Direct comparison between apparent diffusion coefficient and macromolecular proton fraction as quantitative biomarkers of the human fetal brain maturation

Abstract

Background: Apparent diffusion coefficient (ADC) is known as a quantitative biomarker of prenatal brain maturation. Fast macromolecular proton fraction (MPF) mapping is an emerging method for quantitative assessment of myelination that was recently adapted to fetal MRI.

Purpose: To compare the capability of ADC and MPF to quantify the normal fetal brain development.

Study type: Prospective.

Population: Forty-two human fetuses in utero (gestational age [GA] = 27.7 ± 6.0, range 20-38 weeks).

Field strength/sequence: 1.5 T; diffusion-weighted single-shot echo-planar spin-echo with five b-values for ADC mapping; spoiled multishot echo-planar gradient-echo with T₁ , proton density, and magnetization transfer contrast weightings for single-point MPF mapping.

Assessment: Two operators measured ADC and MPF in the medulla, pons, cerebellum, thalamus, and frontal, occipital, and temporal cerebral white matter (WM).

Statistical tests: Mixed repeated-measures analysis of variance (ANOVA) with the factors of pregnancy trimester and brain structure; Pearson correlation coefficient (r); Hotelling-Williams test to compare strengths of correlations.

Results: From the 2^nd to 3^rd trimester, ADC significantly decreased in the thalamus and cerebellum (P < 0.005). MPF significantly increased in the medulla, pons, thalamus, and cerebellum (P < 0.005). Cerebral WM had significantly higher ADC and lower MPF compared with the medulla and pons in both trimesters. MPF (r range 0.83, 0.89, P < 0.001) and ADC (r range -0.43, -0.75, P ≤ 0.004) significantly correlated with GA and each other (r range -0.32, -0.60, P ≤ 0.04) in the medulla, pons, thalamus, and cerebellum. No significant correlations or distinctions between regions and trimesters were observed for cerebral WM (P range 0.1-0.75). Correlations with GA were significantly stronger for MPF compared with ADC in the medulla, pons, and cerebellum (Hotelling-Williams test, P < 0.003) and similar in the thalamus. Structure-averaged MPF and ADC values strongly correlated (r = 0.95, P < 0.001).

Data conclusion: MPF and ADC demonstrated qualitatively similar but quantitatively different spatiotemporal patterns. MPF appeared more sensitive to changes in the brain structures with prenatal onset of myelination.

Level of evidence: 2 Technical Efficacy Stage: 2 J. Magn. Reson. Imaging 2019;50:52-61.

Keywords: apparent diffusion coefficient; brain maturation; fetal MRI; macromolecular proton fraction; myelin.

FIGURE 1:

Representative brain MPF and ADC maps obtained from the fetuses with GA of 20 (a) and 35 (b) weeks with the scheme of ROI placement. MPF maps (top rows) are presented with the grayscale range 0–10%. ADC maps (bottom rows) are presented with the grayscale range 0–2800×10⁻⁶ mm/s. ROI labels correspond to the medulla (M), cerebellum (C), pons (P), temporal WM (TWM), thalamus (T), frontal WM (FWM), and occipital WM (OWM).

FIGURE 2:

Box-whisker plots representing the distributions of ADC (a) and MPF (b) measurements across the fetal brain anatomical structures and pregnancy trimesters. Asterisks indicate significant differences between the 2^nd and 3^rd trimesters for the same structure. Shared letters indicate no significant differences between structures within the same trimester.

FIGURE 3:

Scatterplots of correlations between ADC and GA (a), between MPF and GA (b), and between ADC and MPF (c) in the fetal brain anatomic structures. The lines depict linear regression plots.

FIGURE 4:

Scatterplots of correlations between ADC and MPF in the fetal brain assessed for the global dataset (a) and structure-averaged data (b). The lines depict linear regression plots. Blue and red colors indicate data for the 2^nd and 3^rd trimester fetuses, respectively. The black lines correspond to regression over the data for both trimesters. Structure labels in panel (b) correspond to the medulla (M), pons (P), thalamus (T), cerebellum (C), temporal WM (TWM), frontal WM (FWM), and occipital WM (OWM).