Simultaneous high-resolution T2 -weighted imaging and quantitative T2 mapping at low magnetic field strengths using a multiple TE and multi-orientation acquisition approach

Abstract

Purpose: Low magnetic field systems provide an important opportunity to expand MRI to new and diverse clinical and research study populations. However, a fundamental limitation of low field strength systems is the reduced SNR compared to 1.5 or 3T, necessitating compromises in spatial resolution and imaging time. Most often, images are acquired with anisotropic voxels with low through-plane resolution, which provide acceptable image quality with reasonable scan times, but can impair visualization of subtle pathology.

Methods: Here, we describe a super-resolution approach to reconstruct high-resolution isotropic T₂ -weighted images from a series of low-resolution anisotropic images acquired in orthogonal orientations. Furthermore, acquiring each image with an incremented TE allows calculations of quantitative T₂ images without time penalty.

Results: Our approach is demonstrated via phantom and in vivo human brain imaging, with simultaneous 1.5 × 1.5 × 1.5 mm³ T₂ -weighted and quantitative T₂ maps acquired using a clinically feasible approach that combines three acquisition that require approximately 4-min each to collect. Calculated T₂ values agree with reference multiple TE measures with intraclass correlation values of 0.96 and 0.85 in phantom and in vivo measures, respectively, in line with previously reported brain T₂ values at 150 mT, 1.5T, and 3T.

Conclusion: Our multi-orientation and multi-TE approach is a time-efficient method for high-resolution T₂ -weighted images for anatomical visualization with simultaneous quantitative T₂ imaging for increased sensitivity to tissue microstructure and chemical composition.

Keywords: child brain development; low field MRI; magnetic resonance imaging; pediatric neuroimaging.

FIGURE 1

Example illustration of SR image reconstruction using images acquired in the axial, sagittal, and coronal orientations (the frequency encoding/readout and lower resolution directions are labeled) and the final reconstructed image

FIGURE 2

Analysis workflow from the acquisition of the source anisotropic T₂‐weighted data with coronal, axial, and sagittal orientations (left panel); SR reconstruction of an isotropic T₂‐weighted image from the acquired data (bottom, left panel); and calculation of qT₂ maps from the aligned and resampled multi‐orientation data (top, right panel). For comparison, qT2 maps calculated from multiple TEs acquired in a single orientation and using Hyperfine‐provided five‐TE approach (middle, right panel) are also shown. As well, we show a multi‐orientation + multi‐TI qT2 calculated from data that was preprocessed using adaptive denoising (bottom, right panel)

FIGURE 3

Comparison of qT₂ values from four brain regions of interest (top) in each healthy volunteer. Brain regions included cerebellar white matter (green), corpus callosum (light blue), anterior internal capsule (red), and posterior thalamic radiations (dark blue). The ICC between the noise filtered multi‐orientation + multi‐TE and reference Hyperfine 5‐TE qT₂ values was 0.93. Example monoexponential fits to the corpus callosum and cerebellar white matter data from one of the healthy volunteers is also shown

FIGURE 4

(Left) Comparison of qT₂ images through the multi‐element phantom and (Right) Mean phantom multi‐orientation + multi‐TE and reference Hyperfine 5‐TE qT₂ values. The solid line corresponds to the line of unity, and errors bars in the phantom measurements denote 1 SD. ICC values for the phantom data were 0.97 (with denoising) and 0.96 (without denoising)