OBJECTIVE: Cerebrovascular abnormalities are serious but underrecognized complications of neurofibromatosis type 1 (NF1). The aim of this study was to investigate the prevalence, clinical presentation, imaging findings, and prognosis of cerebral arteriopathies in childhood NF1.
METHODS: Patients followed at the NF1 clinic at the Hospital for Sick Children, Toronto, Ontario, Canada, between 1990 and 2007 were studied. Patients with confirmed NF1 diagnosis and neuroimaging results were included. All neuroimaging studies were reviewed for the presence of arteriopathy by 2 study pediatric neuroradiologists blinded to clinical information. Clinical records of children with cerebral arteriopathy were reviewed.
RESULTS: Among 419 children with confirmed NF1, 266 (63%) received neuroimaging. Among children with neuroimaging results, 17 had cerebral arteriopathy (minimum prevalence rate of 6%). Among the 35 patients who received magnetic resonance angiography (MRA), arteriopathy was more common in patients with NF1 with optic gliomas (11 of 21) compared with those without optic glioma (4 of 14). Forty-seven percent of children developed focal deficits months to years after the diagnosis of the arteriopathy. Follow-up at a mean of 7 years after diagnosis of arteriopathy showed that 35% (6 of 17) had progressive arteriopathy requiring revascularization surgery. Seven patients received aspirin for primary stroke prevention. On retrospective review of imaging studies, a mean delay of 51 months to clinical radiographic reporting of these findings was observed.
CONCLUSIONS: The prevalence of cerebral arteriopathy in children with NF1 in this study was at least 6% and was associated with young age and optic glioma. Arteriopathy causes stroke with resultant neurologic deficits. Medical and/or surgical interventions may prevent these complications. Therefore, the addition of vascular imaging (MRA/conventional angiography) to brain imaging studies for early detection of arteriopathy should be considered for children with NF1, particularly young patients with optic glioma.
Neurofibromatosis type 1 (NF1) is a multisystem, progressive autosomal dominant genetic disorder that affects ∼1 in 3500 individuals.1–3 The diagnosis of NF1 may be established according to the criteria developed by a National Institutes of Health Consensus Development Conference.4 The diagnostic criteria, potential for serious complications, and recommendations for health supervision of children with NF1 have been outlined by the American Academy of Pediatrics Committee on Genetics.5
Case reports6 and 2 case series3,7 of cerebral arteriopathies in children with NF1 have been noted in the literature. It has been suggested that NF1-associated cerebral arteriopathies may be underrecognized, because not all children with NF1 undergo neuroimaging, and of those that do, not all undergo magnetic resonance angiography (MRA).3 The aim of our study was to investigate the prevalence, clinical presentation, management, and prognosis of cerebral arteriopathies in a tertiary care referral center population of children with NF1.
The Pediatric Neurofibromatosis Clinic at the Hospital for Sick Children, Toronto, Ontario, Canada, provides health supervision for children with confirmed or probable NF1. Children are referred to the clinic by clinical geneticists, primary pediatricians, and a diverse group of pediatric and surgical subspecialists. It provides multidisciplinary care for this patient population and refers them to pediatric subspecialties such as neuroophthalmology, neurosurgery, and neurology as clinically needed. The clinic practices targeted rather than universal neuroimaging screening, in keeping with current recommendations from the National Neurofibromatosis Foundation Optic Pathway Task Force8 and the American Academy of Pediatrics Committee on Genetics.5 The study population included all children who met National Institutes of Health diagnostic criteria4 for confirmed NF1 from 1990 to 2007 and had neuroimaging performed (MRI with or without MRA, computed tomography with or without angiogram or conventional angiography). The clinical information abstracted from the health record of each child included gender, age at NF1 diagnosis, clinical symptoms as reported by the primary pediatrician in the NF1 clinic and/or the pediatric neurologist, presence of neuroimaging, age at first neuroimaging, age at which arteriopathy was diagnosed, medical or surgical therapies for arteriopathy, duration of clinical follow-up, and treatment complications that were recorded. Patients who had a previous history of intracranial radiotherapy were excluded from the study. Therefore, none of the vascular changes reported in this study could be attributed to radiation therapy.
All neuroimaging studies for children with ≥1 neuroimaging studies were reviewed independently by 2 pediatric neuroradiologists (Drs Rea and Armstrong) blinded to the clinical presentation. During the study period, an MRA was included for children with MRI findings suggestive of arteriopathy who were scheduled for follow-up neuroimaging. Arteriopathy was defined as any abnormality in the intracranial arterial system that could not be defined as a normal variant. Specific morphologic findings that were examined included any abnormal change in the caliber of the intracranial arterial system, arterial occlusion, any abnormal collateral arterial supply, or the presence of any intracranial aneurysm. In cases of disagreement, consensus opinion based on discussion between the 2 neuroradiologists was reported. Information regarding location, type, and severity of arteriopathy, presence of diffuse leptomeningeal increased signal on fluid-attenuated inversion recovery (FLAIR) MRI sequence representing engorged pial network via leptomeningeal anastomosis (the “ivy” sign), which is well described in moyamoya arteriopathy, 9 and the presence of cerebral infarct were documented. A review of the original neuroimaging interpretation, compared with the current interpretation, identified missed or delayed diagnosis of arteriopathy. Nonvascular findings such as focal areas of abnormal signal intensity (FASI), optic pathway glioma, and sphenoid wing dysplasia impacting on the course of the internal carotid or vertebrobasilar arteries were also documented. This study protocol was approved by the institutional research ethics board.
Descriptive statistics were used in the study. Two-group comparison was conducted in the cohort of children with NF1 who had MRA studies between the group of patients with optic glioma and those without optic glioma. The presence of arteriopathy was compared between the 2 groups by using the Fisher's exact test.
During the study interval, 419 children with a confirmed diagnosis of NF1 were screened. Of these, 266 (63%) had undergone neuroimaging and were included in the study, and 112 of these children had optic glioma. Review of indications for neuroimaging in these 266 children (Table 1) did not identify any bias in their symptomatology to suggest that they were more likely to have an arteriopathy identified by neuroimaging. The neuroimaging studies included MRA for 35 of 266 children. Of the 266 children who had undergone neuroimaging, 17 (6% [95% confidence interval: 3.8% to 9.5%]) had an arteriopathy identified by MRI and MRA (15 of 17) or MRI only (2 of 17).
The clinical features of the 17 children with arteriopathy are summarized in Table 2. Arteriopathy was nearly twice as common in boys compared with girls (11 boys versus 6 girls). The age at diagnosis of NF1 ranged from birth to 11 years with a median age of 2 years. Nine of the 17 (52%) children had no focal neurologic deficit. For these 9 patients, indications for neuroimaging included diagnosis or follow-up for optic glioma, headache, amblyopia, proptosis, and altered visual acuity. The majority of the patients with focal neurologic deficits developed their symptoms after the diagnosis of their arteriopathy (Fig 1). Seven children were treated with aspirin as primary stroke prevention, 9 children were untreated, and 6 children have undergone indirect revascularization surgery (pial synangiosis) except patient 8 who underwent both direct and indirect revascularization.
Cerebrovascular Findings on Imaging
All of the 266 patients with NF1 included in the study had MRI performed. Arteriopathy was observed on MRI and then the diagnosis was confirmed by MRA in 15 of 17 patients. Arteriopathy was diagnosed by standard nonangiographic MRI sequences only in 2 of 17 patients. The first neuroimaging study was performed at ≤6 years of age in 14 of 17 patients (82%; Table 3). Neuroimaging findings for the 17 patients are summarized in Table 3, and a representative illustration of these findings is shown in Fig 2. Stenotic lesions were noted in 16 of 17 children including unilateral moyamoya (12), bilateral moyamoya (1), and supraclinoid internal carotid artery stenosis (3). One child had a 1-cm fusiform aneurysm in the cavernous right internal carotid artery without evidence of moyamoya. Eleven of the 16 (69%) children with stenotic lesions demonstrated the “ivy sign” on FLAIR or T1 postgadolinium sequences (Fig 2A). In 7 children, the arteriopathy was diagnosed at the time of the initial imaging at a median age of 5.2 years (range: 1.8–11.2 years). In 10 children, the arteriopathy was diagnosed at the time of the current study at a median age of 5.3 years (range: 1.8–13.9 years). The mean (SD) delay between onset of vessel flow signal abnormalities on standard T2-weighted imaging and prospective documentation of these findings was 51 ± 33.3 months (range: 13–144 months).
Optic glioma was present in 13 of the 17 (76%) patients with arteriopathy. Of the 266 patients who received neuroimaging, 35 patients had MRA including 21 with and 14 without optic glioma. Among the 35 children who had MRA, arteriopathy was more common in patients with NF1 with optic gliomas (11 of 21) compared with those without optic glioma (4 of 14) in the cohort of patients with MRA but did not reach statistical significance (P = .10).
The mean (SD) duration of radiology follow-up was 7 ± 3.5 years (range: 1–12.7 years). Twelve of the 17 (70%) children with arteriopathy remained stable without radiologic evidence of progression. Six of the 17 (35%) children demonstrated disease progression, of whom 4 showed progressive vessel stenosis, 1 had a new infarct (patient 17), and 1 had severe bilateral nonprogressive moyamoya on serial MRA assessments but had worsening of symptoms (patient 14). In our institution, patients are candidates for revascularization surgery if symptoms are worsening, if progressive vessel narrowing is found during radiologic follow-up, and/or cerebral perfusion is significantly reduced shown by a cerebrovascular reactivity (CVR) study (Fig 3). All 6 children subsequently underwent revascularization surgery (Table 2). One of the 6 patients underwent revascularization surgery despite having no focal neurologic deficit (patient 5), because she had progressive arteriopathy on serial MRAs over 8 years of follow-up.
To our knowledge, this is the largest case series investigating the prevalence, radiologic findings, and prognosis of cerebral arteriopathy in a pediatric population with NF1. We report a minimum prevalence rate of 6% for childhood NF1-associated cerebral arteriopathy. This is likely an underestimate. First, not all of our patients with NF1 received neuroimaging. Second, even in those who received neuroimaging, vascular imaging was not routinely performed. Unilateral moyamoya was the most common arteriopathy in our study. Of the patients with arteriopathy, 52% did not have a focal neurologic deficit. Most of the symptomatic patients presented with deficits months to years after the diagnosis of cerebral arteriopathy. Arteriopathy was more common in children with NF1-associated optic glioma presenting in the first 6 years of life. Childhood NF1-associated arteriopathy was progressive in more than one third of the patients requiring revascularization surgery. Patients with progressive arteriopathy tend to be symptomatic. The diagnosis of arteriopathy in patients with childhood NF1 was often missed or delayed in almost 60% of patients, highlighting the need for a higher index of suspicion for this complication of NF1.
Arteriopathy has been an underreported potential complication of NF1. Rosser et al3 reported on their experience at the Children's National Medical Center in Washington, DC, of 353 patients with NF1, of whom 316 (90%) underwent MRI as part of routine screening and 8 (2.5%) patients were identified with an abnormality of the cerebrovascular system. Cairns and North7 reported on 698 patients with NF1 at the Children's Hospital at Westmead in Sydney, Australia, of whom 144 (21%) underwent MRI for a clinical indication and 7 (5%) patients were identified with cerebrovascular dysplasia. In our study, 266 of 419 (63%) patients with NF1 underwent neuroimaging for a clinical indication, and 17 (6%) were identified with arteriopathy. Although our study included a selected group of patients, 1 strength was the prospective and blinded assessment of all neuroimaging studies.
Although the number of patients with focal neurologic deficits was much higher in our study (47%) than that reported by Rosser et al3 (12%) and Cairns and North7 (29%), the majority of patients did not have focal deficits as a result of their cerebral arteriopathy in all 3 studies. Therefore, the absence of neurologic deficit does not rule out the presence of arteriopathy in most patients with NF1. Rosser et al3 followed patients over 10 months to 5 years and identified 1 patient with additional strokes. Cairns and North7 followed patients over 4 to 44 months and identified 3 patients with progression of the stenosis. In our study, with follow-up from 1 to 12 years, 6 (35%) patients demonstrated progression, including 5 with vascular progression and 1 with clinical progression.
Sobata et al6 classified the vascular lesions identified in NF1 into 3 types on the basis of morphology: stenotic lesions, aneurysmal lesions, and mixed stenotic and aneurysmal lesions. Sixteen of the 17 (94%) children included in our study demonstrated stenotic-type vasculopathy with 1 aneurysmal type. Our findings are in keeping with the published review by Sobata and colleagues6 where 89% of children had stenotic lesions with the remainder demonstrating only aneurysms.
The pathophysiology of NF1 vasculopathy is not yet fully understood. Neurofibromin, the protein product of the tumor suppressor NF1 gene, has been shown to be expressed in mammalian endothelium.10 Loss of neurofibromin function may cause smooth muscles to proliferate causing vascular stenosis.11
The development of clinical neurologic symptoms caused by ischemic changes secondary to childhood NF1-associated arteriopathy may occur months to years after the diagnosis of the arteriopathy, which may become progressive with or, rarely, without clinical symptoms. This raises the possibility that early treatment intervention with an antiplatelet agent such as aspirin can lower the risk of ischemic changes as a complication of the arteriopathy. This study was not powered to answer this question given that 5 of 7 of the treated patients were already symptomatic when treatment was initiated. However, it is conceivable that an antiplatelet agent may reduce the risk of clot formation in narrowed arteries where high flow is the main mechanism of platelet-rich clot formation.12 A prospective study to investigate the effect of antiplatelet agents on the prognosis of moyamoya would shed light on the role of antiplatelet agents in NF1-associated cerebral arteriopathy in children.
This study indicated that factors that may predict the presence of arteriopathy include young age and concurrent optic glioma. Arteriopathy was more common in patients with optic glioma compared with those without optic glioma. We have shown in our cohort that NF1-associated arteriopathy tends to develop during the first 6 years of life. The mean age for the identification of cerebrovascular abnormalities was 7.3 years in the study by Rosser et al3 and 6.8 years in the study by Cairns and North.7 It is notable that both NF1-associated arteriopathy and optic glioma8 seem to develop during the first 6 years of life. However, the increased prevalence of arteriopathy in patients with optic glioma could be a result of selection bias where these patients undergo high-resolution MRI through the optic nerve, chiasm, and tracts as well as the internal carotid artery and its terminal branches. Interestingly, most optic gliomas in the arteriopathy group (6 of 13) were more extensive (category C of Dodge classification13) and involved the posterior optic tracts and hypothalamus, whereas the majority (58 of 99) of optic gliomas in the nonarteriopathy group were more anterior and involved the optic nerve (category A13). These results may suggest that the presence of a more extensive optic glioma may trigger the formation of specific growth factors leading to abnormal vascular endothelial proliferation.
Early identification of cerebral arteriopathy and close follow-up of its progression by neuroimaging may lead to early medical or surgical intervention and prevention of significant neurologic complications. This raises the potential for universal neuroimaging of young children with confirmed or probable NF1, which contrasts with current published guidelines for the diagnosis and management of patients with NF1.5,14 However, the management of cerebrovascular disease as a specific potential complication of NF1 was not specifically addressed by these guidelines, and unlike optic glioma, clinical examination may not identify children at risk for cerebral arteriopathy. In September 2008, the American Heart Association published a scientific statement on the management of stroke in infants and children and recommended that routine vascular screening may be considered in individuals with relatively common and high-risk disorders such as NF1, Down syndrome, and sickle cell disease.15
In view of the relatively high incidence of NF1 affecting 1 in 3500 individuals, providing MRI/MRA for young children diagnosed with NF1 at a national level may be problematic in terms of feasibility and cost. In our institution, performing an MRA as 1 additional sequence would increase the required scan and anesthesia time by ∼12%. With increasing numbers of scans, routine performance of MRI/MRA for asymptomatic children may slowly begin to impact on the ability to deliver timely imaging for all children who require an MRI. However, the observation that a significant number of asymptomatic children with NF1 may have an unidentified arteriopathy can not be ignored. NF1-associated arteriopathy may silently progress or continue to be undetected until it presents with arterial ischemic stroke leading to lifelong morbidity for the child and serious burden to parents and society who may have to provide lifelong care to a child with stroke. Unlike asymptomatic optic glioma, early detection of asymptomatic arteriopathy by neuroimaging may impact the management of the patient with NF1 by initiating primary stroke prevention in the form of antiplatelet therapy. The financial cost of frequent neuroimaging for patients with NF1 may be significant. Data regarding medical treatment costs for adult patients with NF1 confirmed increasing costs with increased interventions and disease severity.16 No specific data are available on imaging costs for children with NF1. The cost of care for a child with a stroke in only the first year after diagnosis was estimated to be $42 338.17 Therefore, some of these new costs could potentially be offset in reduced costs for long-term care for children with lifelong disability from stroke.
Delays in identifying NF1 arteriopathy have previously been described in large series7 and in individual case reports.10 Difficulties in radiologic interpretation arise when white matter signal abnormalities arising from ischemia are incorrectly assumed to be caused by FASI seen in the cerebellar peduncles, brainstem, basal ganglia, thalami, internal capsules, and splenium of prepubescent children with NF1. Typically, FASI are represented by focal areas of hyperintensity on both T2-weighted and FLAIR images without any associated mass effect or contrast enhancement. White matter lesions arising consequent to cerebral ischemia arise in atypical locations for FASI and are more likely to be hyperintense on FLAIR and be associated with the presence of the ivy sign. The imaging findings associated with NF1-related arteriopathy may be subtle and can easily be overlooked. This is particularly challenging in a complex condition such as NF1 because of the need to examine multiple areas of parenchymal brain abnormality and compare with previous examinations. In the absence of an MRA, for all children with NF1 undergoing neuroimaging, careful assessment of the standard imaging sequences, vessel caliber and continuity on standard T2-weighted sequences, presence of collateral vessels in the expected distribution of a known vessel, or presence of the ivy sign on FLAIR or T1 postcontrast images is advocated. When standard magnetic resonance techniques raise the possibility of arteriopathy, magnetic resonance or conventional angiography should be subsequently undertaken.
We recommend that all children with NF1 who are undergoing neuroimaging should have MRI/MRA included in their studies, because NF1 arteriopathy may be asymptomatic and potentially progressive. MRI/MRA should be considered in children with NF1 who are 6 years of age or younger, particularly those with optic glioma at the time of diagnosis. These children should be closely followed with periodic MRI/MRA for early detection of cerebral arteriopathy and should start aspirin or other antiplatelet medication after an arteriopathy is detected. If early signs of progressive arteriopathy are found clinically or on imaging, patients may be considered for surgical intervention. Finally, prospective study of the clinical and radiographic progression of cerebral arteriopathy will help guide recommendations regarding the need for universal screening of all children with NF1.
- Accepted April 10, 2009.
- Address correspondence to Rand Askalan, MD, PhD, Hospital for Sick Children, Division of Neurology, 555 University Ave, Toronto, Ontario, Canada M5G 1X8. E-mail:
Financial Disclosure: The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject:
Cerebral arteriopathy is a rare but underrecognized complication of childhood NF1. We investigated the prevalence, clinical presentation, management, and prognosis of cerebral arteriopathy in a tertiary care referral center population of children with NF1.
What This Study Adds:
To our knowledge, this is the largest case series investigating the prevalence, radiological findings, and prognosis of cerebral arteriopathy in a pediatric population with NF1. The study recommendations will have significant impact on the management of these patients.
- ↵DeBella K, Szudek J, Friedman JM. Use of the National Institutes of Health criteria for diagnosis of neurofibromatosis 1 in children. Pediatrics.2000;105 (3 pt 1):608– 614
- ↵Rosser TL, Vezina G, Packer RJ. Cerebrovascular abnormalities in a population of children with neurofibromatosis type 1. Neurology.2005;64 (3):553– 555
- ↵Hersh JH; American Academy of Pediatrics, Committee on Genetics. Health supervision for children with neurofibromatosis. Pediatrics.2008;121 (3):633– 642
- ↵Cairns AG, North KN. Cerebrovascular dysplasia in neurofibromatosis type 1. J Neurol Neurosurg Psychiatry.2008;79 (10):1165– 1170
- ↵Flaumenhaft R, Tanaka E, Graham GJ, et al. Localization and quantification of platelet-rich thrombi in large blood vessels with near-infrared fluorescence imaging. Circulation.2007;115 (1):84– 95
- ↵Taylor T, Jaspan T, Milano G, et al. Radiological classification of optic pathway glioms: experience of a modified functional classification system. Br J Radiol.2008;81 (970):761– 766
- ↵Ferner RE, Huson SM, Thomas N, et al. Guidelines for the diagnosis and management of individuals with neurofibromatosis 1. J Med Genet.2007;44 (2):81– 88
- ↵Roach ES, Golomb MR, Adams R, et al; Council on Cardiovascular Disease in the Young. Management of stroke in infants and children: a scientific statement from a Special Writing Group of the American Heart Association Stroke Council and the Council on Cardiovascular Disease in the Young. Stroke.2008;39 (9):2644– 2691
- ↵Lo W, Zamel K, Ponnappa K, et al. The cost of pediatric stroke care and rehabilitation. Stroke.2008;39 (1):161– 165
- Copyright © 2009 by the American Academy of Pediatrics