Prematurity is associated with white matter T2 MRI visible perivascular spaces in very preterm-born neonates

Perivascular spaces (PVSs), as biomarkers of the brain's neurofluid clearance system, remain largely unexplored in the brains of very preterm and term-born infants. In this retrospective cross-sectional study, we investigate PVSs in very preterm and term-born neonates and explore potential associations with preterm birth, maturation, brain injury and developmental outcome. T2-weighted fast spin echo MRI data of 86 very preterm (<32 gestational weeks) and 43 term-born neonates were acquired at term-equivalent age (mean postmenstrual age = 41.63 ± 2.09 weeks) with a 3T GE HD.xt scanner, using an 8-channel head coil, with echo time/repetition time = 109/5700 ms, field of view = 25.6 cm, acquisition matrix = 256 × 256, reconstruction matrix = 512 × 512 and slice thickness = 2 mm. PVS counts were estimated visually in the basal ganglia and centrum semiovale (CSO) using a validated scoring system. Developmental outcome was evaluated using the Bayley Scales of Infant Development II or III. White and grey matter injuries were evaluated using Woodward's score. Groupwise differences in PVS counts between term-born and very preterm-born neonates and associations with postmenstrual age, preterm birth, brain injury and developmental outcomes were evaluated with Mann-Whitney tests and multiple regression. PVS counts in the basal ganglia did not differ between groups, whereas PVS counts in the CSO were significantly higher in very preterm-born neonates (median = 1, interquartile range: 0-2), compared with term-born neonates (median = 0, interquartile range: 0-0; P < 0.001). CSO PVSs were independently associated with postmenstrual age, incidence rate ratio = 1.44, confidence interval 1.23-1.70), P < 0.001, and preterm birth, incidence rate ratio = 44.87, confidence interval 10.39-212.98, P < 0.001. Including a postmenstrual age-by-group interaction showed that the maturation-related increase of CSO PVSs did not differ between the groups. We found no evidence for an association of PVSs with brain injury (grey matter injury: P = 0.75 for basal ganglia and P = 0.84 for CSO; white matter injury: P = 0.92 for basal ganglia and P = 0.60 for CSO) or developmental outcome (P = 0.31 for basal ganglia, P = 0.24 for CSO). These results demonstrate an association of CSO PVSs with preterm birth and an increase of CSO PVSs with maturation. Longitudinal studies may shed further light on the role of preterm birth on the brain's neurofluid clearance system development.