Background: Extremely preterm infants are at risk for neurodevelopmental impairment (NDI). Early cranial ultrasound (CUS) is usual practice, but near-term brain MRI has been reported to better predict outcomes. We prospectively evaluated MRI white matter abnormality (WMA) and cerebellar lesions, and serial CUS adverse findings as predictors of outcomes at 18 to 22 months' corrected age.
Methods: Early and late CUS, and brain MRI were read by masked central readers, in a large cohort (n = 480) of infants <28 weeks' gestation surviving to near term in the Neonatal Research Network. Outcomes included NDI or death after neuroimaging, and significant gross motor impairment or death, with NDI defined as cognitive composite score <70, significant gross motor impairment, and severe hearing or visual impairment. Multivariable models evaluated the relative predictive value of neuroimaging while controlling for other factors.
Results: Of 480 infants, 15 died and 20 were lost. Increasing severity of WMA and significant cerebellar lesions on MRI were associated with adverse outcomes. Cerebellar lesions were rarely identified by CUS. In full multivariable models, both late CUS and MRI, but not early CUS, remained independently associated with NDI or death (MRI cerebellar lesions: odds ratio, 3.0 [95% confidence interval: 1.3-6.8]; late CUS: odds ratio, 9.8 [95% confidence interval: 2.8-35]), and significant gross motor impairment or death. In models that did not include late CUS, MRI moderate-severe WMA was independently associated with adverse outcomes.
Conclusions: Both late CUS and near-term MRI abnormalities were associated with outcomes, independent of early CUS and other factors, underscoring the relative prognostic value of near-term neuroimaging.
Trial registration: ClinicalTrials.gov NCT00233324.
Keywords: MRI; neurodevelopmental; neuroimaging; preterm infant; ultrasound.
Copyright © 2015 by the American Academy of Pediatrics.