Parenchymal brain injury in the preterm infant: Comparison of cranial ultrasound, MRI and neurodevelopmental outcome

Aim: Magnetic resonance imaging (MRI) is increasingly being used in high-ri sk preterm neonates. Cranial ultrasound (US) was compared with MRI in prete rm patients with parenchymal injury and related to neurodevelopmental outco me. Patients and Methods: Studies were performed in 61 patients. Twelve inf ants with normal US (Group 1) had an MRI within the first 4 weeks of life ( early MRI), and 10 also at term age (late MRI). Eight out of 20 infants wit h intraventricular haemorrhage with parenchymal involvement (IVH + PI) (Gro up 2) had an early as well as a late MRI and 12 a late MRI. Of the 20 patie nts with TI cystic-periventricular leukomalacia (c-PVL) (Group 3), 7 had an TR early MRI, 7 had an MRI on both occasions and 12 had a late US MRI. All 9 children with focal infarction (FI) (Croup 4) had a late MRI. Results: M RI was conform with cranial US in Group 1. Early MRI in Group Z showed cont ralateral c-PVL in one infant and an additional contralateral occipital par enchymal haemorrhage and blood in the posterior fossa in another infant. La te MRI showed an asymmetrical posterior limb of the internal capsule (PLIC) (n = 6), which predicted later hemiplegia. Early MRI in Group 3 showed mor e cysts (n = 5), punctate white matter lesions (n = 6), lesions in the basa l ganglia (n = 1) and once involvement of the cerebellum. Late MRI showed i nvolvement of the centrum semiovale (n = 2) lesions in the basal ganglia (n = 2) and bilateral abnormal signal intensity of the PLIC in 7 infants who all went on to develop cerebral palsy. In Group 4 MRI showed signal intensi ty changes suggestive of cystic lesions compared to persisting echogenicity on US (n = 3) and an asymmetrical PLIC (n = 5), which predicted hemiplegia in 4. Conclusion: Early MRI especially provided additional information in those with c-PVL. MRI at term age could assess the PLIC, which was useful i n children with unilateral parenchymal involvement, for prediction of subse quent hemiplegia and, to a lesser degree, in bilateral c-PVL for prediction of diplegia or quadriplegia.