| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiological Sciences, University of California at Los Angeles
Medical Center, 10833 Le Conte Ave., Los Angeles, CA 90095-1721.
2
Department of Radiology, Lucile Salter Packard Children's Hospital, Stanford
University Medical Center, 725 Welch Rd., Palo Alto, CA 94304.
3
Department of Radiology, Children's Hospital Boston, 300 Longwood Ave.,
Boston, MA 02115.
4
Department of Pediatrics, Oregon Health Science University, 707 S.W. Gaines
Rd., Portland, OR 97201.
5
New England Research Institute, 9 Galen St., Watertown, MA 02472.
6
Department of Biostatistics, University of California at Los Angeles Medical
Center, Los Angeles, CA 90095.
7
Department of Pathology, Children's Hospital, Boston, MA 02115.
8
Department of Neurology, Children's Hospital, Boston, MA 02115.
Received November 2, 2000;
accepted after revision March 1, 2001.
Address correspondence to A. Panigrahy.
Abstract
 Top Abstract IntroductionMaterials And MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. We retrospectively reviewed the MR images of 70 patients who were between the ages of 1 and 5 years and whose images showed PVWM T2-signal hyperintensities. The patients were divided into premature (n = 35 children) and term (n = 35) groups depending on their gestational age at birth. Volumetric analysis was performed on four standardized axial sections using T2-weighted images. Volumes of interest were digitized on the basis of gray-scale densities of signal intensities to define the hemispheric cerebral white matter and lateral ventricles. Age-adjusted comparisons of volumetric measurements between the premature and term groups were performed using analysis of covariance.
RESULTS. The volume of the cerebral white matter was smaller in the premature group (54 ± 2 cm3) than in the term group (79 ± 3 cm3, p < 0.0001). The volume of the lateral ventricles was greater among the patients in the premature group (30 ± 2 cm3) than among those in the term group (13 ± 1 cm3, p < 0.0001). Fifty percent of all the premature children had spastic diplegia or quadriplegia. Thirty-two percent of all the term children had hypotonia. There were patients in both groups whose PVWM T2-signal hyperintensities did not correlate with any neuromotor abnormalities but were associated with seizures or developmental delays.
CONCLUSION. The differences in volumetric measurements of cerebral white matter and lateral ventricles in children with PVWM T2-signal hyperintensities are related to their gestational age at birth. Several neurologic motor abnormalities are found in children with such hyperintensities.
|
TopAbstractIntroductionMaterials And MethodsResultsDiscussionReferences |
|---|
Evidence suggests that there are differences in both clinical outcomes and MR imaging patterns of PVWM injury between infants born prematurely and those born at term. For example, studies have shown that PVWM injury in the premature infant may result in decreased cerebral white matter and ventriculomegaly in the neonatal period [2, 12,13,14]. Separate studies have suggested that PVWM injury in term infants tends to be "milder" in both the types of neuromotor deficits and MR imaging features seen [15,16,17]. Children who were different gestational ages at the time of birth display major differences in pathogenetic associations and clinical patterns of cerebral palsy [18]. However, overall, the pathophysiologic mechanism of these syndromes and differences in their radiologic manifestations in children are not well understood, particularly those appearing after a child's first birthday.
The purpose of this study was to compare the volumes of cerebral white matter and lateral ventricles with gestational age at birth for children with retrospectively identified PVWM T2-signal hyperintensities. We also sought to determine the spectrum of neuro-motor abnormalities associated with such hyperintensities in these children. We hypothesized that children with PVWM T2-signal hyperintensities who had been born prematurely would have decreased cerebral white matter volume and increased lateral ventricle volume compared with children with PVWM T2-signal hyperintensities who had been carried to term, a difference that has been found in children during the neonatal period [2].
|
TopAbstractIntroductionMaterials And MethodsResultsDiscussionReferences |
|---|
The inclusion criteria of the study was the presence of PVWM T2-signal hyperintensities located predominately around the trigones and the frontal horns of the lateral ventricles. Patients in whom the hyperintensities were thought to have been previously described as "terminal myelination zones" were not included [4, 5]. Although diffuse and excessive T2-signal hyperintensities have been described in the cerebral white matter of neonates [19], an abnormal PVWM T2 signal cannot be clearly distinguished from normal T2 hyperintensity of unmyelinated PVWM until approximately the child's first birthday. Patients less than 90 postconceptional weeks old (postconceptual age being calculated as the child's gestational age at birth added to the child's postnatal age) were excluded because it was not possible to distinguish premyelinated PVWM T2 hyperintensity from abnormal PVWM T2 hyperintensity (Fig. 1A,1B) in these patients. Other exclusion criteria included posthemorrhagic hydrocephalus; cerebral cortical abnormalities that would result in relatively abnormal volumetric measurements; known metabolic or infectious disorders associated with periventricular T2 hyperintensities; other intracranial pathologies; and macrocephaly or microcephaly that would also result in relatively abnormal volumetric measurements. Patients who were born at a gestational age of 37 weeks or less were defined as premature whereas patients who were born at a gestational age more than 37 weeks were considered to be term.
|
|
Clinical Data
Clinical data, including the gestational age at birth and predominate
neuromotor abnormality at the time of the MR imaging, were recorded for each
patient. The patient's neuromotor abnormality was retrospectively identified
after a review of the detailed neurologic examination data in the medical
record. A pediatric neurologist evaluated the children taking part in our
study with neuromotor abnormalities in either an outpatient cerebral palsy
clinic or inpatient setting. Other clinical data collected included birth
weight, Apgar scores, history of pregnancy complications, history of
resuscitation (intubation) at birth, and history of seizures
(Table 1). Some of these
clinical variables were not available for every child.
|
MR Imaging Protocol and Parameters
MR imaging was performed on a 1.5-T imaging system to obtain fast spin-echo
axial T2-signal images (TR/TEeff, 3400/84; 1 excitation; 24-cm
field of view; 5-mm section thickness with a 1-mm intersection gap; 256
x 192 acquisition matrix; and echo train length, 8); axial
fluid-attenuated inversion recovery (FLAIR) fast spin-echo images
(TR/TEeff, 10004/162; 1 excitation; TI, 2200; 24-cm field of view;
5-mm section thickness with a 1-mm intersection gap; and 256 x 192
acquisition matrix); and axial proton-density fast spin-echo images
(TR/TEeff, 3400/17; 1 excitation; 24-cm field of view; 5-mm section
thickness with a 1-mm intersection gap; 256 x 192 acquisition matrix;
and echo train length, 4).
Quantitative Volumetric Analysis
Because most of the records identified in the database were not available
on digital archives (only printed films with calibration and section thickness
information were available), previously described three-dimensional MR
automated volumetric programs were not used
[2]. Instead, a technique based
on the calculation of a volume from a standardized sample of serial sections
of boundary contours was used to calculate an estimated volume of cerebral
white matter and lateral ventricle. Because the gray-scale density of a T2
signal is well differentiated between gray matter, mature white matter, and
bright cerebral spinal fluid in the ventricle, a densitometry program (MCID
imaging system; Imaging Research, St. Catherines, Ontario, Canada) was used to
create automated boundary contours on the basis of the differential relative
optical densities. Volumetric analysis was performed on four standardized
axial sections (including the trigones of the lateral ventricles) using
T2-weighted images and, when available, corresponding FLAIR sequences. The
four selected axial sections were digitized from printed films and calibrated
with the scale bar using the imaging software. The four sections were
rigorously matched on all patients by a pediatric neuroradiologist.
Volumes-of-interest boundary contours were digitized to define the hemispheric cerebral white matter and lateral ventricles. The boundary contour of the hemispheric cerebral white matter included the subcortical white matter of the frontal, insular, parietal, and occipital cortices and did not include the corpus callosum (genu and splenium) or the white matter tracts in the basal ganglia and thalamus. The boundary contour of the lateral ventricle included the frontal horns, the third ventricle, the body of the lateral ventricle, the region of the trigone, and the occipital horns. As a comparison, the cerebral cortex was also measured in this manner. The boundary contour of the cerebral cortex included all parts of the cortex. The digitization was performed and monitored by a radiologist, and manual editing of the boundary contours was performed as needed. Because the gray-scale density of the high PVWM T2-signal hyperintensity did not differ significantly from the surrounding high signal in the lateral ventricle, a separate method was used to create the boundary contour of the PVWM signal hyperintensities. Comparable FLAIR and proton-density axial images were toggled with the T2-signal axial images, using an overlay function of the imaging software to help delineate the contour of the PVWM signal hyperintensities from the border of the lateral ventricles. PVWM T2-signal hyperintensities not thought to be pathologic—such as a high signal from Virchow-Robin spaces—were edited out manually. After the volumes of interest were constructed on each of the four slices, a three-dimensional algorithm was used to calculate the volume on the basis of the area of the boundary contour and the appropriate thickness of the sections. The algorithm used for three-dimensional reconstruction and volume calculation from a series of boundary contours was based on both anatomic boundaries and differential densities and has been previously described by us [20, 21].
Statistical Analysis
For the clinical continuous variables, a two-sample-pooled Student's
t test was used to determine the significance between the children in
the premature and term groups. To account for missing data points, a maximum
likelihood technique was used to determine the significance in the continuous
variables between the two groups.
For the clinical categoric variables, a Fisher's exact test was used to determine the significance of differences between the premature and term group. To account for missing data points, a Fisher-Freeman-Halton exact test was used to determine the significance in the categorical variables between the premature and term groups.
For the quantitative analysis, age-adjusted comparisons of volumetric measurements between premature and term infants were performed using analysis of covariance. As previously discussed, a patient's postconceptional age was calculated by adding the gestational age at birth to the postnatal age. Four of the term patients had medical records that clearly indicated that the patients were term but did not give their specific gestational ages at birth. For these patients, the gestational age at birth was assumed to be 40 weeks, and the postconceptional age was calculated accordingly. Crude comparisons (not adjusted for age) of volumetric measurements between the premature and term groups were performed using Student's t test. To meet normality assumptions, the volume of the PVWM intensity was log transformed before analysis.
|
TopAbstractIntroductionMaterials And MethodsResultsDiscussionReferences |
|---|
Effect of Age on the Volumetric Measurements
The premature and the term groups did not differ significantly with regard
to post-conceptional age (155.9 ± 62.1 weeks for term vs 155.2 ±
64.9 weeks for premature, p = 0.774) or postnatal age (p =
0.565). There was a positive effect of age on the volumetric measurements of
the cerebral cortex (p < 0.015) and a marginal negative effect of
age on the volumetric measurement of the PVWM T2-signal hyperintensities
(p = 0.089). Therefore, although the mean post-conceptional age at
the time of imaging of the two groups was similar, age-adjusted means were
used to gain statistical power when comparing these measurements between the
premature and term infants. The effect of age on the volumetric measurements
of the cerebral white matter and the lateral ventricles was different for the
premature and the term groups (p < 0.05). Therefore, raw means
were used to compare these measurements between the two groups.
Comparative Volumetric Analysis Between Premature and Term
Groups
The volume of the cerebral white matter was less in the premature group (54
± 2 cm3) than in the term group (79 ± 3
cm3, p < 0.0001, Figs.
2A,2B,3,4A,4B,5A,5B,5C,6A,6B,7).
The volume of the lateral ventricles was greater in the premature group (30
± 2 cm3) than in the term group (13 ± 1
cm3, p < 0.0001, Figs.
2A,2B,3,4A,4B,5A,5B,5C,6A,6B,
8). The log-transformed volume
of the PVWM T2-signal hyperintensities was slightly greater in the premature
group (1.6 ± 0.1 cm3) than in the term group (1.3 ±
0.1 cm3, p = 0.033, Figs.
2A,2B,3,4A,4B,5A,5B,5C,6A,6B).
Although the volume of the cerebral cortex was less in the premature group
(119 ± 2 cm3) than in the term group (124 ± 2
cm3), the difference was only marginally significant (p =
0.082).
|
|
|
|
|
|
|
|
|
|
|
|
Comparative Developmental Trends in Cerebral White Matter and Lateral
Ventricular Volumes Between Premature and Term Groups
In the cerebral white matter data set, the volume of cerebral white matter
increased with age in both the premature and the term groups
(Fig. 7). However, the volume
of cerebral white matter in the term infants increased at a greater rate
during development than in the premature infants
(Fig. 7). The slope for the
term group was 0.19 cm3 per week (p = 0.0001), whereas the
slope for the premature group was 0.07 cm3 per week (p =
0.034, Fig. 7).
In the lateral ventricle volume data set, the volume of the lateral ventricle decreased slightly with age in the term group (Fig. 8). This decrease is in contrast to findings for the premature group, in which the volume of the lateral ventricle increased with age (Fig. 8). The slope of the term group was -0.02 cm3 per week (p = 0.319; the slope is not statistically different from zero) whereas the slope of the premature group was 0.05 cm3 per week (p = 0.027, Fig. 8).
Neurologic Motor Outcome at Time of MR Imaging
In the premature group, 50% of the children had spastic diplegia or
quadriplegia; 29% had other abnormalities, including hypotonia, hypertonia,
hemiparesis, and ataxia; and 21% had no neuromotor abnormalities. In the term
group, 9% of the children had spastic diplegia or quadriplegia; 32% had
hypotonia; 32% had other abnormalities, including hypertonia, hemiparesis, and
ataxia; and 27% had no neuromotor abnormality.
|
TopAbstractIntroductionMaterials And MethodsResultsDiscussionReferences |
|---|
Our data suggest that the children with a history of premature birth and PVWM injury sustained a greater degree of injury than did the children with a history of term birth and PVWM injury, as manifested by greater diminution of cerebral white matter volume and more pronounced ventriculomegaly. Although the differences are clear, there is overlap of the findings between term and premature children.
The spectrum of neuromotor disturbances most likely is related to the complex interplay of three major factors: the severity, timing, and cause of the insult to the developing cerebral white matter [13]. The severity of the insult is likely to be more marked in the premature infant because of the presence of cerebral and systemic circulatory disturbances, in part related to the infant's respiratory disease and consequent need for mechanical ventilation. The effect of the timing of the insult is intrinsically related to the vulnerability of the infant's cerebral white matter. For example, perhaps premature infants sustain greater injury to cerebral white matter because of the predominance of oligodendrocyte precursors, which are more vulnerable to hypoxia-ischemia or infection than the mature white matter of term infants [13]. Studies of oligodendrocytes in cultures suggest a maturation-dependant vulnerability of oligodendroctye precursors to free-radical attack; in contrast, mature oligodendrocytes appear resistant to such attacks [22]. The decreased cerebral white matter volume with secondary ventriculomegaly that we detected in our study in the premature children may represent the end result of the "diffuse" component of periventricular leukomalacia, which may be detected in the acute and subacute phase during the neonatal period by diffusion-weighted imaging techniques [1]. In light of our findings, it is interesting to note that pathologic focal periventricular leukomalacia lesions without ventriculomegaly are found in term infants with surgically corrected congenital heart disease [17].
We did find that the premature group with PVWM T2-signal hyperintensities had a slightly greater volume of these hyperintensities than the term group did. The histopathologic correlates of these T2-signal hyperintensities are not entirely clear, given the lack of radiologic—pathologic correlative studies of children between 1 and 5 years old, the age span of the children in our study [23,24,25,26]. Earlier in development, the PVWM is characteristically T2-signal hyperintense, which corresponds to the MR appearance of fiber bundles of unmyelinated white matter as seen in radiologic—pathologic studies of premature infants seen at autopsy [4, 23,24,25,26,27]. The PVWM T2-signal hyperintensities normally decrease across development with the progression of myelination. Barkovich [4] has found that the changes on the T2-weighted images correspond temporally with the formation of the complete myelin sheath and probably reflect a change in water distribution from an intact myelin sheath. The abnormal PVWM T2 signal that persists beyond the first year of life may be the result of two possible pathologic processes: axons in the periventricular region have not been properly wrapped with myelin, suggestive of injury to the oligodendrocytes or their precurors; replacement of normal hydrophobic myelin by gliosis or hypertrophic astrocytes (increased cellularity) has caused an increased signal; or both of these pathologies are present. Studies have shown that quantitative assessment of water diffusion by diffusion-tensor MR imaging mirrors the microstructural development in cerebral white matter that may account for changes in myelination in the PVWM [28, 29]. Both the decreased volume of cerebral white matter and the increased volume of PVWM T2-signal hyperintensities in the premature infants compared with term infants support the idea that the injury to immature white matter (of the premature infants) is more pronounced than injury to the mature white matter (of the term infants).
Periventricular white matter T2-signal hyperintensities are associated with a broad spectrum of neuromotor abnormalities. The relationship between premature birth and the presence of decreased cerebral white matter volume, secondary ventriculomegaly, and spastic diplegia has been described by others [13, 15, 16]. The finding of hypotonia as a type of cerebral palsy among the term infants in our study has also been described previously [11, 18]. The relationship between the hypotonia and PVWM injury is not clearly understood. Also, our data do not show a correlation between all patients with PVWM T2-signal intensity and the presence of a clinical neuromotor abnormality. However, these patients did have a considerable history of seizures or history of developmental delay suggestive of abnormal neurologic development. Other studies have shown that periventricular leukomalacia in term children can present with various neurologic abnormalities including seizures, developmental delay, and heterogeneous motor findings [30]. Some researchers have suggested that although children with evidence of periventricular white matter disease are at increased risk for seizures, there is no clear correlation between findings on MR images and seizure patterns [31].
Possibly, some children in our study had PVWM T2-signal hyperintensities that represented "exaggerated" terminal myelination zones. The criteria by which one could differentiate terminal myelination zone from subtle abnormal periventricular signal are not entirely clear [4, 5]. Baker et al. [5] noted that a layer of myelinated white matter is normally present between the trigone of the lateral ventricle and the terminal zone of hyperintensities on images. It has been suggested that this terminal zone of hyperintensity corresponds to the known delayed myelination of the fiber tracts involving the association areas of the posterior and inferior parietal and posterior temporal cortices [5]. Yakovlev and Lecours [32] have described these areas as terminal zones because some of the axons in these regions do not stain for myelin until the patient reaches the fourth decade of life. However, the equally extensive myelination studies by Brody et al. [33] and Kinney et al. [34] do not support the existence of these unmyelinated fiber tracts in the first two years of life. More detailed radiologic—pathologic correlation studies are needed to determine conclusively how to differentiate terminal myelination zones from subtle focal periventricular injuries with no neurologic consequences.
Potential limitations of our study include its retrospective nature and the inability to accurately quantify and standardize the neuromotor assessments. Our three-dimensional technique is limited in that it provides an estimated volume based on a standardized series of sample sections. However, this technique does allow direct digitization and manual placement of boundary contours around areas with which a purely automated program would have difficulty. Moreover, the four standardized axial levels used in our study would be the key diagnostic axial sections that neuroradiologists would routinely use to determine associated characteristics of PVWM injury. The volumetric differences seen in this analysis may reflect the effect of clinical variables that were not specifically correlated with volumetric measurements in our study. Also, the developmental trends described in our study reflect a group of pediatric patients in which each patient represents a single point in time in a child's development.
In conclusion, our data suggest that there are volumetric differences in cerebral white matter and lateral ventricles between children with PVWM T2-signal hyperintensities related to a child's gestational age at birth. Also, a spectrum of neuromotor abnormalities are found in children with such hyperintensities. Future research aimed at correlating specific neuromotor abnormalities with volumetric data will be important in characterizing which patterns of PVWM T2-signal hyperintensities are clinically important and likely to result in neuromotor abnormalities in both premature and term infants.
Acknowledgments
We thank Philip F. Morway, Virginia Grove, and Lolita Lewis for technical
assistance and Donald Sucher for help with photography.
|
TopAbstractIntroductionMaterials And MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Bava, S. L. Archibald, and D. A. Trauner Brain Structure in Prenatal Stroke: Quantitative Magnetic Resonance Imaging (MRI) Analysis J Child Neurol, July 1, 2007; 22(7): 841 - 847. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Fraser, L. Bennet, R. Helliwell, S. Wells, C. Williams, P. Gluckman, A. J. Gunn, and T. Inder Regional Specificity of Magnetic Resonance Imaging and Histopathology Following Cerebral Ischemia in Preterm Fetal Sheep Reproductive Sciences, February 1, 2007; 14(2): 182 - 191. [Abstract] [PDF]
|
|
|
|
 |
|
 C. Garel, A.-L. Delezoide, M. Elmaleh-Berges, F. Menez, C. Fallet-Bianco, E. Vuillard, D. Luton, J.-F. Oury, and G. Sebag Contribution of Fetal MR Imaging in the Evaluation of Cerebral Ischemic Lesions AJNR Am. J. Neuroradiol., October 1, 2004; 25(9): 1563 - 1568. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. H. Hoon JR, K. M. Belsito, and L. M. Nagae-Poetscher Neuroimaging in Spasticity and Movement Disorders J Child Neurol, January 1, 2003; 18(1_suppl): S25 - S39. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
= term children.