Nonperioperative strokes in children with central nervous system tumors

1094

Nonperioperative Strokes in Children with Central Nervous System Tumors Daniel C. Bowers, M.D.1 Arlynn F. Mulne, M.D.2 Joan S. Reisch, Ph.D.3 Roy D. Elterman, M.D.2 Louis Munoz, M.D.2 Timothy Booth, M.D.4 Kenneth Shapiro, M.D.2 Deborah L. Doxey, Ph.D.2 1

Department of Pediatrics, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas.

2

Neuro-Oncology Program, Children’s Medical Center of Dallas, Dallas, Texas.

3

Department of Academic Computing Services, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas.

4

Department of Radiology, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas.

BACKGROUND. Nonperioperative strokes are rare yet potentially devastating events for children with central nervous system (CNS) tumors. The incidence of and risk factors for nonperioperative strokes in children with CNS tumors is unknown. METHODS. The authors performed a retrospective review of children from their institution with CNS tumors. The incidence of stroke in the nonperioperative period and the influence of patient demographic factors, coexisting genetic diseases, tumor type, and treatment modality on the subsequent occurrence of a stroke were determined. RESULTS. Eight hundred seven consecutive patients from the authors’ institution with CNS tumors were observed for a combined 3224 nonperioperative years. Thirteen patients (1.6%) had a nonperioperative stroke, for an incidence of 4.03 strokes/1000 years of nonperioperative patient follow-up. Eight patients were males, and the median age at diagnosis of a CNS tumor was 4.8 years (range, 0.3–18.6 years). The median duration from diagnosis of a CNS tumor until the occurrence of stroke was 2.3 years (range, 0.3–15.8 years). Among numerous potential risk factors individually examined by chi-square analysis, only treatment with radiation therapy was associated with the subsequent development of a stroke (chi-square, P ⫽ 0.007). By logistic regression analysis, treatment with radiation therapy and a diagnosis of an optic pathway glioma were the only statistically significant variables associated with a stroke. CONCLUSIONS. Strokes are much more common among children with CNS tumors. Children treated with radiation therapy and those with optic pathway gliomas have a higher association with the occurrence of a subsequent nonperioperative stroke. Because children with optic pathway gliomas may be at particularly high risk of stroke after radiation therapy, the desired beneficial therapeutic effects of irradiation must always be weighed against its potentially adverse effects, including stroke. Cancer 2002;94:1094 –101. © 2002 American Cancer Society. DOI 10.1002/cncr.10353

KEYWORDS: cerebrovascular accidents, brain tumors, optic pathway gliomas, radiation therapy, late effects.

Presented at the International Symposium on Pediatric Neuro-Oncology. San Francisco, California, June 9 –13 2000. The authors thank the Children’s Brain Tumor Foundation of the Southwest and the Children’s Cancer Fund of Dallas for their generous financial contributions. Received March 13, 2001; revision received October 12, 2001; accepted October 24, 2001. © 2002 American Cancer Society

S

trokes are rare yet potentially devastating events during childhood. The incidence of stoke during childhood in the United States is 2.7–3.3 strokes per 100,000 patient years.1,2 Strokes are believed to be more common among children with central nervous system (CNS) tumors.3–18 However, existing reports of stroke among children with CNS tumors do not include a control study population and therefore do not allow for the determination of the incidence of stroke nor identify risk factors for their occurrence. The objectives of this report are to identify the previously unreported incidence of stroke in children with CNS tumors and to identify significant risk factors for their occurrence. We reviewed the

Strokes in Pediatric Central Nervous System Tumor Patients/Bowers et al.

medical records of all patients from our institution with CNS tumors and identified those who also experienced strokes. We then determined the incidence of and risk factors for strokes in this population. Identification of risk factors for stroke may facilitate better understanding of the pathogenesis of stroke and perhaps influence treatment decisions for children with CNS tumors.

1095

enced a stroke were reviewed for the following: age at diagnosis, gender, presence of coexisting phakomatoses (e.g., neurofibromatosis, tuberous sclerosis), tumor type, and age at treatment with surgery, radiation therapy, and chemotherapy. All available radiographic examinations of patients who experienced a stroke also were reviewed. Finally, each patient’s event free survival (EFS) and overall survival (OS) also were recorded.

METHODS Review of Medical Records

Statistical Methods

The medical records were reviewed of all patients evaluated at our institution between January 1, 1985 and January 1, 2000 with the diagnosis of a CNS tumor and entered into our institution’s neurooncology database. The electronic pediatric neurooncology database includes summaries of patients’ demographic data, medical histories, physical exams, treatment, summaries, imaging studies, and follow-up evaluations. Patient follow-up was determined as the duration of time from diagnosis of a CNS tumor to the date of last contact or patient death. Nonperioperative patient follow-up was calculated as patient follow-up not including the immediate 3 months after all tumor biopsies or debulking procedures. Patients with optic nerve gliomas and brain stem gliomas did not need to have a biopsy for diagnosis if the patient’s history, physical exam, and imaging studies were consistent with the diagnosis of these tumors. The database was queried to identify patients who had experienced a cerebrovascular accident, infarction, or stroke. The medical records of identified patients were further reviewed regarding the clinical circumstances of their stroke. The clinical diagnosis of a stroke was based on the occurrence of a new, focal, irreversible neurologic deficit resulting from ischemic damage to the brain. New ischemic brain lesions, identified by magnetic resonance imaging (MRI) with or without angiography, confirmed the physical findings of a new neurologic deficit. Neurologic injury as a result of tumor recurrence or progression was considered and excluded as the cause of new neurologic deficits in all cases. Perioperative strokes, defined as a new ischemic brain lesion occurring less than 3 months after any tumor biopsy or resection, were excluded from this report. Also, clinically silent lacunar lesions, which are believed to be a consequence of radiation-induced microvascular disease within the territories of branching arterioles of the major cerebral arteries, were not examined in this report.19 Finally, we did not consider transient ischemic attacks (TIAs) or reversible ischemic neurologic deficits (RINDs) to be strokes. The medical records of patients who had experi-

The mean and standard deviation were calculated for age at diagnosis and duration of follow-up. Frequency distributions for each of the categoric measures were determined (e.g., tumor type, treatment modality). Chi-square contingency analysis was utilized to determine whether there was a relation between the presence of each of the measurements. Stepwise logistic regression analysis was utilized to determine the significant risk factors for subsequent stroke. Possible independent predictors of a stroke included demographic factors, tumor type, and treatment modality. Because the number of strokes was small, we utilized a liberal entry criteria (P ⫽ 0.10) for the inclusion of factors in the statistical model. All analyses were conducted using a SAS statistical package (version 6.12, SAS Institute, Inc. Cary, NC, 1996).

RESULTS Patient Population The medical records of 807 consecutive patients age less than 18 years who were evaluated by the NeuroOncology Program at our institution from the years 1985 to 2000 were reviewed. Median age at diagnosis of this population was 6.0 years (range, 0 –18 years) with 54.9% (443 of 807) being male. The patients’ age, gender, coexisting diseases, tumor histologies, and treatment are summarized in Table 1. Recent documentation of medical status or details regarding patient death were available for 686/807 (83.8 %) patients. The remaining patients were considered lost to follow-up. A combined 3533 years from patient’s diagnosis to date of last contact was available for all patients. A combined 3224 years of nonperioperative patient follow-up was available for all patients. The mean duration of nonperioperative follow-up per patient was 4.6 years (standard deviation, ⫾4.13).

Description of Patients with Strokes Thirteen patients (8 males) with CNS tumors were identified who also had experienced a nonperioperative stroke (Table 2). The incidence of nonperioperative stroke among children with CNS tumors was 4.03 strokes/1000 years of nonperioperative patient follow-

1096

CANCER February 15, 2002 / Volume 94 / Number 4

TABLE 1 The Demographic Features, Coexisting Diseases, Tumor Type, and Treatment of Patient Population of Children with CNS Tumors and Affected Patients Who Also Experienced a Stroke Characteristic

n

Percentage

Affected patients

Percentage of population

Total Demographic features: Male gender Age ⬍5 ys Coexisting diseases Neurofibromatosis Tuberous sclerosis Tumor type Low-grade glioma Medulloblastoma/PNET Optic pathway glioma Ependymoma Brain stem glioma Ganglioglioma High-grade glioma Craniopharyngioma Oligodendroglioma Germ cell tumor Other Treatment Surgery Chemotherapy Radiation therapy

807

100

13

1.6

443 289

54.9 35.8

8 7

1.8 2.4

0.939 0.433

45 9

5.6 1.1

1 0

2.2 0.0

0.737 0.699

202 129 72 57 50 42 40 37 23 17 138

25.7 16.1 8.9 7.2 6.2 5.2 5.0 4.6 2.9 2.4 17.3

2 2 3 2 1 0 0 1 0 0 2

1.0 1.5 4.2 3.4 2.0 0 0 2.7 0 0 1.4

0.135 0.943 0.071 0.249 0.821 0.684 0.406 0.589 0.534 0.572 0.850

721 363 385

88.2 45.0 47.7

12 8 11

1.7 2.2 2.9

0.727 0.076 0.007

P values

CNS: central nervous system; PNET: peripheral neuroectodermal tumor.

up. No patients had more than one stroke. The patients’ median age at diagnosis of a CNS tumor was 4.8 years (range, 0.3–18.6 years). The median duration of time from diagnosis of a CNS tumor until the occurrence of a stroke was 2.3 years (range, 0.3–15.8 years). One patient had neurofibromatosis. All patients presented with new onset of a focal neurologic deficit, seizures, or altered mental status at time of their stroke. As part of their initial medical evaluation for stroke, none of the patients were found to have hypotension, sepsis, metabolic abnormalities, or other acute insults as the cause of their stroke. Although nine patients (Patients 1, 2, 5, 6, and 8 –12) had persistent tumor at the time of their stroke, tumor recurrence or progression was excluded as the cause of new neurologic deficits in all cases. Six of 13 patients had evidence of endocrine dysfunction before their stroke, including hypothyroidism (4 patients), precocious puberty (3 patients), growth hormone deficiency (2 patients), and diabetes insipidus (2 patients). Five of nine patients who had neurocognitive evaluations had deficits identified. Finally, a second malignant neoplasm was found in one patient (Patient 7) who was treated with radiation therapy. A malignant peripheral nerve sheath tumor within the radiation field was

diagnosed in this patient 10 years after the diagnosis of his primary tumor. Ten of 13 patients’ strokes were in a distribution consistent with thrombosis of either the internal carotid artery or major cerebral artery (Table 2); 9 of the major arterial thromboses were within the patient’s radiation fields. The remaining three patients’ cerebral infarctions were in a watershed distribution. After their stroke, four patients (Patients 4, 5, 6, and 13) were evaluated by magnetic resonance angiography (MRA) of the cerebral vessels, and one additional patient (Patient 10) was evaluated by both MRA and cerebral arteriography. Four of five patients (Patients 4, 5, 6, and 10) were found to have obstructive vasculitis with multiple areas of high-grade stenosis of the internal carotid arteries, large cerebral arteries and proximal, primary branches of the large cerebral arteries. Magnetic resonance angiography revealed decreased blood flow through the right medial and lateral lenticulostriate arteries in the remaining patient (Patient 13) who had a stroke in an anterior watershed distribution. Patient evaluations for the presence of a hypercoaguable state were performed inconsistently, reflecting the treating physician’s clinical preferences and the continuing evolution of the understanding of

TABLE 2 Summary of Patients with CNS Tumors Who Subsequently Experienced a Stoke Patient no.

Gender

Age at diagnosis (yrs)

1

M

1.7

2

M

3

Treatment Surgery

Radiation therapy

Chemotherapy

STR

2700 cGy, focal

Yes

11

Right parietal lobe

None

NA

41a

10.4

Anaplastic ependymoma PNET

STR

Yes

6

Right PCA

None

NA

0.4a

M

3.1

Craniopharyngioma

GTR

3520 cGy brain, 4240 cGy spine, 5330 cGy posterior fossa None

None

6

HT, DI

None

3⫹

4

F

4.7

STR

5400 cGy, focal

Yes

27

7⫹

F

0.3

STR

5220 cGy, focal

None

190

Left MCA

Yes

0.6⫹

6

F

5.7

STR

5120 cGy, focal

Yes

54

Right PCA

HT, GH, DI, precocious puberty HT, precocious puberty GH

Yes

5

None

10⫹

7 8

M M

18.6 5.7

STR STR

10 33

Left MCA Left PICA

HT, DI None

None NA

124⫹ 7a

M

15.5

Bx

4500 cGy, focal 6960 cGy, focal, hyperfractionated None

None Yes

9

None

66

Left ACA

None

None

4⫹

10

F

3.8

None

4760 cGy, focal

None

72

Right MCA

Precocious puberty

Yes

109a

11

M

2.7

STR

109

Right parietal lobe

None

Yes

14⫹

F M

5.9 4.8

3000 cGy brain, 3600 cGy spine, 5570 cGy to tumor 7020 cGy, focal 3520 cGy brain and spine, 5320 cGy posterior fossa

Yes

12 13

Low-grade glioma (right temporal lobe) Optic pathway glioma Optic pathway glioma Pituitary adenoma Ependymoma (posterior fossa) Thalamic mass (presumed lowgrade glial neoplasm) Optic pathway glioma (neurofibromatosis) Low-grade glioma (brain stem) Brain stem glioma Medulloblastoma

Right temporal hemorrhagic infarction and inferior watershed infarction Right MCA

Yes Yes

9 4

Watershed distribution Frontal watershed distribution

None None

NA Yes

1.1a 1.2⫹

STR GTR

Distribution of stroke

Endocrinopathies

Neurocognitive deficits

Survival after stroke (mos)

CNS: central nervous system; M: male; STR: subtotal resection; cGy: centigray; NA: not assessed; PNET: primitive neuroectodermal tumor; PCA: posterior cerebral artery; GTR: macroscopic gross total resection; HT: hypothyroidism; DI: diabetes insipidus na; not assessed; F: female; GH: growth hormone deficiency; MCA: middle cerebral artery; Bx: biopsy; PICA: posterior inferior cerebellar artery; ACA: anterior cerebral artery a Patient died.


Primary tumor

Diagnosis (mos) to stroke

1097

1098


hypercoaguable states. All patients who were evaluated for the presence of a hypercoaguable state had normal evaluations. The most common primary tumor diagnosis was an optic pathway glioma (n ⫽ 3). Other tumor diagnoses included low-grade glioma (n ⫽ 2), ependymoma (n ⫽ 2), medulloblastoma/peripheral neuroectodermal tumor (PNET; n ⫽ 2), brain stem glioma (n ⫽ 1), prolactin secreting pituitary adenoma (n ⫽ 1), craniopharyngioma (n ⫽ 1), and a thalamic mass lesion for which a biopsy specimen was nondiagnostic for tumor (presumed low-grade glial neoplasm, n ⫽ 1). Initial treatment of the 13 patients had consisted of debulking surgery in 11 patients. The patient who had a stereotactic biopsy of a thalamic lesion and another patient with an optic pathway glioma and neurofibromatosis type 1 were the only patients who did not have a biopsy of their tumors. Two of the 13 patients received neither radiation nor chemotherapy. Eleven patients received adjuvant treatment with radiation therapy (craniospinal, 4; focal, 7), and 8 received chemotherapy (nearly all chemotherapy protocols contained combinations of alkylating agents, either cisplatin or carboplatin, etoposide, and vincristine). Eight of the 13 patients were alive at a median of 3.8 years (range, 0.6 –12.7 years) after their stroke. Four of five patient deaths were a result of tumor progression at a median of 6.6 months after their stroke (range, 0.4 – 41 months). The cause of death in the remaining patient with an optic pathway glioma 109 months after diagnosis was undetermined. The four patients who died of tumor progression had an ependymoma (two patients), PNET, and brain stem glioma. No patients had an autopsy.

Risk Factors for Strokes in Children with CNS Tumors In the current series, prior radiation therapy was the only univariate risk factor identified for the subsequent development of a stroke in children with CNS tumor (chi-square, P ⫽ 0.007; Table 1). A tumor type of an optic pathway glioma and treatment with chemotherapy were of borderline significance. No other tumor types, coexisting phakomatoses, or treatment modalities were associated with an increased risk of a stroke. Various measurements that were considered as possible independent predictors of stroke were entered into a stepwise logistic regression analysis. The only independent variables found to be associated with stroke were the treatment modality of radiation therapy and tumor type of optic pathway glioma (odds ratios and 95% confidence limits are listed in Table 3). Based on the Hosmer–Lemeshow goodness-of-fit test, the logistic regression fit the data well (P ⫽ 0.308).

TABLE 3 Stepwise Logistic Regression Analysis of Significant Risk Factors for Stroke in Children with CNS Tumors Predictor Treatment modality Radiation therapy Tumor type Optic pathway glioma

Odds ratio

95% confidence interval

6.99

1.52–32.17

4.14

1.08–15.87

CNS: central nervous system.

DISCUSSION Strokes are rare yet potentially devastating events for children. On the basis of existing case reports, there is a suggestion that strokes are more common among children with CNS tumors. These reports have not included a defined study population to serve as a denominator and have not reported the incidence of stroke. In our series, the incidence of nonperioperative stroke among children with CNS tumors was 4.03 strokes/1000 years of nonperioperative patient followup. This incidence is comparable to the incidence of stroke found among patients with sickle cell anemia and more than 1000 times the 2.7–3.3 cerebral infarctions and subarachnoid hemorrhages per 100,000 children per year.1,20,21 To our knowledge, this is the first report of the incidence of nonperioperative stroke among children with CNS tumors and demonstrates that they are much more frequent in this patient population. Adults with malignancies are at increased risk of stroke as a result of a paraneoplastic hypercoaguable state, which may include nonbacterial thrombotic endocarditis (NBTE).22,23 Although NBTE is very rare among children with cancer, they also have a predisposition to strokes. Packer et al. reviewed the records of 700 children with non-CNS tumors over a 4-year period and identified 26 patients who had also suffered a cerebrovascular accident.24 In this series, strokes were more frequent among patients with hematologic malignancies than solid tumors. However, children with CNS tumors were excluded from this series, and few patients were treated with cranial radiation therapy. Most of the strokes were believed to be a direct effect of the malignancy or a consequence of cancer treatment. Strokes occur even more rarely among children without cancer.2,20,25 Lanthier et al. reported 46 ischemic and 21 hemorrhagic strokes among children without cancer.25 The most frequent risk factors in their report were vascular abnormalities, hypercoaguable states, metabolic disorders, and congenital heart disease. The specific hypercoaguable states most frequently identified among children with


thrombotic strokes included the presence of anticardiolipin antibodies, plasminogen deficiency, resistance to activated protein C, protein C deficiency, lupus anticoagulant, and protein S.20 The only significant risk factors for stroke identified in our series, by logistic regression analysis, were treatment with radiation therapy and the presence of an optic pathway glioma. No other patient demographic variables, tumor histologies, treatment modalities, or coexisting phakomatoses were identified as risk factors for stroke. Among previously reported case series, the most frequent associations for nonperioperative stroke include the presence of an optic pathway glioma,4 –7,9 –17,26,27 coexisting neurofibromatosis,5,9,11,14,16,26 and treatment with radiation therapy.4 –7,9 –11,14 –18,26,27 Prior case series describe strokes among children with other types of CNS tumors, after other treatment modalities and in the presence of other phakomatoses. Strokes have been reported among children with medulloblastoma,12,17 craniopharyngioma,8,17 non-Hodgkin lymphoma,17 supratentorial germinoma,17 pituitary adenoma,8 and highgrade glioma.28 Furthermore, other etiologies, including tumor encasement of major cerebral arteries by tumor,3,5,8,29 coexisting tuberous sclerosis,5 surgical manipulation of the carotid artery,29 and treatment with chemotherapy,17 have been proposed as causes of strokes. None of the patients from our series underwent autopsy, and therefore we were unable to examine and characterize any histologic abnormalities of their cerebral vessels. However, the vascular distribution, by MRI, MRA and angiography, supports the strokes being a result of radiation injury to the carotid arteries and large cerebral arteries. This is consistent with Brandt–Zawadzki and coworkers’ description of autopsy findings from two patients with brain tumors treated who subsequently suffered thrombotic strokes after treatment with radiation therapy.30 Both of these patients had occlusive vasculopathy that was restricted to the internal carotid and large cerebral arteries. Histologically, endothelial proliferation and thickening of tunicas intima and muscularis were observed in the patients’ large cerebral arteries. Also, there was redundancy and thickening of the internal elastic lamina. The remaining smaller intracranial vessels, arterioles, capillaries, and veins did not have any identifiable abnormalities. However, radiation therapy was not the cause of all of our patients’ strokes because two patients were never treated with this modality. Also, two patients treated with radiation therapy and one patient who was not treated with radiation therapy had strokes in a “watershed distribution,” which is an atypical distribution for radia-

1099

tion-induced vasculopathy. As a result, we conclude that most of our patients’ strokes were a result of radiation-induced vasculopathy, but that clearly other conditions contribute to the occurrence of a stroke among patients with CNS tumors. Strokes have been reported among patients with optic pathway gliomas and neurofibromatosis type 1, and there is a well-known association of these two conditions.10,11,13,14,16 In this report, we identified the presence of an optic pathway glioma, but not neurofibromatosis, as a significant risk factor for subsequent stroke. We were unable to identify neurofibromatosis as a risk factor for stroke in our series, even though 13 patients with neurofibromatosis also were treated with radiation therapy (data not shown). Moya-moya is a radiographic syndrome resulting from either stenosis or occlusion of the distal internal carotid and proximal large cerebral arteries.3,14,31 A radiographic “moya-moya pattern” of cerebral arterial occlusion has been observed among children with CNS tumors who have experienced either TIAs or strokes.17 Although most patients with moya-moya have no identifiable etiology, it is reported most frequently among patients with neurofibromatosis and sickle cell disease and after cranial irradiation.31 Of note, strokes and the moya-moya syndrome have been reported among patients with neurofibromatosis in the absence of a CNS tumor or treatment with radiation therapy.27,32–35 Rudoltz et al. reported five children with CNS tumors and strokes.3 They proposed that patients with neurofibromatosis may have an exaggerated vascular injury after treatment with radiation therapy. In our series, however, an increased occurrence of stroke was not observed among children with neurofibromatosis and CNS tumors. Alternatively, we believe that the intimate proximity of optic pathway tumors and the large cerebral arteries may enhance the subsequent risk of stroke, especially after treatment with radiation therapy. Limitations of this report include the relatively short median duration of follow-up for our patient population. Indeed, our data does not support and we have no reason to believe that the incidence of stroke might decrease with longer duration of patient survival. We believe that an evaluation of a larger group of adult survivors of childhood CNS tumors over a longer period of time may be able to better address this issue. Heikens et al. reported that survivors of childhood medulloblastoma were at increased risk of cardiovascular disease.36 They identified an association between growth hormone deficiency and elevation of LDL cholesterol and apolipoprotein B, and weight: height ratio among medulloblastoma survivors. Perhaps the growth hormone deficiency observed in two

1100


of our patients may have contributed to their subsequent cerebrovascular disease and stroke. Finally, over the past 15 years, there has been a impressive evolution of the understanding of acquired and inherited prothrombotic states and their influence on cerebrovascular disease. However, this report is inadequately able to address the influence of such states on the subsequent occurrence of a stroke in this patient population. Given the heterogeneity of our patients and their clinical situations, clinical preferences of treating physicians, and the continuing evolution of understanding of the causes of strokes during the period of this review, it is not surprising that were widely disparate evaluations for a hypercoaguable state. Patients in our series were inconsistently evaluated for the presence of hypercoaguable states after their strokes, and those investigations that were performed failed to identify inherited hypercoaguable states. At this time, we recommend the following evaluation for the presence of a hypercoaguable state after a nonperioperative stroke in any patient with a brain tumor: protein C activity, protein S activity, antithrombin, prothrombin 20210A mutation, factor V Leiden mutation, plasma homocysteine, and antiphospholipid antibodies.2 In summary, this report defines the incidence of nonperioperative stroke in children with CNS tumors as 4.03 strokes/1000 years of nonperioperative patient follow-up. Also, children with optic pathway gliomas and those treated with radiation therapy were identified as having a greater association with a subsequent nonperioperative stroke. Of course, radiation therapy remains an integral component of the multidisciplinary treatment of many childhood CNS tumors, despite its association with nonperioperative stroke, neurocognitive deficits, somatic growth impairment, and hearing loss. Because children with optic pathway gliomas may be at particularly high risk of stroke after radiation therapy, the desired beneficial therapeutic effects of irradiation must always be weighed against its potentially adverse effects, including stroke, in this patient population.

5.

6.

7. 8.

9. 10.

11.

12. 13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

REFERENCES 1.

2.

3.

4.

Broderick J, Talbot GT, Prenger E, Leach A. Stoke in children within a major metropolitan area: the surprising importance of intracerebral hemorrhage. J Child Neurol 1993;8:250 –5. deVeber G, Roach ES, Riela AR, Wiznitzer M. Stroke in children: recognition, treatment, and future directions. Semin Pediatr Neurol 2000;7:309 –17. Rudoltz MS, Regine WF, Langston JW, Sanford R, Kovnar EH, Kun LE. Multiple causes of cerebrovascular accidents in children with tumors of the suprasellar region. J NeuroOncol 1998;37:251– 61. Rajakulasingam K, Cerullo LJ, Raimondi AJ. Childhood moyamoya syndrome. Childs Brain 1979;5:467–75.

23. 24.

25.

26.

Hilal SK, Soloman GE, Gold AP. Cerebral occlusive disease in children. I. Acute acquired hemiplegia. II. Neurocutaneous syndromes. Radiology 1971;99:71–94. Debrun G, Sauvengrain J, Aicardi J, Goutieres F. Moyamoya, a non-specific radiologic syndrome. Neuroradiology 1975;8:241– 4. Lee KF, Hades PJ. Intracranial ischemic lesions. Radiol Clin North Am 1967;5:363–93. Mori K, Takeuchi J, Ishikawa M, Handa H, Toyama M, Yamaki T. Occlusive arteriopathy and brain tumor. J Neurosurg 1978;49:2–35. Servo A, Puranen M. Moyamoya syndrome as a complication of radiation therapy. case report. J Neurosurg 1978;48:1026–9. Bataini JP, Delanian S, Ponvert D. Chiasmal gliomas: results of irradiation management in 57 patients and review of literature. Int J Radiat Oncol Biol Phys 1991;21:615–23. Kovalic JJ, Grigsby PW, Shepard MJ, Fineberg BB, Thomas PR. Radiation therapy for gliomas of the optic nerve and chiasm. Int J Radiat Oncol Biol Phys 1990;18:927–32. Painter MJ, Chutorian AM, Hilal SK. Cerebrovasculopathy following irradiation in childhood. Neurology 1975;25:189–94. Beyer RA, Paden P, Sobel DF, Flynn FG. Moyamoya pattern of vascular occlusion after radiotherapy for glioma of the optic chiasm. Neurology 1986;36:1173– 8. Okuno T, Prensky AL, Gado M. The moyamoya syndrome associated with irradiation of an optic glioma in children: report of two cases and review of the literature. Pediatr Neurol 1985;1:311– 6. Flickinger JC, Torres C, Deutch M. Management of lowgrade gliomas of the optic nerve and chiasm. Cancer 1988; 61:635– 42. Pierce SM, Barnes PD, Loeffler JS, McGinn C, Tarbell NJ. Definitive radiation therapy in the management of symptomatic patients with optic glioma. Cancer 1990;65:45–52. Mitchell WG, Fishman LS, Miller JH, Nelson M, Zeltzer PM, Soni D, et al. Stroke as a late sequela of cranial irradiation for childhood brain tumors. J Child Neurol 1991;6:128 –33. Grenier Y, Tomita T, Marymont MH, Byrd S, Burrowes DM. Late postradiation occlusive vasculopathy in childhood medulloblastoma. J Neurosurg 1998;89:460 – 4. Fouladi M, Langston J, Mulhern R, Jones D, Xiong X, Yang J, et al. Silent lacunar lesions detected by magnetic resonance imaging of children with brain tumors: a late sequela of therapy. J Clin Oncol 2000;18:824 –31. deVeber G, Monagle P, Chan A, MacGregor D, Curtis R, Lee S, et al. Prothrombotic disorders in infants and children with cerebral thromboembolism. Arch Neurol 1998;55:1539 – 43. Ohene-Frempong K, Weiner SJ, Sleeper LA, Miller ST, Embury S, Moohr JW, et al. Cerebrovascular accidents in sickle cell disease: rates and risk factors. Blood 2001;91:288 –94. Rogers LR, Cho ES, Kempin S, Posner JB. Cerebral infarction from non-bacterial thrombotic endocarditis. Clinical and pathological study including the effects of anticoagulation. Am J Med 1987;83:746 –56. Graus F, Rogers LR, Posner JB. Cerebrovascular complications in patients with cancer. Medicine 1985;64:16 –35. Packer RJ, Rorke LB, Lange BJ, Siegel KR, Evans AE. Cerebrovascular accidents in children with cancer. Pediatrics 1985;76:194 –201. Lanthier S, Carmant L, David M, Larbrisseau A, de Veber G. Stroke in children: the coexistence of multiple risk factors predicts poor outcome. Neurology 2000;54:371– 8. Klatte EC, Franken EA, Smith JA. The radiographic spectrum in neurofibromatosis. Semin Roentgenol 1976;11:17–33.

Strokes in Pediatric Central Nervous System Tumor Patients/Bowers et al. 27. Hirata Y, Matsukado Y, Mihara Y, Kochi M, Sonoda H, Fukumura A. Occlusion of the internal carotid artery after radiation therapy for the chiasmal lesion. Acta Neurochirur 1985;74:141–7. 28. Janss AJ, Grundy R, Cnaan A, Savino PJ, Packer RJ, Zackai EH, et al. Optic pathway and hypothalamic/chiasmatic gliomas in children younger than age 5 years with a 6-year follow-up. Cancer 2000;75:1051–9. 29. Sutton LN, Gusnard D, Bruce DA, Fried A, Packer RJ, Zimmerman RA. Fusiform dilatation of the carotid artery following radical surgery of childhood craniopharyngiomas. J Neurosurg 1991;74:695–700. 30. Brant-Zawadzki M, Anderson M, DeArmond SJ, Conley FK, Jahnke RW. Radiation-induced large intracranial vessel occlusive vasculopathy. Am J Roentgenol 1980;134:51–5. 31. Riela AR, Roach ES. Etiology of stroke in children. J Child Neurol 1993;8:201–20.

1101

32. de Kersaint-Kelly A, Zenethe L, Dabouis G, Mussini JM, Lajat Y, Robert R, et al. Abnormalities of the intercerebral vasculature in a case of neurofibromatosis. J Neuroradiol 2000;7: 193– 8. 33. Erickson R, Woolliscroft J, Allen RJ. Familial occurrence of intracranial arterial occlusive disease (moya-moya) in neurofibromatosis. Clin Genet 2000;18:191– 6. 34. Toboada D, Alonso A, Moreno J, Muro DF. Occlusion of the cerebral arteries in recklinghausen’s disease. Neuroradiology 1979;18:281– 4. 35. Tomsick TA, Luskin RR, Chambers AA, Benton C. Neurofibromatosis and intercranial arterial occlusive disease. Neuroradiology 1976;11:229 –34. 36. Heikens J, Ubbink MC, van der Pal HP, Bakker PJ, Fliers E, Smilde TJ, et al. Long term survivors of childhood brain cancer have an increased risk for cardiovascular disease. Cancer 2000;88:2116 –21.