements.

Purpose of the Study:

• To evaluate the reliability of SR reconstructed fetal brain volumes for biometric assessments.
• To compare different reconstruction toolkits (NiftyMIC, MIALSRTK, SVRTK) and MRI sequences (T2w TSE, b-FFE).
• To assess the impact of SR reconstruction on the accuracy and reproducibility of fetal biometry.

Main Methods:

• Analysis of 17 second-trimester fetal MR exams with T2w TSE and b-FFE sequences.
• Reconstruction of high-resolution 3D volumes using NiftyMIC, MIALSRTK, and SVRTK.
• Comparison of 15 biometric measurements from 2D images and SR volumes using statistical methods.

Main Results:

• NiftyMIC and MIALSRTK demonstrated reliable SR volumes suitable for fetal brain biometry.
• NiftyMIC improved operator intraclass correlation for quantitative biometric measures.
• T2w TSE sequences yielded more robust reconstructions against artifacts than b-FFE sequences.

Conclusions:

• Automatic toolkits like NiftyMIC and MIALSRTK are valuable for fetal brain reconstruction and biometry.
• SR reconstruction enhances the evaluation of fetal brain development in early pregnancy.
• T2w TSE sequences are recommended for more robust fetal brain reconstructions.