 |
|
Brain Morphometry, T2-Weighted Hyperintensities, and IQ in Children With Neurofibromatosis Type 1
Robert S. Greenwood, MD;
Larry A. Tupler, PhD;
J. Kenneth Whitt, PhD;
Anne Buu, PhD;
Carrie B. Dombeck, BS;
Amanda G. Harp;
Martha E. Payne, MPH, RD;
James D. Eastwood, MD;
K. Ranga R. Krishnan, MD;
James R. MacFall, PhD
Arch Neurol. 2005;62:1904-1908.
ABSTRACT
| |
Background Larger gray matter (GM) volume in healthy children is correlated with higher IQ. Children with neurofibromatosis type 1 (NF1) have larger brains, their magnetic resonance images frequently show T2-weighted hyperintensities, and their IQs are lower.
Objectives To confirm the hypotheses that (1) children with NF1 have larger GM and white matter volumes, (2) the greatest volume differences are in the frontal and parietal regions and in children with NF1 with hyperintensities, and (3) GM volume is inversely related to IQ in children with NF1.
Design Wechsler Intelligence Scale for Children–Third Edition IQ testing and measurement of cerebral volumes and hyperintensities in brain magnetic resonance images were performed on 36 children with NF1 and on 36 matched relatives who served as control subjects.
Results Gray matter and white matter volumes were significantly larger in children with NF1. The greatest difference was observed in cerebral white matter volume, predominantly in the frontal lobes, whereas the greatest difference in GM volume was in the temporal, parietal, and occipital regions. In controls, IQ was significantly related to GM volume, but in children with NF1, IQ was not inversely associated with GM volume, although IQs of children with NF1 were significantly lower.
Conclusions Children with NF1 do not have the normal relationship between GM volume and IQ. Larger GM volume in the posterior brain regions and larger white matter volumes in the frontal brain regions contribute to the larger brain volume in children with NF1.
INTRODUCTION
Patients with neurofibromatosis type 1 (NF1) often have a large head size,1 and structural abnormalities, particularly signal hyperintensities, are commonly found in their brain magnetic resonance (MR) images.2-4 Children with NF1, on average, have lower IQs than those of healthy children.5 Brain volume, especially gray matter (GM) volume, in healthy children correlates positively with general intelligence.6 Increased brain volume in specific brain regions has been associated with cognitive dysfunction.7-8 Because NF1 involves some loss of growth control, we may gain a better understanding of the cognitive deficits in NF1 by defining whether the greater brain volume in NF1 is of GM or white matter (WM) origin, by specifying which brain regions comprise the volume changes, and by determining whether IQ is related to these changes.
Four studies3-4,9-10 examined brain MR imaging volumetric differences between patients with NF1 and control subjects. These studies confirmed that children with NF1 have significantly larger brain size, but each reached different conclusions regarding the contribution of GM and WM to the brain volume increase. Only one of these, a study4 limited to boys with NF1, attempted to define which regions of the brain have significant increases in WM or GM volume. Two studies3-4 found that increased brain volume was correlated with the presence of T2-weighted hyperintensities in children with NF1. Results from these previous studies have limited applicability to many children with NF1 because they enrolled special populations and did not have genetic controls for all of the subjects with NF1. For these reasons, we undertook a larger prospective study of cerebral lobar and regional GM and WM volumes and IQ in children with NF1 compared with family-matched genetic controls.
We hypothesized that whole-brain volumes, primarily supratentorial WM volume and the cross-sectional area of the corpus callosum, would be larger in children with NF1 and that IQ would be inversely related to brain volume in children with NF1. Because motor, visuospatial, and attentional processes are typically compromised in NF1, we also hypothesized that significant differences in WM volume between children with NF1 and family controls would be observed in the anterior brain and in predominantly parietal brain regions. We hypothesized further that children with NF1 with hyperintensities in the globus pallidus would have significantly larger WM volume.
METHODS
The human subjects research review boards at the University of North Carolina School of Medicine, Chapel Hill, and Duke University Medical Center, Durham, NC, approved the study protocols. All subjects (age range, 6-13 years) were from North Carolina or adjacent states and had been evaluated by a pediatric geneticist or a child neurologist to confirm the NF1 diagnosis.11 Children were excluded if they had the following: (1) a history of a clinically significant medical disease or condition that might severely compromise their ability to participate; (2) a history of serious psychiatric illness; (3) continuing seizures or significant neurologic deficit; (4) a diagnosis of brain tumor, except stable optic gliomas; or (5) a documented IQ less than 70. The order of priority for matching controls to children with NF1 was as follows: (1) sibling, (2) age within 3 years of the child with NF1, (3) same sex as the child with NF1, and (4) second-degree relative, chosen in priority order of first cousin, aunt, or uncle of the same sex.
We identified 85 potential subjects with NF1. Forty-six subjects were excluded for the following reasons: (1) they did not have a closely related control subject (24 subjects), (2) they were not interested in participating in the study (10 subjects), (3) they could not undergo MR imaging (5 subjects), or (4) they met 1 of the medical exclusion criteria (7 subjects). The final cohort consisted of 36 children with NF1 (21 male; mean ± SD age, 9.3 ± 2.3 years) and 39 siblings or second-degree relatives (20 male; mean ± SD age, 9.5 ± 2.5 years). After study entry, we excluded MR imaging data from 2 NF1 and control pairs because 1 subject was too fearful to undergo MR imaging and because the MR image of 1 subject could not be analyzed because of technical difficulties.
MR IMAGING AND ANALYSIS
Unsedated subjects underwent MR imaging with a 1.5-T whole-body MR imaging system (Signa; GE Medical Systems, Milwaukee, Wis) using the standard head (volumetric) radiofrequency coil. High-resolution images were obtained with a dual-echo fast spin-echo sequence (repetition time, 4000 milliseconds; echo times, 30 milliseconds [first echo] and 100 milliseconds [second echo]; and echo train length, 16 echoes [8 for each echo]), with a 256 x 256-pixel matrix, 3-mm section thickness with no gaps between sections, 1 signal acquired per phase-encoding increment, and a 22-cm field of view.
The MR images were processed using the Duke University Medical Center Neuropsychiatric Imaging Research Laboratory–modified version of MRX software (GE Corporate Research and Development, Schenectady, NY) for image segmentation and the GRID program developed by the laboratory, as previously described.12 The segmentation protocol used by the laboratory13 was modified from that developed by Kikinis et al14 and Byrum et al.15 Reliability was established by repeated measurements on multiple MR images before raters were approved to process the study data.
Total hemisphere volumes included GM and WM in cortical and subcortical structures, excluding the cerebellum and infratentorial structures. The GM and WM components were separately assessed after segmentation. Cortical GM volume was assessed after excluding subcortical structures by means of MRX software.
The volumes of the left and right hemispheres were further subdivided using the GRID program into brain regions that approximated lobar regions of the brain. We realigned the 3-dimensional images to a standard orientation, including making the anterior commissure–posterior commissure plane horizontal, and defined planes according to a previously described regional stereotactic coordinate system16 (Figure). To measure the area of the corpus callosum, it was outlined in a midsagittal MR image reformatted from an axial T2-weighted fast spin-echo image on a commercial imaging workstation (Advantage Windows Workstation; GE Medical Systems).
|
|
|
|
Figure. Drawings of the left and right sides of the brain showing planes defining neuroanatomic sections used to subdivide the brain.16 A, The parcellation white matter volumes that were significantly larger in children with neurofibromatosis type 1 (NF1) (P<.05) are shaded. B, The parcellation gray matter volumes that were significantly larger in children with NF1 (P<.05) are shaded. The solitary numerals indicate the 16 parcellation regions.
|
|
|
INTELLECTUAL PERFORMANCE
Each subject’s intellectual performance was assessed using the full version of the Wechsler Intelligence Scale for Children–Third Edition (WISC-III). The WISC-III yielded measures of full-scale IQ, verbal IQ, and performance IQ.17
STATISTICAL ANALYSIS
Paired t tests and Pearson  2 tests were used to compare the children with NF1 and the controls for demographic variables, brain volume measurements, and intellectual performance. SAS statistical software version 9.1 (SAS Institute Inc, Cary, NC) was used to analyze cortical GM and WM in relation to 16 parcellation regions in a repeated-measures analysis of variance, with group (NF1 vs control) and sex as between-subjects variables and with 3 within-subjects variables (2 levels of laterality, 4 levels of the anteroposterior plane, and 2 levels of the superoinferior plane). Given any significant interaction between group status and any of these within-subjects variables, individual parcellation regions were examined to better define the basis for the differences, with P<.05 indicating statistical significance. A general linear mixed-effects model18 was used to test group differences in the associations between brain volumes and WISC-III full-scale IQ. Under the assumption of equal variances for the children with NF1 and the controls, these common variances, as well as the covariance between a child with NF1 and his or her genetic control, were included in the model. Assuming linear relationships between brain volumes and intellectual performance, this model specifies the following 4 fixed effects: the intercept and the slope for children with NF1 and the intercept and the slope for the controls. The estimates of the 2 slopes and the corresponding t tests were of primary interest in these statistical test results.
RESULTS
The mean ages (t=–0.49, P = .63) and sex composition (2 = 0.05, P = .82) of the groups were comparable, and maternal education was not significantly different between the groups (2 = 1.63, P = .80). The matched pairs consisted of 20 full-sibling pairs, 9 half-sibling pairs, and 10 second-degree relative pairs, all first cousins.
BRAIN VOLUMES
Right and left hemisphere volumes, total GM volume, and total WM volume were significantly greater in children with NF1 (Table 1). The greatest percentage difference between the children with NF1 and their close relatives was in WM volume in both hemispheres. The greater WM volume was further reflected in a significantly larger mean area of the corpus callosum in the children with NF1. We did not find a differential effect for right vs left hemisphere volumes in the children with NF1.
|
|
|
|
Table 1. WM Volume, GM Volume, and CC Area in Control Subjects and in Children With NF1*
|
|
|
Results of the repeated-measures analysis of variance demonstrated significant interactions between group status and within-subjects variables for GM and WM. For GM, a significant interaction with group was observed for the anteroposterior plane (F3,68 = 9.09), anteroposterior x laterality planes (F3,68 = 8.36), and anteroposterior x superoinferior planes (F3,68 = 10.02) (P<.001 for all). For WM, an interaction with group was observed for the anteroposterior x superoinferior planes (F3,68 = 5.47, P<.001). The significantly larger GM volumes were predominantly posterior to plane 5 (roughly corresponding to the parietal, occipital, and temporal regions of the brain) (Figure), but the significantly greater WM volumes were predominantly anterior to plane 5 (roughly corresponding to the posterofrontal, orbitofrontal, and anterotemporal regions of the brain). The parcellation volumes with greater WM volume were principally anterior to the parcellation volumes greater GM.
The mean ± SD brain volumes of the 25 children with NF1 with hyperintensities (GM, 986.83 ± 21.80 cm3; and WM, 435.15 ± 15.72 cm3) were not significantly different from those of the 11 children without hyperintensities (GM, 938.88 ± 35.65 cm3; and WM, 389.66 ± 23.28 cm3).
INTELLECTUAL PERFORMANCE AND BRAIN VOLUME
The mean WISC-III full-scale, verbal, and performance IQs were significantly lower in children with NF1 (P<.001 for all) (Table 2). Total GM volume but not WM volume was significantly positively related to WISC-III full-scale IQs for the controls, but not for the children with NF1 (Table 3). This positive relationship held for GM volumes in the right and left hemispheres.
|
|
|
|
Table 2. Wechsler Intelligence Scale for Children–Third Edition IQs in Control Subjects and in Children With NF1*
|
|
|
|
|
|
|
Table 3. GM and WM Volume and Wechsler Intelligence Scale for Children–Third Edition Full-Scale IQs in Control Subjects and in Children With NF1
|
|
|
COMMENT
The results of this study confirm previous findings that the larger brain volume of children with NF1 reflects greater GM and WM volumes compared with controls.9 In children with NF1, GM volume was significantly greater in parcellation regions that included most of the parietal, occipital, and temporal lobes, whereas greater WM volume was predominantly in parcellation regions corresponding to the posterofrontal and orbitofrontal lobe and the anterotemporal lobe. We also established for the first time, to our knowledge, that the expected relationship between cortical GM volume and cognitive function is disturbed in children with NF1. Children with NF1 had lower general cognitive function than their relatives, but they did not have the hypothesized inverse relationship between IQ and brain volume. Instead, only the controls showed a significant positive relationship between cortical GM volume and IQ.
The significantly larger brains of children with NF1 comprise greater GM and WM volumes, as demonstrated in the previous study,9 using different volume measurement methods, and as shown in a study4 of boys with NF1 without attention-deficit/hyperactivity disorder. One study3 of brain volume in children with NF1 attributed the greater brain size to a significant increase only in GM volume, and 2 other studies4, 10 attributed the greater brain size to a significant increase in WM volume. These previous studies did not use closely related control subjects as we did,3-4,10 included younger subjects3 or only male subjects,4 or allowed inclusion of subjects with optic glioma.3 The measurements of brain volume in the latter study3 included the brainstem and cerebellum. We found the greatest percentage difference in cerebral WM volume, and post hoc analysis showed that significant changes were predominantly in frontal regions of the brain, as reported by Cutting et al4 in boys with NF1. They found a significant increase in frontal lobe WM volume, and sublobar analysis showed that the significant changes were in the premotor region of the frontal cortex. Contrary to our hypothesis and the findings of Cutting et al, WM volumes in the parietal lobe were not significantly different from those of controls in the present study. Instead, in our study, the GM volumes were significantly greater posteriorly. In the study by Cutting et al,4 GM volume in the frontal lobe was significantly greater than that in controls only in the boys with NF1 without attention-deficit/hyperactivity disorder, and this was attributed to an increase in the prefrontal region of the frontal lobe. We conclude that, when school-aged children with NF1 are compared with family members, their larger brain sizes reflect significantly greater GM and WM volumes, but the greatest difference is in WM volume in the posterior aspect of the frontal lobe. Furthermore, GM volume was significantly larger in the parietal and occipital cortices, areas of importance for understanding the visuospatial deficits that have usually been found in children with NF1.5
Magnetic resonance imaging hyperintensities were not significantly related to any aspect of brain volume or IQ. Other investigators found a significant relationship between hyperintensities and brain volume, but as already noted, those studies3-4 utilized different types of subjects with NF1, and the findings were different. One study3 found an association between thalamic hyperintensities and GM volume, and the other study4 found an association between globus pallidus hyperintensities and WM volume. In view of our negative findings and the inconsistency of the findings in the previous studies,3-4 we conclude that the relationship between T2-weighted hyperintensities and brain volume is unclear and remains of unknown significance.
To our knowledge, our study is the largest study of family members paired with children with NF1 to show reduced IQ among the children with NF1. Previous studies19-23 demonstrated similar findings in children with NF1, but they were smaller and were not pair matched. Our study also indicates that the normal positive relationship between IQ and GM volume6 is absent in children with NF1. Our findings suggest that alteration in brain growth may be a factor in the cognitive deficits among children with NF1. A prospective longitudinal study design would be required to confirm such a hypothesis. Our study clearly raises the possibility that understanding what cellular elements cause the abnormal brain growth in NF1 may lead to a better understanding of the pathophysiology of learning disabilities in children with NF1.
AUTHOR INFORMATION
Correspondence: Robert S. Greenwood, MD, Department of Neurology, The University of North Carolina at Chapel Hill, 3100 Bioinformatics Bldg, Chapel Hill, NC 27514-7025 (greenwor{at}neurology.unc.edu).
Accepted for Publication: April 19, 2005.
Author Contributions: Study concept and design: Greenwood, Tupler, Whitt, Krishnan, and MacFall. Acquisition of data: Greenwood, Tupler, Whitt, Dombeck, Harp, Eastwood, and MacFall. Analysis and interpretation of data: Greenwood, Tupler, Whitt, Buu, Payne, Eastwood, Krishnan, and MacFall. Drafting of the manuscript: Greenwood, Whitt, Buu, and Eastwood. Critical revision of the manuscript for important intellectual content: Greenwood, Tupler, Whitt, Dombeck, Harp, Payne, Eastwood, Krishnan, and MacFall. Statistical analysis: Tupler, Whitt, and Buu. Obtained funding: Greenwood, Tupler, and Whitt. Administrative, technical, and material support: Greenwood, Tupler, Whitt, Dombeck, Harp, Payne, Krishnan, and MacFall. Study supervision: Greenwood, Tupler, Whitt, Payne, and MacFall.
Funding/Support: This study was supported in part by grants R01 NS36829, GCRC RR00046, and K08 MH01460, from the National Institutes of Health, Bethesda, Md.
Acknowledgment: We gratefully acknowledge Chris Byrum, MD, PhD, and Denise Fetzer, MA, for methods development and image processing; and the study coordinators, research assistants, and psychology interns for assistance with subject recruitment, family liaison, data collection, and study coordination.
Author Affiliations: Departments of Neurology (Dr Greenwood and Ms Harp), Pediatrics (Drs Greenwood and Whitt and Ms Harp), and Psychiatry (Dr Whitt), University of North Carolina School of Medicine, and Curriculum of Neurobiology and Neurodevelopmental Disorders Center (Dr Greenwood), The University of North Carolina at Chapel Hill; Departments of Psychiatry and Behavioral Sciences (Drs Tupler and Krishnan and Mss Dombeck and Payne) and Radiology (Drs Eastwood and MacFall), Duke University Medical Center, Durham, NC; and Department of Statistics, University of Michigan at Ann Arbor (Dr Buu).
REFERENCES
| |
1. Riccardi VM. Type 1 neurofibromatosis and the pediatric patient. Curr Probl Pediatr. 1992;22:66-106.
FULL TEXT
| PUBMED
2. Ferner RE, Chaudhuri R, Bingham J, Cox T, Hughes RA. MRI in neurofibromatosis 1: the nature and evolution of increased intensity T2 weighted lesions and their relationship to intellectual impairment. J Neurol Neurosurg Psychiatry. 1993;56:492-495.
FREE FULL TEXT
3. Moore BD III, Slopis JM, Jackson EF, De Winter AE, Leeds NE. Brain volume in children with neurofibromatosis type 1: relation to neuropsychological status. Neurology. 2000;54:914-920.
FREE FULL TEXT
4. Cutting LE, Cooper KL, Koth CW, et al. Megalencephaly in NF1: predominantly white matter contribution and mitigation by ADHD. Neurology. 2002;59:1388-1394.
FREE FULL TEXT
5. North KN, Riccardi V, Samango-Sprouse C, et al. Cognitive function and academic performance in neurofibromatosis 1: consensus statement from the NF1 Cognitive Disorders Task Force. Neurology. 1997;48:1121-1127.
FREE FULL TEXT
6. Wilke M, Sohn JH, Byars AW, Holland SK. Bright spots: correlations of gray matter volume with IQ in a normal pediatric population. Neuroimage. 2003;20:202-215.
FULL TEXT
|
ISI
| PUBMED
7. Reiss AL, Eckert MA, Rose FE, et al. An experiment of nature: brain anatomy parallels cognition and behavior in Williams syndrome. J Neurosci. 2004;24:5009-5015.
FREE FULL TEXT
8. Cody H, Pelphrey K, Piven J. Structural and functional magnetic resonance imaging of autism. Int J Dev Neurosci. 2002;20:421-438.
FULL TEXT
|
ISI
| PUBMED
9. Said SM, Yeh TL, Greenwood RS, Whitt JK, Tupler LA, Krishnan KR. MRI morphometric analysis and neuropsychological function in patients with neurofibromatosis. Neuroreport. 1996;7:1941-1944.
ISI
| PUBMED
10. Steen RG, Taylor JS, Langston JW, et al. Prospective evaluation of the brain in asymptomatic children with neurofibromatosis type 1: relationship of macrocephaly to T1 relaxation changes and structural brain abnormalities. AJNR Am J Neuroradiol. 2001;22:810-817.
FREE FULL TEXT
11. Neurofibromatosis: conference statement: National Institutes of Health Consensus Development Conference. Arch Neurol. 1988;45:575-578.
FREE FULL TEXT
12. Steffens DC, Byrum CE, McQuoid DR, et al. Hippocampal volume in geriatric depression. Biol Psychiatry. 2000;48:301-309.
FULL TEXT
|
ISI
| PUBMED
13. Payne ME, Fetzer DL, MacFall JR, Provenzale JM, Byrum CE, Krishnan KR. Development of a semi-automated method for quantification of MRI gray and white matter lesions in geriatric subjects. Psychiatry Res. 2002;115:63-77.
FULL TEXT
|
ISI
| PUBMED
14. Kikinis R, Shenton ME, Gerig G, et al. Routine quantitative analysis of brain and cerebrospinal fluid spaces with MR imaging. J Magn Reson Imaging. 1992;2:619-629.15. Byrum CE, MacFall JR, Charles HC, et al. Accuracy and reproducibility of brain and tissue volumes using a magnetic resonance segmentation method. Psychiatry Res. 1996;67:215-234.
FULL TEXT
|
ISI
| PUBMED
16. Zipursky RB, Lim KO, Sullivan EV, Brown BW, Pfefferbaum A. Widespread cerebral gray matter volume deficits in schizophrenia. Arch Gen Psychiatry. 1992;49:195-205.
FREE FULL TEXT
17. Wechsler D. Wechsler Intelligence Scale for Children–Third Edition (WISC-III). New York, NY: Psychological Corp; 1991.18. Littell RC, Milliken GA, Stroup WW, Wolfinger RD. SAS System for Mixed Models. Cary, NC: SAS Institute Inc; 1996.19. Dilts CV, Carey JC, Kircher JC, et al. Children and adolescents with neurofibromatosis 1: a behavioral phenotype. J Dev Behav Pediatr. 1996;17:229-239.
ISI
| PUBMED
20. Eliason MJ. Neurofibromatosis: implications for learning and behavior. J Dev Behav Pediatr. 1986;7:175-179.
ISI
| PUBMED
21. Es SV, North KN, McHugh K, Silva MD. MRI findings in children with neurofibromatosis type 1: a prospective study. Pediatr Radiol. 1996;26:478-487.
FULL TEXT
|
ISI
| PUBMED
22. Hofman KJ, Harris EL, Bryan RN, Denckla MB. Neurofibromatosis type 1: the cognitive phenotype. J Pediatr. 1994;124:S1-S8.
FULL TEXT
|
ISI
| PUBMED
23. North K, Joy P, Yuille D, Cocks N, Hutchins P. Cognitive function and academic performance in children with neurofibromatosis type 1. Dev Med Child Neurol. 1995;37:427-436.
ISI
| PUBMED
 CiteULike  Connotea  Del.icio.us  Digg  Reddit  Technorati  Twitter
What's this?
|
