| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Neuroradiology |
1 From the Departments of Radiology (J.D.E., D.J.F., J.F.M., J.M.P.), Biomedical Engineering (J.F.M.), and Community and Family Medicine, Division of Biometry (D.M.D.), Duke University Medical Center, Box 3808, Durham, NC 27710; and Department of Neurology, University of North Carolina School of Medicine, Chapel Hill (R.S.G.). Received May 31, 2000; revision requested July 17; revision received August 14; accepted September 11. J.D.E., J.M.F., R.S.G. supported in part by funds from National Institutes of Health grant NS 36-829. Address correspondence to J.D.E. (e-mail: eastw004@mc.duke.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Twenty highly related pairs of children consisting of one child with NF 1 and one unaffected child were examined. Prospective comparisons of isotropic apparent diffusion coefficient (ADC) values at five anatomic locations were performed, with and without T2-hyperintense lesions included. Retrospective analysis of hyperintense globus pallidus lesions in 16 children and in the paired control subjects also was performed.
RESULTS: Significant increases in ADC values were seen in all five anatomic locations in the NF 1 group. The greatest increases were seen in the globus pallidus (14%; P = .002) and brachium pontis (10.8%; P = .003). With exclusion of hyperintense lesions, significant ADC increases were measured in four locations. Significant ADC increases were seen in hyperintense globus pallidus lesions in the NF 1 group compared with ADC values in the normal-appearing contralateral globus pallidus (4.9%; P = .02) and those in the globus pallidus of the paired control subjects (16%; P = .003).
CONCLUSION: Significant ADC increases were measured both in the hyperintense lesions and in the normal-appearing areas of the brain in children with NF 1.
Index terms: Brain, abnormalities, 10.1831 • Brain, MR, 10.121411, 10.121413, 10.121416, 10.12144 • Children, central nervous system, 10.1831 • Magnetic resonance (MR), diffusion study, 10.12144 • Neurofibromatosis, 10.1831
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
It has been shown recently that diffusion-weighted (DW) MR imaging may depict brain water diffusion abnormalities in the brains of individuals with conditions such as Alzheimer disease, vascular dementia, and schizophrenia (5–9). The brain water diffusion abnormalities shown in these studies have been attributed to differences in axon density—that is, how closely packed the fibers are, to myelin abnormality, and, in some cases, to differences in the white matter connections that may be present in the brains of individuals with these diseases (7,9). The finding of brain diffusion abnormalities in children with NF 1 might represent evidence of microstructural abnormalities in the brains of these children.
The aim of the first and prospective part of this investigation was to compare the water diffusibility measured at five predetermined anatomic locations in children with NF 1 with that measured at the same locations in highly related (ie, siblings, half siblings, or first cousins) unaffected control subjects. We sought to test the null hypothesis that no difference between the children with NF 1 and the control subjects would be found.
The aim of the second part of this investigation was to retrospectively determine the water diffusibility properties of the nonneoplastic hyperintense (at T2-weighted maging) basal ganglia lesions that are commonly seen on MR images obtained in children with NF 1 (10–14). The available histopathologic data have shown that these hyperintense lesions are associated with myelin vacuoles, which are small areas that do not show staining at histologic examination and are therefore thought to contain fluid in vivo (15). Increases in fluid-containing spaces are expected to permit greater water diffusion. For this reason, we sought to test the hypothesis that increases in water diffusibility, as measured by using apparent diffusion coefficient (ADC) values, are associated with hyperintense lesions of the basal ganglia.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Ten pairs were first-degree relatives (ie, full siblings), and 10 pairs were second-degree relatives (ie, half siblings or first cousins). Thus, the NF 1 and control groups were similar in mean age, genetic makeup, and environmental background. Written informed consent was obtained from the families of all subjects involved in the study. Verbal consent to the study, including MR imaging, was obtained from all of the children before MR imaging. The human subjects research review boards at both institutions approved the protocol.
MR imaging was performed in all the children with NF 1 and control subjects by using a single 1.5-T commercial unit (Signa Horizon; GE Medical Systems, Milwaukee, Wis). After undergoing other imaging examinations, which included fast spin-echo intermediate-weighted (repetition time, 4,000 msec), T2-weighted (4,000/30 and 100 [repetition time msec/effective echo time msec), and fluid-attenuated inversion-recovery T2-weighted fast spin-echo (10,000/157; inversion time, 2,200 msec) sequences, the children underwent a fluid-attenuated inversion-recovery (inversion time, 2,200 msec), single-shot, spin-echo, echo-planar DW sequence (effective echo time, 102 msec; DW factor b, 0 and 1,000 mm2/sec; one excitation). Diffusion gradients were applied in each of three orthogonal directions (x, y, z), and trace-weighted DW images were generated. As a control standard for diffusion measurements, small cylindrical tubes of water were applied to each side of the head during imaging.
MR Image Analysis
All DW image analyses were performed by a single neuroradiologist (J.D.E.), who was blinded to the diagnoses of NF 1, on an imaging workstation (Advantage Windows; GE Medical Systems) by using commercial software (Functool; GE Medical Systems). Small circular regions of interest of uniform size (88 mm2) were generated and manually placed in the water tubes bilaterally and in five predetermined anatomic locations. By studying a subset of subjects, we have found the intraobserver variability of this method for determining brain ADC values to be 1%.
Two white matter locations were studied: one supratentorial region—the frontal lobe white matter—and one infratentorial region—the brachium pontis. Three gray matter locations were studied: the thalamus, globus pallidus, and hippocampus. The anatomic locations were chosen to represent the sites that are likely to appear abnormal at imaging in children with NF 1—that is, the globus pallidus and brachium pontis—and those that are less commonly involved—that is, the frontal white matter and hippocampus. The cortical gray matter was not studied because it was assumed that volume-averaging artifacts in this region would probably prevent accurate measurements.
The anatomic locations were consistently identified on the DW images in the following order: frontal white matter, thalamus, globus pallidus, hippocampus, and brachium pontis. Because anatomic detail was better on the DW images than on the ADC maps, regions of interest were placed on the DW images and in turn automatically superimposed on the ADC maps by the software that was used in the study. This strategy was especially useful for locating the globus pallidus, which often could not be differentiated from adjacent structures on the ADC maps.
After the globus pallidus was located, the region of interest was placed centrally within its nucleus. The frontal white matter was studied on the first section above the ventricles, anterior to the frontal horns (Fig 1). The thalamus location studied was in the central portion of the anterior one-half of the thalamus. The hippocampus was identified posterior to the uncus and medial to the temporal horn of the lateral ventricle. The brachium pontis measurements were made within the central portion of this structure.
|
The ADC data were compared in two ways: First, the ADC values at each anatomic location were compared between the children with NF 1 and control subjects by using all of the collected data. In a second comparison, the normal-appearing locations were analyzed separately. In this second comparison—that of the normal-appearing locations in the NF 1 group with those in the control group—each anatomic location presented three possibilities: The first possibility was that at a given anatomic location (eg, thalamus), the right- and left-sided sites both were normal appearing. In this case, the ADCmean for the child with NF 1 was compared with the ADCmean for the paired control subject. The second possibility was that only one normal-appearing site was present in a given anatomic location—for example, the thalamus on one side was normal appearing, and that on the other side contained a lesion. In this case, the ADC of the normal-appearing site was compared with the ADCmean for the paired control subject. The third possibility was that neither the left- nor right-sided site was normal appearing—for example, both thalami contained lesions. In this case, neither the ADC values at the given anatomic location in the child with NF 1 nor those at that location in the control subject were used in the second paired analysis.
For the retrospective investigation of globus pallidus lesions in the second part of this study, the subjects with globus pallidus NF 1 lesions were first separated into two groups: those with unilateral lesions and those with bilateral lesions (Fig 2). In the subjects with unilateral lesions, the ADC value of the lesion was compared with the ADC value of the normal-appearing contralateral side as well as with the globus pallidus ADCmean value for the paired control subject. In the subjects with bilateral globus pallidus lesions, the globus pallidus ADCmean value for the child with NF 1 was compared with the globus pallidus ADCmean value for the paired control subject.
|
correlations to determine possible trends associated with age and sex; and (c) repeated measures linear models for regional outcomes based on sex, age, and NF 1 status. SAS statistical software (SAS, Cary, NC) was used, and the mixed linear models procedure within this software was used to perform regression analysis. |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
No significant differences in mean water tube ADC values (2,792 ± 184 vs 2,813 ± 155 x 10-6 mm2/sec; 0.3%; P = .63) or in water tube values between the left and right sides were observed between the NF 1 and control groups.
ADC values in NF 1 versus control group.—Significant increases in mean ADC values were found in all five anatomic locations in the NF 1 group compared with these values in the same anatomic locations in the paired control subjects. The largest increases were seen in the globus pallidus (877 ± 132 vs 769 ± 40; 14.0%; P = .002) and brachium pontis (766 ± 74 vs 691 ± 59; 10.8%; P = .003) locations. The data are summarized in Table 1.
|
In the NF 1 group, no lesions were detected in frontal lobe white matter locations. Seven lesions were detected in thalamus locations; 22, in globus pallidus locations; eight, in brachium pontis locations; and six, in hippocampus locations. At analysis of only the normal-appearing sites, increases in mean ADC values were again measured at all five anatomic locations in the children with NF 1 compared with mean ADC values in the control subjects. Significant increases were found at four locations: frontal white matter, brachium pontis, hippocampus, and globus pallidus. The data are summarized in Table 2.
|
Regression Analysis of Age and ADC
A significant negative linear relationship between age and frontal white matter ADC was observed in the frontal lobe white matter (Fig 3) in both the NF 1 group and the control group. The Kendall rank b correlation coefficient for the NF 1 group was -0.46 (P = .005). This correlation coefficient for the control subjects was -0.48 (P = .003). The slope obtained by using model fitting with both groups (n = 40) was -10 x 10-6 mm2/sec/y. There was no significant difference in slope between groups. A significant correlation was observed also at the brachium pontis location (r = -0.60; P = .005) in the NF 1 group. A smaller correlation that approached significance was observed at the brachium pontis location (r = -0.31; P = .06) in the control group. Correlations at the other locations were not significant.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
In the present study, significant increases in brain water diffusion were found to be associated with the diagnosis of NF 1. Recent MR spectroscopy data in children with NF 1 have shown absolute increases in choline and absolute decreases in N-acetylaspartate, both in the hyperintense lesions and in the normal-appearing thalamus and globus pallidus (17). The authors concluded that a widespread myelin disorder, such as demyelination followed by remyelination, could be present in children with NF 1.
Existing pathologic data have shown an association between hyperintense basal ganglia lesions and myelin vacuoles (15), which may represent further evidence of a demyelinating process. Thus, hyperintense lesions could represent focally severe disease within a more widespread myelin disorder. Thus, interpreted in this context, our data may represent further evidence that a myelin disturbance, such as diminished myelin amount, increased myelin turnover, or demyelination, is present in the brains of children with NF 1. In accordance with this hypothesis, the increases in ADC values that were measured in the frontal white matter and brachium pontis may be explained by the fact that these regions consist mainly of myelinated fibers that are presumably affected by the proposed myelin disorder. Thus, the increases in ADC values that are associated with gray matter–containing locations such as the globus pallidus may be related to the content of abnormally myelinated fibers in these regions.
Other possible explanations for the increased ADC values seen in the children with NF 1 in our study are that differences in cellularity or factors related to axon fibers, such as decreased organization or axon number, exist in individuals with NF 1. Decreases in overall cellularity or axon number or diminished white matter tract organization are expected to permit greater diffusibility. Within this model, hyperintense lesions could still be due primarily to a myelin disorder, consistent with the findings of Dipaolo et al (15), and confer additional increases in ADC independent of the disorders causing the ADC increases in the normal-appearing areas of the brain.
To our knowledge, there are few published normative ADC data on children of the age range that we studied, and most of the studies available involved methods that are not directly comparable to those used in the present study (18–20). The studies in which ADC values were measured by using a method that is comparable to that used in our study involved assessment of ADC values in infants and adults (21,22). In the present study, the mean ADC value in the frontal white matter in the control subjects was 795 x 10-6 mm2/sec compared with a previously reported range of 710–790 x 10-6 mm2/sec in adults without brain abnormalities (21,22).
The significant negative correlations between frontal white matter ADC values and age observed in both the children with NF 1 and the control subjects in this study are consistent with previously obtained data on subjects without brain abnormalities, which showed decreasing frontal lobe white matter ADC values and increasing anisotropy indexes with age (18,20). The authors of those studies suggested that frontal lobe myelination and organization continue into the 2nd decade of life (18–20). If this is true, then the similar regression slopes of the NF 1 and control groups (Fig 3) in the present study suggest that frontal lobe myelination in children with NF 1 proceeds at nearly the same rate as does that in control subjects. We intend to test this hypothesis in a longitudinal study. The correlations between brachium pontis ADC values and age measured in both the children with NF 1 and the control subjects suggest that myelination may be occurring in this location as well as in the frontal lobe white matter.
The difference in ADC values between the right and left sides of the thalamus in the NF 1 group was significant but small. We will study differences in diffusion parameters between the right and left sides in future investigations, although sustainable differences are thought to be unlikely.
Full-tensor acquisition is a DW imaging method that provides knowledge of the axes of favored diffusion, such as the direction of the white matter tracts (23). Tensor DW imaging might facilitate a better understanding of the abnormalities seen in the present study, be more accurate, and enable us to better test the hypothesis that a demyelinating process is present by enabling assessment of diffusional anisotropy. We intend to study diffusion changes in children with NF 1 further by using full tensor data.
A possible limitation of the present study was that true blinding of the radiologist to the subject’s NF 1 status was not always possible, because in some cases, obvious focal lesions were present. However, this did not affect the placement of the regions of interest in the predetermined locations, so we think that it is unlikely that that this factor substantially influenced the data.
In conclusion, brain ADC values—both in the hyperintense lesions of the basal ganglia and in the normal-appearing locations—were found to be higher in children with NF 1. The precise mechanisms of these increases are not known. However, on the basis of evidence from prior studies, our findings may represent further evidence that a microstructural abnormality, such as myelin abnormality, is present in individuals with NF-1.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, J.D.E.; study concepts, all authors; study design, J.D.E., J.M.P., J.F.M., R.S.G.; literature research, J.D.E., D.J.F., R.S.G.; clinical studies, R.S.G.; experimental studies, J.D.E., J.F.M.; data acquisition, J.D.E.; data analysis/interpretation, all authors; statistical analysis, D.M.D., J.D.E.; manuscript preparation, J.D.E., D.J.F.; manuscript definition of intellectual content, all authors; manuscript editing, J.D.E., D.J.F., J.M.P.; manuscript revision/review, J.D.E., D.J.F.; manuscript final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S.J.P.M. van Engelen, L.C. Krab, H.A. Moll, A. de Goede-Bolder, S.M.F. Pluijm, C.E. Catsman-Berrevoets, Y. Elgersma, and M.H. Lequin Quantitative Differentiation Between Healthy and Disordered Brain Matter in Patients with Neurofibromatosis Type I Using Diffusion Tensor Imaging AJNR Am. J. Neuroradiol., April 1, 2008; 29(4): 816 - 822. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.L. Zamboni, T. Loenneker, E. Boltshauser, E. Martin, and K.A. Il'yasov Contribution of Diffusion Tensor MR Imaging in Detecting Cerebral Microstructural Changes in Adults with Neurofibromatosis Type 1 AJNR Am. J. Neuroradiol., April 1, 2007; 28(4): 773 - 776. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Alkan, A. Sigirci, R. Kutlu, S. Doganay, G. Erdem, and C. Yakinci Giant Axonal Neuropathy: Diffusion-Weighted Imaging Features of the Brain J Child Neurol, October 1, 2006; 21(10): 912 - 915. [Abstract] [PDF]
|
|
|
|
 |
|
 S. M. Kealey, T. Aho, D. Delong, D. P. Barboriak, J. M. Provenzale, and J. D. Eastwood Assessment of Apparent Diffusion Coefficient in Normal and Degenerated Intervertebral Lumbar Disks: Initial Experience Radiology, May 1, 2005; 235(2): 569 - 574. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Bennett, T. A. Rizvi, S. Karyala, R. D. McKinnon, and N. Ratner Aberrant Growth and Differentiation of Oligodendrocyte Progenitors in Neurofibromatosis Type 1 Mutants J. Neurosci., August 6, 2003; 23(18): 7207 - 7217. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L.E. Cutting, K.L. Cooper, C.W. Koth, S.H. Mostofsky, W.R. Kates, M.B. Denckla, and W.E. Kaufmann Megalencephaly in NF1: Predominantly white matter contribution and mitigation by ADHD Neurology, November 12, 2002; 59(9): 1388 - 1394. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
= data from the control group, 
= data from the NF-1 group. The solid line represents the linear trend of the control group, and the broken line represents the linear trend of the NF-1 group. The difference in the y intercept was -55 ± 8 x 10-6 mm2/sec (P = .001). The difference in y intercepts between the linear regressions represents the difference in ADC between the groups and was similar to the difference in measured mean ADC values between the groups (-53 x 10-6 mm2/sec). The variation in frontal white matter ADC with age was probably due to progressive myelination.