Automated Brain Segmentation to Establish Reference Ranges for Regional Brain Volumes in Normally Developing Very...

Brain volume is an important marker of early brain development in preterm infants. Automated brain segmentation offers a standardized alternative to manual measurements. The aim of this study was to establish reference ranges for regional brain volumes in normally developing very preterm infants without brain injury at term-equivalent age (TEA) and to explore associations with key neonatal factors.

This cross-sectional study included very preterm infants (<32 weeks of gestation) with structurally normal magnetic resonance imaging at TEA and normal neurodevelopmental outcomes up to 2 years. Magnetic resonance imaging scans were acquired on a 3-Tesla system and processed using Infant FreeSurfer, yielding 26 regional brain structures. Only high-quality scans, confirmed by expert consensus, were included. Associations between regional brain volumes, gestational age, head circumference, and sex were assessed using Pearson correlations and Mann-Whitney U tests.

The final cohort comprised 55 infants (24 males) with a median gestational age of 29.4 weeks (interquartile range: 27.6-31.0). Reference volumes and sex-stratified centiles are provided. Male infants had significantly larger volumes of the putamen (p = 0.031) and hippocampus (p = 0.003). Gestational age showed weak or no correlations with brain volumes. Head circumference at discharge showed moderate to strong correlations with several brain structures, whereas correlations with head circumference at birth were weaker.

This study establishes normative data on regional brain volumes in a well-defined cohort of very preterm infants without brain injury at TEA. These data provide a reference for future research and may support studies investigating early deviations from typical brain development.