ORIGINAL CONTRIBUTION Brain Morphometric Analysis in Neurofibromatosis 1 Francis J. DiMario, Jr, MD; Gale R. Ramsby, MD; Joseph A. Burleson, PhD Rationale and Objectives: To investigate the rela- Results: Therewere27patients(20boys,aged1.0-17.7years; tionships between brain and skull base growth in pa- mean age, 8.8 years) and 43 controls (22 boys, aged 0.1-17.7 tients with neurofibromatosis 1 (NF1) compared with years; mean age, 5.9 years). The mean ages between groups healthy control subjects using brain magnetic reso- (boys, girls, and totals) were not statistically different. Sig- nance imaging (MRI) for morphometric analysis. nificant differences were appreciated for 6 of 24 measures. Patients with NF1 had a significantly larger bicaudate width Methods: Evaluated patients included children who (P = .002), biatrial width (P,.001), and biparietal diameter underwent T1- and T2-weighted or dual-echo proton (P = .003), but not hemispheric length. They also had sig- density axial and T1-weighted sagittal brain MRI from nificantly increased iter measures (P = .004), descending sig- January 1, 1988, to December 31, 1995. Study subjects moid sinus (P,.001), and an age-specific increase in brain- (n = 27) received a diagnosis of NF1 by accepted stem height (P = .03) not seen in controls. National Institutes of Health clinical criteria and were compared with an age- and sex-matched control group Conclusions: Patients with NF1 experience dynamic (n = 43). Twenty-four predetermined ventricular and changes in brain morphometry, resulting in a predomi- brain parenchymal dimensions and area calculations nant lateral volume expansion of the supratentorial com- were evaluated. Data were analyzed using 2-tailed t tests, partment and an increasing velocity of brainstem growth x2 analysis, analysis of variance, and analysis of covari- as they age. These data underscore brain-region– ance adjusted for age and sex. Correlational analyses specific parenchymal overgrowth potential. with respect to subject type and age were performed separately. Arch Neurol. 1999;56:1343-1346 EUROFIBROMATOSIS 1 (NF1) lencephalywouldappeartobethemostcom- is a relatively common (in- mon identifiable underlying reason for mac- cidence, 1 in 3000) autoso- rocephaly seen in patients with NF1. Ow- mal dominant disorder with ing to the fact that the NF1 gene plays a role aspontaneousmutationrate in regulating cell growth and differentiation, Nof about 60%.1,2 The diagnosis is based on an it is likely that a direct effect on tissue growth individual demonstrating at least 2 of the fol- and ultimate size is linked to tissue-specific lowing 7 clinical criteria: 6 or more cafe´-au- expression of this gene. Since not all cell and lait spots (pubertal-dependent size criteria), tissue types need be dependent on this in- 2 or more neurofibromas or plexiform neu- tracellularregulatorygeneproduct,itislikely rofibromas,axillaryoringuinalfreckling,dis- that not only a brain-specific effect could be tinctive osseous lesions, optic nerve glioma, postulatedbutalsothatbrain-region–specific more than 2 iris Lisch nodules, or a first- effects could be identified. degree relative with NF1.1,2 The NF1 gene on We undertook a morphometric analy- chromosome 17 encodes for the protein sis of the brain to clarify the structural re- neurofibromin.2 Mutations in the NF1 gene lationships of brain and skull base growth result in the production of a nonfunctional with attention to region-specific brain From the Departments of neurofibromin protein or the absence of its growth patterns in patients with NF1 com- Pediatrics (Dr DiMario), expression. pared with healthy control subjects. We ex- Radiology (Drs DiMario and pected that the analysis would reveal a logi- Ramsby), and Behavioral For editorial comment cal sequence of consistently measurable Sciences and Community see page 1322 neural and cranial distortions. These iden- Health (Dr Burleson), tified regions may correlate further to ar- University of Connecticut, Macrocephaly, a common clinical eas of highest potential for the develop- Farmington; Connecticut manifestation in patients with NF1, can be ment of hamartomas or neoplasia. Children’s Medical Center, ascribed to several mechanisms that are not Hartford (Dr DiMario); and the 3-5 University of Connecticut mutually exclusive. This may be second- RESULTS Health Center, Farmington ary to communicating or obstructive hydro- (Drs DiMario, Ramsby, and cephalus, megalencephaly (large brain), or There were 27 patients with NF1 (20 boys; Burleson). intracranial mass or fluid collections. Mega- aged 1.0-17.7 years; mean age, 8.8 years) ARCH NEUROL / VOL 56, NOV 1999 WWW.ARCHNEUROL.COM 1343 ©1999 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 SUBJECTS AND METHODS Magnetic resonance images were examined by one of us who is a neuroradiologist (G.R.R.) and who was un- aware of the clinical condition of the subjects. All mea- SUBJECTS AND COMPARISONS surements were confirmed by another of us (F.J.D.). Mea- surements were made in millimeters to the nearest single All patients referred to the Neurogenetics Clinic at The Uni- millimeter for all linear measures, and angular measures versity of Connecticut Health Center, Farmington, and were obtained to the nearest single degree. Twenty prede- Connecticut Children’s Medical Center, Hartford, were con- termined ventricular and brain parenchymal dimensions sidered eligible for evaluation. Entry criteria included the avail- were evaluated, including foramen magnum anterior- ability of brain magnetic resonance images (MRI) obtained posterior (AP) diameter in sagittal plane, the sinojugular from January 1, 1988, to December 31, 1995. Patients selected transition zone, and the ascending sigmoid sinus widths for evaluation included study subjects with a diagnosis of NF1 in axial plane. In addition, the sum of the maximum width by National Institutes of Health–accepted clinical and radio- of the left- and right-sided sinojugular transition zones was logical criteria and a control group matched by sex and age. calculated. The product of the maximum AP and lateral Controls were only included if the indication for MRI was for widths of the descending sinuses on the left and right sides the evaluation of headaches or idiopathic seizures and if re- was calculated. These products (left and right) were then sults of the clinical examination and MRI were normal. The summed for a combined total. The predetermined ana- MRIstudywasconsideredcompletewithaT1-andT2-weighted tomical regions examined were identified consistently within 6 or dual-echo proton density axial image set and a T1-weighted the landmark identification key (Figure 1). Lastly, the sagittal image set. Poor-quality studies due to motion or presence or absence of ventriculomegaly was determined other artifacts excluded the potential control from study. In and further quantified as grade 1 (mild), 2 (moderate), or patients with NF1 who subsequently underwent neuro- 3 (severe) in the NF1 group. This was done by comparing surgicalinterventions,onlypreinterventionimageswereevalu- all axial T1-weighted images together. Controls had no ven- ated. The presence or absence of malformations, high-signal– triculomegaly by definition. intensity lesions, and masses were noted by location. We were interested in identifying regions of growth or size differ- ence that would be appreciated readily, as they related to ob- STATISTICAL ANALYSIS servable changes in skull growth. Previous studies indicated increased skull growth in the region of the sella turcica and Data were analyzed using the statistical software SPSS for in each skull dimension (height, length, and width).5 Pre- Windows Release 5.0 (SPSS, Inc, Chicago, Ill, 1992). Analy- vious studies demonstrated excellent correlation between sis of covariance with adjustments for age and sex was used skull base distortion and brain morphometry in patients to assess mean differences between groups, whereas Pear- with achondroplasia using these measures and methods,6 son correlational analysis was used to assess the relation- and we therefore applied this approach in our present study. ship between measures. and 43 controls (22 boys; aged 0.1-17.7 years; mean age, Although 15 of the 27 patients with NF1 had MRI high- 5.9 years). The mean ages between groups (boys, girls, signal T2-weighted lesions identified during this study, there and totals) were not statistically different. were no significant differences noted on any measure when Significant differences were appreciated in 6 of 24 separate group analyses of the patients with NF1 with and measures (Table). Patients with NF1 had a signifi- without high-signal lesions were performed independently. cantly enlarged bicaudate width (P = .002), biatrial width We studied no patients with identified tumors or masses. (P,.001), and biparietal diameter (P = .003), but not hemispheric length. Patients with NF1 also had signifi- COMMENT cantly increased iter measures (P = .004) and AP dimen- sion of the sigmoid sinus (P,.001). Finally, there was Our analysis of morphometric measures showed a series an interaction between patient group and age such that of predictable consequences of brain growth in patients with an age-dependent increase in brainstem height was seen NF1. The well-known macrocephaly identified in pa- among patients with NF1 but not in controls (P = .03). tients with NF1 can be quantified as having a primary ba- For the NF1 group, the older the patient, the