Assessing Early Brain Development in Neonates by Segmentation of HighResolution 3T MRI

1
Assessing Early Brain Development in Neonates by
Segmentation of High-Resolution 3T MRI 1,2G
Gerig, 2M Prastawa, 3W Lin, 1John Gilmore
Departments of 1Psychiatry, 2Computer Science,
3Radiology University of North Carolina, Chapel
Hill,NC 27614, USA gerig_at_cs.unc.edu /
http//www.cs.unc.edu/gerig
RESULTS
SUMMARY
T1-only segmentation

• Research Quantitative MRI to study unsedated
  newborns at risk for neurodevelopmental
  disorders.
• Clinical Study 120 newborns recruited at UNC,
  age at MRI about 2 weeks
• Motivation Early detection of abnormalities ?
  Possibility for early intervention and therapy.
• Imaging High field (3T Siemens Allegra), high
  resolution (T1 1mm3, FSE 0.9x0.9x3mm3),
  high-speed imaging (12 for T1, FSE and DTI).

hyper- intense motor cortex
early myelinated corticospinal tract
Neonate
Adult

• Challenge Very low CNR, heterogeneous tissue,
  early myelination regions, reverse contrast
  wm/gm.
• Standard brain tissue segmentation fails.

PD/T2 segmentation
T1 3D MPRage 1x1x1 mm3
FSE T2w 1x1x3 mm3
FSE PDw 1x1x3 mm3
METHODS

• Approach
• Atlas-moderated EM segmentation (cf. Leemput and
  Warfield)
• Tissue intensity model for white matter
  (non-myelinated and myelinated wm form bimodal
  distribution) (cf. Cocosco, Prastawa)

High-resolution PD/T2 data courtesy of Petra
Hueppi, Univ. of Geneva.
Preliminary Results UNC Neonate Study

• So far 20 normal neonates (10 males, 10 females)
• Age 16 4 days
• Siemens 3T head-only scanner
• Neonates were fed prior to scanning, swaddled,
  fitted with ear protection and had their heads
  fixed in a vac-fix device
• A pulse oximeter was monitored by a physician or
  research nurse
• Most neonates slept during the scan
• Motion-free scans in 13-15 infants

Building of Atlas Template
CONCLUSIONS

• It is feasible to study brain development in
  unsedated newborns using 3T MRI
• Study will likely provide a vastly improved
  understanding of early brain development and its
  relationship to neuropsychiatric disorders.
• Novelty Tissue model for segmentation of
  myelinated/nonmyel. white matter.

Literature

• Gilmore JH, Gerig G, Specter B, Charles HC,
  Wilber JS, Hertzberg BS, Kliewer MA (2001a)
  Neonatal cerebral ventricle volume a comparison
  of 3D ultrasound and magnetic resonance imaging.
  Ultrasound Med and Biol 271143-1146.
• Huppi PS, Warfield S, Kikinis R, Barnes PD,
  Zientara GP, Jolesz FA, Tsuji MK, Volpe JJ
  (1998b) Quantitative magnetic resonance imaging
  of brain development in premature and normal
  newborns. Ann Neurol 43 224-235.
• Zhai G, Lin W, Wilber K, Gerig G, Gilmore JH
  (2003) Comparisons of regional white matter
  fractional anisotrophy in healthy neonates and
  adults using a 3T head-only scanner. Radiology
  (in press).
• Warfield, S., Kaus, M., Jolesz, F., Kikinis, R.
  Adaptive template moderated spa-tially varying
  statistical classification. In Wells, W.M.e.a.,
  ed. Medical Image Computing and
  Computer-Assisted Intervention (MICCAI98).
  Volume 1496 of LNCS., Springer 1998
• Van Leemput, K., Maes, F., Vandermeulen, D.,
  Suetens, P. Automated model-based tissue
  classification of MR images of the brain. IEEE
  Transactions on Medical Imaging 18 (1999) 897908
  
• Cocosco, C.A., Zijdenbos, A.P., Evans, A.C.
  Automatic generation of training data for brain
  tissue classification from mri. In Dohi, T.,
  Kikinis, R., eds. Medical Image Computing and
  Computer-Assisted Intervention MICCAI 2002.
  Volume 2488 of LNCS., Springer Verlag (2002)
  516523
• Prastawa, M., Bullitt, E., Gerig, G., Robust
  Estimation for Brain Tumor Segmentation, MICCIA
  2003, Nov. 2003

Supported by NIH Conte Center MH064065,
Neurodevelopmental Disorders Research Center HD
03110 and the Theodore and Vada Stanley
Foundation
MICCAI Nov. 2003