High resolution and contrast 7 tesla MR brain imaging of the neonate

Abstract

Introduction: Ultra-high field MR imaging offers marked gains in signal-to-noise ratio, spatial resolution, and contrast which translate to improved pathological and anatomical sensitivity. These benefits are particularly relevant for the neonatal brain which is rapidly developing and sensitive to injury. However, experience of imaging neonates at 7T has been limited due to regulatory, safety, and practical considerations. We aimed to establish a program for safely acquiring high resolution and contrast brain images from neonates on a 7T system.

Methods: Images were acquired from 35 neonates on 44 occasions (median age 39 + 6 postmenstrual weeks, range 33 + 4 to 52 + 6; median body weight 2.93 kg, range 1.57 to 5.3 kg) over a median time of 49 mins 30 s. Peripheral body temperature and physiological measures were recorded throughout scanning. Acquired sequences included T2 weighted (TSE), Actual Flip angle Imaging (AFI), functional MRI (BOLD EPI), susceptibility weighted imaging (SWI), and MR spectroscopy (STEAM).

Results: There was no significant difference between temperature before and after scanning (p = 0.76) and image quality assessment compared favorably to state-of-the-art 3T acquisitions. Anatomical imaging demonstrated excellent sensitivity to structures which are typically hard to visualize at lower field strengths including the hippocampus, cerebellum, and vasculature. Images were also acquired with contrast mechanisms which are enhanced at ultra-high field including susceptibility weighted imaging, functional MRI, and MR spectroscopy.

Discussion: We demonstrate safety and feasibility of imaging vulnerable neonates at ultra-high field and highlight the untapped potential for providing important new insights into brain development and pathological processes during this critical phase of early life.

Keywords: brain; infant; magnetic resonance imaging (MRI); neonate; neuroradiology; ultra-high field MRI.

Figure 1

Image quality assessment. (A) Shown are representative images depicting quality of the native images as assessed using a previously described score (20), where (1) is a poor quality image (2) image contains significant motion artefact, (3) image with negligible motion artefact and (4) good quality image. (B) Image quality assessment results for the native T2-weighted images and (C) for the slice-to-volume reconstructed T2-weighted images.

Figure 2

Infant temperature during scanning on the 7T system. There was no significant difference between infant axillary temperature at the start (left) and end (right) of the imaging session on the 7T system (paired two tailed t-test: p = 0.76). Box and whisker plots showing data median (bold line), 25th and 75th centile (box borders) and data range (whiskers). Data outliers are denoted by circles.

Figure 3

B₁⁺ maps from 15 infants. (A) Example B₁⁺ map obtained with AFI sequence, shown in relative units (i.e., relative to nominal B₁⁺). The white contour marks the approximate brain outline. (B) Histograms of relative B₁⁺ magnitude within the brain mask in 15 subjects. The yellow box shows the range of median relative B₁⁺ (0.69−0.78).

Figure 4

High resolution T2 weighted 7T image acquired from a preterm infant. Images show excellent visualisation of the: (A) cerebellar vermis and folia (red arrows); (B) deep gray matter nuclei within the basal ganglia and thalamus (blue arrows); (C) hippocampus (yellow arrow).

Figure 5

Examples of intracranial pathology identified on T2 weighted images at 7T. (from left to right: illustrative slices in the sagittal, coronal, and axial planes). Shown are: (A) Cystic PVL; (B) Complete agenesis of the corpus callosum.

Figure 6

Slice to volume reconstructed T2 weighted images acquired from a preterm infant shown in the (A) sagittal, (B) coronal and (C) axial planes. These high contrast, high resolution images are amenable for further processing such as the generation of (D) (from left to right) inflated, pial, and white matter surfaces.

Figure 7

Susceptibility weighted images (SWI). (A) Example MIP axial slices from the SWI data acquired from a healthy neonate imaged at term equivalent age. (B) A cystic lesion (blue arrow) noted on T2-weighted images (bottom left) in a preterm infant. SWI demonstrated the hemorrhagic origin of the lesion (red arrow) and gives possible insight into the underlying pathophysiology through its adjacent location to the deep medullary veins. (C) Preterm infant with extensive intracerebral hemorrhage in the distribution of the medullary veins. The extent of this is visualized more clearly on the SWI image (bottom right).

Figure 8

Resting state functional MRI data acquired from a preterm infant at 7T. (A) High spatial resolution (1 mm isotropic) high contrast whole brain BOLD fMRI data acquired at 7T from a preterm infant. (B) The sensorimotor resting state network derived using independent component analysis. Activation can be seen to clearly localize to the cortical ribbon, following the configuration of the sulci.

Figure 9

STEAM MRS spectrum acquired from the left thalamus of an infant at term equivalent age on the 7T system. A narrow linewidth (4.8Hz) and high SNR (26) as estimated by LCModel are observed. The acquisition time was just 2’24”.