Brain MRI before and at term equivalent age predicts motor and cognitive outcomes in very preterm infants

Abstract

Brain Magnetic Resonance Imaging (MRI) of high-risk infants in the neonatal period (from 26 weeks postmenstrual age to Term Equivalent Age (TEA)) is increasingly being used for the detection of brain injuries, and the early prognostication of adverse outcomes such as Cerebral Palsy (CP). While most imaging is performed around TEA in clinical practice for infants born preterm (<37 weeks of gestation), this would often require families to return to hospital for imaging. In this work, we extract structural biomarkers from MRI acquired both before and at TEA in a cohort of very preterm infants from the PPREMO and PREBO studies (n = 100), to determine if either time-point, or both combined, are predictive of both Bayley Scales of Infant and Toddler Development - Third Edition (Bayley-III) and the Neuro-sensory Motor Developmental Assessment (NSMDA) at 2 years. Using multivariable regression, moderately strong and statistically significant associations were found between brain structure on both early and TEA MRIs with 2-year outcomes (r = 0.39-0.55 for early MRI, r = 0.37-0.49 for Term MRI, r = 0.37-0.56 for early and TEA MRI combined). Importantly, brain biomarkers associated with early childhood outcomes from MRIs were identified, including white and grey matter volumes, deep grey matter and cerebellar volumes, and gyrification and surface area across the whole cortex. Early MRI showed the best prognostic accuracy along with combining timepoints, indicating the potential clinical benefit of Early MRI in predicting adverse outcomes.