Neuropediatrics 2005; 36(2): 78-89
DOI: 10.1055/s-2005-837574
Original Article

Georg Thieme Verlag KG Stuttgart · New York

Predictive Value of Neonatal MRI with Respect to Late MRI Findings and Clinical Outcome. A Study in Infants with Periventricular Densities on Neonatal Ultrasound

L. T. L. Sie1 , 6 , A. A. M. Hart5 , J. van Hof4 , L. de Groot4 , W. Lems4 , H. N. Lafeber2 , J. Valk3 , M. S. van der Knaap1
  • 1Department of Child Neurology, VU University Medical Center, Amsterdam, The Netherlands
  • 2Department of Neonatology, VU University Medical Center, Amsterdam, The Netherlands
  • 3Department of Radiology, VU University Medical Center, Amsterdam, The Netherlands
  • 4Department of Movement Sciences, Vrije Universiteit, Amsterdam, The Netherlands
  • 5Department of Clinical Epidemiology and Biostatistics, Academic Medical Center, Amsterdam, The Netherlands
  • 6Present address: Department of Child Neurology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands
Further Information

Publication History

Received: April 16, 2003

Accepted after Revision: January 29, 2005

Publication Date:
18 March 2005 (online)

Zoom Image

Abstract

Purpose: The aim of this study was to correlate hypoxic-ischemic white matter damage on neonatal MRI with MRI appearance and neurological outcome at the age of 1œ years. Patients and Methods: A sequential cohort of infants with periventricular densities on neonatal ultrasound was studied with neonatal MRI. Images of 46 infants with a mean gestational age of 31 weeks were obtained at a mean age of 20 days after birth and at 1œ years. To establish agreement between the neonatal and follow-up MRI (general, motor, and visual scores), the weighted Cohen's kappa test was used. To establish the predictive power of neonatal MRI with respect to the neurologic indices at the age of 1œ years, the sensitivity, specificity, and positive and negative predictive values were calculated. Results: There was a moderately good to good agreement between the general, motor, and visual neonatal and follow-up MRI scores: weighted kappa = 0.59 (95 % CI: 0.44 - 0.74), 0.82 (95 % CI: 0.72 - 0.93), and 0.70 (95 % CI: 0.56 - 0.84), respectively. Neonatal MRI scores provided a good prediction of the three neurological outcome measures (developmental delay, cerebral palsy, and cerebral visual impairment): sensitivity, specificity, and predictive values were high, with little difference between the three MRI scores. The 32 patients with (nearly) normal neonatal MRI scores were neurologically (nearly) normal at 1œ years on all three outcome measures, whereas 8 patients with seriously abnormal neonatal MRI scores were neurologically abnormal at 1œ years on all three outcome measures. Conclusion: Neonatal MRI is able to predict the precise localization and size of perinatal leukomalacia on follow-up MRI and provides a good prediction of neurological outcome at 1œ years.