Quantifying Fetal Periventricular White Matter Development Using Multimodal MRI

Helena S. Sousa, Abi Fukami-Gartner, Alena Uus, Sian Wilson, Vanessa Kyriakopoulou, Jana Hutter, Lisa Story, Megan Hall, Grainne M. McAlonan, Joseph V. Hajnal, Mary A. Rutherford, Alexander Hammers, Jacques-Donald Tournier, Maria Deprez

Abstract: The periventricular white matter (PVWM) in the developing brain is of high clinical relevance due to its vulnerability to injury during fetal and neonatal life, particularly in preterm infants, where damage in these regions can disrupt early brain connectivity and lead to important clinical consequences in postnatal life. Despite its importance, there are currently no standardised methods for delineating the PVWM in fetal MRI. In this study, we address this gap by characterising the normative anatomical and microstructural development of the PVWM across 21–36 weeks of gestational age (GA) using multimodal MRI data. We define the PVWM in a 0.5 mm isotropic T2-weighted fetal brain atlas from the developing Human Connectome Project (dHCP) and train an Attention U-Net model for automated PVWM segmentation. Using MRI data from multiple fetal cohorts, we analyse developmental trends in PVWM macro- and microstructure and discuss possible correlations with the underlying neurobiological processes occurring in the fetal brain. Our findings establish a normative reference framework for in utero PVWM maturation, providing a foundation for identifying atypical developmental patterns.

External IDs:dblp:conf/miccai/SousaFUWKHSHMHRHTD25