Multiresolution comparison of fetal real-time and cine magnetic resonance imaging at 0.55T

ABSTRACT: Background: Dynamic fetal cardiovascular MRI (CMR) enables visualization of moving structures to assess congenital heart disease and plan treatment. Low field MRI systems can provide more comfortable platforms for fetal CMR. Here, we demonstrate the feasibility and utility of motion corre...

Full description

Saved in:
Bibliographic Details
Main Authors: Datta Singh Goolaub, Ye Tian, Joshua F.P. van Amerom, John Wood, Jon Detterich, Krishna S. Nayak, Christopher K. Macgowan
Format: Article
Language:English
Published: Elsevier 2025-01-01
Series:Journal of Cardiovascular Magnetic Resonance
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S1097664725000183
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:ABSTRACT: Background: Dynamic fetal cardiovascular MRI (CMR) enables visualization of moving structures to assess congenital heart disease and plan treatment. Low field MRI systems can provide more comfortable platforms for fetal CMR. Here, we demonstrate the feasibility and utility of motion corrected fetal cardiac cine CMR and compare it with real-time CMR at multiple spatial resolutions at 0.55 T. Methods: Ten human pregnancies were scanned at 0.55T on a derated MAGNETOM Aera (Siemens Healthineers, Erlangen, Germany) with spiral steady-state free precession imaging. Real-time images were reconstructed and used for motion correction and fetal cardiac gating followed by cine reconstructions. The signal-to-noise ratio (SNR), image quality, blood-to-myocardium contrast, and contrast-to-noise ratio (CNR) from real-time and cine reconstructions were compared. The effect of acceleration on cine accuracy was assessed by retrospectively undersampling the data and measuring the reconstruction error with the normalized root-mean-squared difference (NRMSD) in five fetuses. Reproducibility of the measurements was assessed by reconstructing cines from two independent windows of data and computing the NRMSD relative to the reference image in five fetuses. Results: The SNR, CNR, and image quality were better for cines than their corresponding real-time reconstructions. The blood-to-myocardium contrast had no significant difference between real-time and cine reconstructions. With finer spatial resolution, real-time images degraded, and cardiac structures were less conspicuous. NRMSD in cines decreased with increasing scan times across all resolutions (NRMSD = 10 ± 2% for 7 s scan duration). Good consistency (NRMSD = 11 ± 3%) was achieved between independent reconstruction windows. Conclusion: While this study was performed on an experimental scanner (derated; not commercially available), we have shown that fetal cine CMR is feasible at 0.55T and provides high-quality fetal cardiac images at high spatiotemporal resolutions.
ISSN:1097-6647