T1 and T2 measurements of the neonatal brain at 7 T

Abstract

Purpose: To determine the expected range of NMR relaxation times (T₁ and T₂) in the neonatal brain at 7 T.

Methods: Data were acquired in a total of 40 examinations on infants in natural sleep. The cohort included 34 unique subjects with postmenstrual age range between 33 and 52 weeks and contained a mix of healthy individuals and those with clinical concerns. Single-slice T₁ and T₂ mapping protocols were used to provide measurements in white matter, cortex, cerebellum, and deep gray matter. Automatic image segmentation of a separate T₂-weighted brain volume was used to define regions of interest for analysis.

Results: Linear regression was used to estimate relaxation times at term equivalent age (40 weeks postmenstrual age). $T_1^{40\mathrm{wk}}$ with 95% confidence intervals was measured to be 2933 [2893, 2972] ms in white matter; 2653 [2604, 2701] ms in cerebellum; and 2486 [2439, 2532] ms in basal ganglia. $T_2^{40\mathrm{wk}}$ was estimated as 119 [116, 121] ms in white matter, 99 [96, 102] ms in cerebellum, and 90 [89, 92] ms in basal ganglia. Most tissue-relaxation times showed a significant negative correlation with postmenstrual age, with the strongest correlation seen in cerebellum.

Conclusions: We describe neonatal brain tissue and age-specific T₁ and T₂ relaxation values at 7 T. The presented values differ substantially from both adult values at 7 T and neonate values measured at lower field strengths, and will be essential for pulse-sequence optimization for neonatal studies.

Keywords: neonatal imaging; relaxation times; ultrahigh field.

FIGURE 1

Top row: Example of planning single oblique slice for relaxometry. Left: Localizer image showing the planned slice (yellow) in three planes (green box is the volume used for shimming). Bottom row: Left: example T₂‐weighted image from this oblique slice. Right = T₂w image overlaid with example image segmentation from this subject, used for region of interest analysis (details of segmentation methods are given in section 2.5). GM, gray matter; PVFWM, periventricular frontal white matter.

FIGURE 2

Results from T₁ mapping on 1 subject (id#23; see Table S1). (A) Fitted T₁ map. (B) Inversion inefficiency ($ϵ$) parameter. (C) Equivalent slice from relative B₁ map.

FIGURE 3

T₂ estimation results from the same infant featured in Figure 2 (id#23). (A) T₂ estimated using dictionary with fixed T₁ = 2.6 s. (B) Dictionary estimation including voxel‐wise T₁ information. (C) Difference between these two estimates.

FIGURE 4

Region‐of‐interest measurements of relaxation parameters (regions of interest as defined by segmentation, illustrated in Figure 1). Top panels of (A) plot T₁ and bottom panels of (B) give equivalent results for T₂. Both parameters are plotted against postmenstrual age (PMA), with color representing postnatal age (PNA = PMA‐GA). Linear trend line is superimposed (solid line) with 95% confidence interval (shading). The number of observations used for each regression (N) is indicated on each plot. N varies between anatomical regions, because some are not detected in the slice used for imaging in some subjects; there are also fewer subjects for T₂ measurements due to missing $B_1^{\mathit{rel}}$. The samples indicated with triangles were excluded from regression (details in text). GA, gestational age; GM, gray matter; PNA, postnatal age; PVFWM, periventricular frontal white matter.