Image Processing Reference
In-Depth Information
growth trajectories. Results confirm qualitative descriptions of growth
patterns in pediatric radiology studies and our new quantitative mod-
eling scheme has the potential to advance understanding of variability
of brain tissue maturation and to eventually differentiate normal from
abnormal growth for early diagnosis of pathology.
1
Introduction
Early brain development involves a complex sequence of biophysical and chemi-
cal changes occurring in systematic progression. These changes including cortical
folding, premyelination changes in white matter, and myelination, can be clearly
seen in brain MR (Magnetic Resonance) images [
1
]. Several qualitative studies
have attributed significant changes in MR image appearance seen in the first
two years of life to myelination [
2
,
3
]. Myelination manifests as changing relative
contrast between white matter and gray matter tissues in T1W (T1-Weighted)
and T2W (T2-Weighted) MR images. MR studies of neurodevelopment confirm
that these appearance changes are highly modality-specific. Each modality cap-
tures different phases of myelination resulting in differential timing of contrast
change trajectories [
1
-
3
].
Most quantitative studies of the early brain have focused on volumetric and
morphometric indicators, as well as microstructural parameters such as diffu-
sion [
4
,
5
]. The usage of image appearance as a complementary indicator of brain
maturation is relatively much less explored although a key feature in pediatric
radiological exams. A few recent studies have modeled spatiotemporal changes
in signal intensity (SI) to better understand neurodevelopmental processes [
6
,
7
].
However, using SI for appearance analysis increases dependence on effective nor-
malization schemes to account for variability in intensity range between scans.
The primary goal of this work is to jointly model appearance changes in MR
images of the developing brain across multiple modalities using relative white-
gray matter contrast. The usage of the distance between white matter and gray
matter intensity distributions to quantify white-gray contrast provides greater
stability to inter-scan variations compared with signal intensity analysis [
8
]. We
apply this method to imaging studies of early brain development that are longi-
tudinal and multimodal by design.
Most imaging studies use a combination of different modalities of MRI such
as T1W and T2W to study the early developing brain. Patterns of appearance
change seen in commonly used MPRAGE (Magnetization Prepared Rapid Acqui-
sition GRE) T1W and FSE (Fast Spin Echo) T2W scans of the early developing
brain are shown in Fig.
1
. The T1W scans show increasing brightness of white
matter relative to gray matter with progression of age while the T2W scans show
the reversed pattern (decreasing brightness of white matter). The signal intensi-
ties and tissue properties captured by T1W and T2W scans depend on T1 and
T2 relaxation times. The shortening of T1 relaxation time occurs in relation to
an increase in cholesterol and glycolipids during the early phase of myelin forma-
tion, while the shortening of T2 relaxation time is associated with later stages of
Search WWH ::
Custom Search