Peak Width of Skeletonized Water Diffusion MRI in the Neonatal Brain.

Manuel Blesa,James P Boardman,Alan J Quigley,Emily N Wheater,Gemma Sullivan,Mark E Bastin,David Q Stoye,Michael J Thrippleton,Paola Galdi,Gillian J Lamb

doi:10.3389/fneur.2020.00235

Abstract

Preterm birth is closely associated with cognitive impairment and generalized dysconnectivity of neural networks inferred from water diffusion MRI (dMRI) metrics. Peak width of skeletonized mean diffusivity (PSMD) is a metric derived from histogram analysis of mean diffusivity across the white matter skeleton, and it is a useful biomarker of generalized dysconnectivity and cognition in adulthood. We calculated PSMD and five other histogram based metrics derived from diffusion tensor imaging (DTI) and neurite orientation and dispersion imaging (NODDI) in the newborn, and evaluated their accuracy as biomarkers of microstructural brain white matter alterations associated with preterm birth. One hundred and thirty five neonates (76 preterm, 59 term) underwent 3T MRI at term equivalent age. There were group differences in peak width of skeletonized mean, axial, and radial diffusivities (PSMD, PSAD, PSRD), orientation dispersion index (PSODI) and neurite dispersion index (PSNDI), all p < 10−4. PSFA did not differ between groups. PSNDI was the best classifier of gestational age at birth with an accuracy of 81±10%, followed by PSMD, which had 77±9% accuracy. Models built on both NODDI metrics, and on all dMRI metrics combined, did not outperform the model based on PSNDI alone. We conclude that histogram based analyses of DTI and NODDI parameters are promising new image markers for investigating diffuse changes in brain connectivity in early life.

Highlights

Preterm birth is closely associated with a phenotype that includes cognitive impairment in childhood and cerebral white matter disease
White matter disease is apparent as diffuse changes in signal intensity on conventional MRI [1, 2], and alterations in diffusion MRI parameters based on the diffusion tensor [fractional anisotropy (FA), and mean, axial, and radial diffusivities, (MD, AD, RD)], and more recently, metrics based on biophysical models, such as neurite orientation and dispersion imaging (NODDI) [3, 4]
Deriving whole brain estimations of these parameters is often computationally expensive, there are uncertainties about which metric or combination of metrics best captures generalized white matter disease associated with preterm birth, and which is likely to be most useful for prognosis

Summary

Introduction

Preterm birth is closely associated with a phenotype that includes cognitive impairment in childhood and cerebral white matter disease. White matter disease is apparent as diffuse changes in signal intensity on conventional MRI [1, 2], and alterations in diffusion MRI parameters based on the diffusion tensor [fractional anisotropy (FA), and mean, axial, and radial diffusivities, (MD, AD, RD)], and more recently, metrics based on biophysical models, such as neurite orientation and dispersion imaging (NODDI) [3, 4] These metrics have proven useful for making inferences about microstructural alteration of white matter that characterizes dysmaturity associated with preterm birth, for investigating upstream pathways to typical / atypical brain development, and for studying the anatomical bases of subsequent cognitive function in early life [5,6,7,8,9,10,11,12,13]. The framework is readily extensible to other DTI metrics (FA, RD, and AD) and to NODDI metrics

Results

Discussion

Conclusion