Clinical manifestations

Presentation and course

The clinical manifestations of an associated primary central nervous system neoplasm in a child with neurofibromatosis are dependent on the type of tumor present and its location within the nervous system. The most common neoplasm that occurs in patients with neurofibromatosis is the visual pathway glioma. This tumor may involve one or both optic nerves, the chiasm, or other portions of the visual pathway (12; 41). Clinical manifestations of optic nerve involvement include unilateral or bilateral proptosis, decreased vision in 1 or both eyes, optic nerve pallor, and restricted extraocular movements (primarily due to a local restrictive process). On clinical examination, there may be difficulty in retropulsing the involved eye because of increased tissue within the orbit. There is often a significant discrepancy between the apparent size of the tumor and visual acuity. Patients with optic nerve neoplasms usually have decreased acuity in the involved eye, and are often blind; however, some patients will have usable vision in the involved eye. Optic nerve meningiomas may occur in patients with neurofibromatosis 1 and may cause findings identical to those seen in patients with optic nerve gliomas. Another abnormality found in patients with neurofibromatosis 1, dysplasia of the sphenoid wing, may also be associated with optic nerve dysfunction, but is usually distinguishable from an optic nerve tumor.

Tumors of the optic chiasm are usually associated with a variable degree of loss of visual acuity and visual field (24). Pure chiasmatic tumors will not cause proptosis. Visual loss in patients with chiasmatic or retroorbital gliomas is variable and may be unilateral or bilateral. Central scotomas, a measurable depression of central vision, occur in approximately 70% of patients. Peripheral field defects are common, but they are also variable and include quadrantic or hemianopic fields. Bitemporal hemianopic visual field loss occurs in less than one half of patients.

As stated previously, funduscopic examination in a child with an optic nerve glioma usually shows optic nerve pallor, although optic nerve hypoplasia has been noted. Chiasmatic lesions may also result in unilateral or bilateral optic nerve atrophy or pallor. Ocular motor difficulties may also occur, and include abduction deficits or a fine rapid nystagmus. The pathophysiology of this fine, rapid nystagmus is unclear and may be due, at least in part, to decreased visual acuity.

If the tumor is large enough, hypothalamic dysfunction may occur with resultant precocious puberty or diabetes insipidus. This is relatively infrequent in patients with neurofibromatosis and visual pathway tumors. Similarly the diencephalic syndrome, manifest by emaciation, failure to thrive, and a seemingly over-alertness, is a relatively common-presenting symptom complex in patients with hypothalamic tumors and no stigmata of neurofibromatosis 1; it is less frequent in children with neurofibromatosis 1 (48).

The signs and symptoms of intracranial chiasmatic and hypothalamic involvement in children less than five years of age with neurofibromatosis 1 are variable. When proptosis is present, the diagnosis is usually made quickly. However, there is a greater delay between the onset of symptoms and diagnosis in children with retroorbital involvement. Some children have been misdiagnosed as having spasm nutans, a condition affecting children between four and eight months of age, in which there is an acquired uniocular, or at times biocular, nystagmus and head nodding (01).

The primary differential diagnosis of a child with neurofibromatosis 1 and a primary central nervous system tumor is the distinction between tumors and other abnormal regions of signal intensity within the brain seen on MR. The most common nontumorous abnormal findings in patients with neurofibromatosis 1 are focal areas of signal intensity, appearing as bright regions on T2-weighted images within the brain (09). These areas of increased signal intensity have no apparent mass effect and primarily occur in the cerebellar peduncles, globus pallidus, brainstem, thalamus, and centrum semiovale. They may be solitary, multiple, or confluent. These areas are not thought to represent tumorous growths, although histological confirmation has rarely been obtained. Theories postulate that these are dysmature regions of brain, hamartomas, or dysmyelinated areas of white matter. Because some of these lesions may apparently disappear over time (the incidence of such lesions is much higher in younger children than in older children or adults), the concept that the areas may represent immature areas of brain or dysmyelinated areas of brain, rather than true hamartomas, has taken on greater credence. It is unclear how often, if ever, areas of increased signal intensity will grow or transform into neoplastic lesions.

Similarly, there may be confusion between optic nerve sheath dysplasia and the presence of an optic nerve glioma. Both may show apparent increased tissue within the optic canal, but careful evaluation of the intraorbital space can usually allow distinction between a dysplastic optic nerve sheet and a true optic nerve glioma. Similarly, optic nerve meningiomas, which are significantly less frequent in children with neurofibromatosis 1 than optic nerve gliomas, may mimic an optic nerve glioma.

Patients with neurofibromatosis 1 and chiasmatic tumors will often show abnormal signal streaking along the optic tracts, optic radiations, or both. Although histologic confirmation is usually not obtained, it has been thought that these abnormal regions of signal intensity in patients with chiasmatic tumor, optic nerve tumor, or both represent infiltrating tumor (48). It is also possible that they do not represent infiltrating tumor, but rather represent dysplastic regions of brain.

The clinical manifestations of other tumor types that occur in higher incidence in children with neurofibromatosis 1, such as astrocytomas in other regions of the brain, ependymomas, neurofibrosarcomas, primitive neuroectodermal tumors, and meningiomas are similar to the manifestations of these tumors in patients without neurofibromatosis (74).

Because of the higher incidence of intracranial visual pathway gliomas in children with neurofibromatosis 1, children with neurofibromatosis 1 are often screened with CT or MR at the time of recognition of neurofibromatosis 1. This has led to the discovery of abnormal tissue within the brain, at times believed to represent gliomas in children who are asymptomatic at the time of diagnosis (48). In children with so-called asymptomatic optic nerve tumors, careful clinical examination often reveals some degree of optic nerve pallor or restricted extraocular movement. However, many patients are truly asymptomatic, and others have apparently static neurologic or visual deficits.

Prognosis and complications

The prognosis of patients with visual pathway gliomas and neurofibromatosis 1 seems somewhat better than the prognosis for patients with similar tumors and no evidence of neurofibromatosis 1 (48; 37; 38; 02). Less than one in five patients with neurofibromatosis 1 static and presumed low-grade visual pathway gliomas at the time of diagnosis will develop progressive disease over a 5-year period of observation. In patients with visual pathway gliomas and progressive disease, the overall prognosis seems the same as for those children without neurofibromatosis 1 (45). Approximately 90% of children with visual pathway gliomas and neurofibromatosis 1 will be alive five years following diagnosis; however, probably less than 70% will be alive 10 and 15 years later (20). The causes of death in children with visual pathway gliomas and neurofibromatosis 1 have not been well documented, but have included progressive neurologic dysfunction due to continued growth of a low-grade malignancy and occasionally malignant transformation of the malignancy. To date, there is no evidence that low-grade lesions in children with neurofibromatosis 1 are more likely to transform to malignant high-grade gliomas than in children without neurofibromatosis 1. Children with neurofibromatosis 1 and low-grade visual pathway gliomas are at significant risk for increasing visual morbidity, including unilateral or bilateral blindness and significant visual field loss. Visual outcomes may be suboptimal, even after treatment with chemotherapy (42; 68). Diencephalic gliomas may also result in obstructive hydrocephalus, hemiparesis, cognitive dysfunction, and obtundation. Cognitive dysfunction may be a result of the treatment utilized to control progressive tumor growth, especially the use of radiotherapy.

The outcomes of children with other primary central nervous system tumors and neurofibromatosis 1 has never been shown to be different than the outcomes of children without neurofibromatosis 1 (14) except possibly in the situation of patients with apparent brainstem tumors; where once again some children with neurofibromatosis 1 may have prolonged disease stability without intervention (56; 73). There have been concerns raised that children with neurofibromatosis 1 may be more sensitive to the detrimental effects of radiotherapy needed to control diseases such as medulloblastoma and ependymoma, especially the possibility that such treatment may cause secondary tumors including sarcomas, high-grade gliomas, and meningiomas. In a study of 80 patients with neurofibromatosis 1 and optic nerve gliomas, nine of 18 patients who received radiotherapy developed 12 second tumors, in comparison to eight of 40 not treated with radiotherapy, with a relative risk increase of 3.04 (66).

Plexiform neurofibromas, although not a primary central nervous tumor, may cause severe neurologic dysfunction, primarily by compression of the spinal cord, spinal nerves, or peripheral nerves (44). Studies have suggested that molecular targeted therapies, such as the mTOR inhibitors, PDGF receptor inhibitors, and drugs, which block signaling through the RAS-MAPK pathway, may be effective in slowing the growth of plexiform neurofibromas (26; 25; 78; 76; 77). However, by far the most effective are the MEK inhibitors that block signaling through hyperactive RAS-MAPK pathway (18).

Clinical vignette

A 2-year-old boy was referred to the neurologist's office because his parents were concerned that he could not see out of the right eye. The family had always been aware that his right eye was "bulging," but they had not been concerned about his vision until he was seen by a pediatrician who felt that the child could not see well.

On examination, the child was awake and alert. His head circumference was at the 98th percentile for age. His general physical examination was noteworthy for multiple pigmented areas (cafe-au-lait spots) over his trunk. There was axillary freckling.

Cranial nerve examination disclosed that his right eye was proptotic and extended 7 mm outward as compared to his left. The right eye had full movement. The eyelid over the eye was not swollen. The eye was not painful on movement. Funduscopic examination noted that the right optic nerve was pale. The left seemed normal in color. The child's visual acuity out of the right eye was difficult to fully assess and was probably no better than 20/800. On the left, acuity seemed normal, and the visual fields seemed full. The child's cranial nerve examination was otherwise within normal limits. Coordination testing was normal except for mild hypotonia and a mild degree of fine motor delay. Reflexes were 2 out of 4 and symmetric in the upper extremities and 2 out of 4 and symmetric in the lower extremities.

On neuroimaging, the child had a large mass behind the right eye that extended posteriorly and merged into a large chiasmatic hypothalamic abnormality. There was also apparent involvement of the hypothalamus. On MRI scan, the mass clearly extended posteriorly and there was streaking along the visual pathway.

Biological basis

Etiology and pathogenesis

The neurofibromatosis 1 gene resides on the long arm (q portion) of chromosome 17 and that the neurofibromatosis 1 locus encodes a tumor suppressor gene (75). It has been postulated that all patients with neurofibromatosis 1 inherit one nonfunctional allele. The loss of one functional neurofibromatosis 1 allele may have direct consequences, however, when the second neurofibromatosis 1 allele is lost as a somatic event, as postulated by the two-hit proposal of Knudsen, tumor formation occurs. If the loss is in a Schwann cell, the result is a neurofibroma. If the loss occurs in a glial cell, a glioma develops.

The pathogenesis and pathophysiology of tumor growth in neurofibromatosis 1 is secondary to the loss of the neurofibromatosis 1 gene involving a tumor suppressor locus (75). The gene for neurofibromatosis 1 stretches across approximately 300 kb of the 17q 11.2 region. The gene contains 50 axons and encodes a protein of 2818 amino acids. As stated previously, it is believed that the loss of one functioning allele occurs in all cells of patients with neurofibromatosis 1. Although this may have some direct consequences, such as the association of neurofibromatosis 1 with cognitive disabilities, it has been postulated that a tumor would only form when the remaining neurofibromatosis 1 allele is lost as a somatic event (27). It has also been postulated that if mutations occur in other tumor suppressor genes or oncogenes, there would be a higher incidence of true malignancy, possibly explaining the apparent transformation of some neurofibromas to neurofibrosarcomas or possibly lower grade gliomas to higher grade gliomas. The severity of disease, especially the variable incidence of tumor formation, may be influenced not only by the presence of neurofibromatosis 1 but by other genetic factors and environmental factors, and to some degree random chance.

The protein product of the neurofibromatosis 1 gene, neurofibromin, can be found in virtually all tissues, but is expressed at highest levels in the brain and spinal cord. At least one of the isoforms of neurofibromin is believed to play a role in central nervous system differentiation. Neurofibromin is a member of the GTPase-activating protein family. There is evidence that neurofibromin interacts with the ras signaling pathways, and this contributes to the development of tumors (36). BRAF mutations found in a high percentage of children with pilocytic astrocytomas do not occur in those with neurofibromatosis 1 (55; 32; 31; 28). However, a subgroup of children with neurofibromatosis 1 will have low-grade gliomas with bi-allelic neurofibromatosis inactivation and concomitant FGFR1 mutation (22). Molecular analysis has demonstrated that specific subtypes of mutations on the NF1 gene are associated with a higher likelihood of glial tumor development (67). The RAS/MAPK cellular signaling pathway is a therapeutic target for progressive low-grade gliomas (64; 25) and MEK inhibitors have been shown to be effective in recurrent low-grade gliomas (05). Farnesyl transferase protein inhibitors have been studied, and results have been disappointing (25). The neurofibromatosis 1 tumor suppressor gene regulates mTOR pathway activity, with hyperactivation of the mTOR pathway in neurofibromatosis 1 tumors, especially gliomas (15). It is activated in the majority of neurofibromatosis 1-related gliomas. mTORC2 is differentially active in visual pathway gliomas and is a potential therapeutic target (30; 25).

In addition, in older children and adults with neurofibromatosis 1, anaplastic astrocytomas with piloid features occur. Such tumors may be a transformation of the more common juvenile pilocytic astrocytoma. This variant is more clinically aggressive and in addition to gene alternations affecting the NF1 gene, have concomitant other gene alterations; most commonly mutations of the CDKN2A/B gene and mutations or loss of ATRX (60).

Epidemiology

Neurofibromatosis 1 is one of the most common genetic disorders affecting the central nervous system. Inherited as an autosomal dominant condition, it affects approximately 1 in 3500 individuals in all ethnic groups. The spontaneous mutation rate of this gene is also extremely high, and one-half of all affected patients have no apparent relatives with the disease. The majority of central nervous system tumors in patients with neurofibromatosis 1 will become manifest within the first two decades of life. The exact incidence of CNS tumors with neurofibromatosis is difficult to ascertain due to the variability of the neurofibromatosis 1 phenotype, as affected individuals may escape detection. Visual pathway astrocytomas are the most common intracranial neoplasm in neurofibromatosis 1 and occur in between 5% to 20% of patients with neurofibromatosis 1 (13; 14). With the advent of screening programs at the time of diagnosis, the higher incidence figure is probably more accurate. The percentage of children with documented chiasmatic gliomas in neurofibromatosis 1 have varied between series, although approximately 20% to 25% of patients with chiasmatic gliomas will be noted to have neurofibromatosis 1 (20; 48). This percentage is probably lower when only children with progressive visual dysfunction or tumor growth are included in the group of patients with chiasmatic or visual pathway gliomas. The association between optic nerve gliomas and neurofibromatosis 1 is even stronger; neurofibromatosis 1 is present in up to 50% of children with intraorbital gliomas. Some have even suggested that an isolated optic nerve glioma in a child under 10 years of age is diagnostic of neurofibromatosis 1, with or without other clinical manifestations of the disease. In the consensus conference criteria, however, patients with optic nerve gliomas needed to have at least one other criterion for diagnosis of neurofibromatosis 1.

Other astrocytic tumors occur more frequently in patients with neurofibromatosis 1 than in the general population. However, the exact incidence of this increase is unclear. In an integrated molecular and clinical analysis of low-grade tumors, gangliogliomas also were noted to be relatively frequent (22).

Similarly, it has been reported that patients with neurofibromatosis 1 have a higher incidence of ependymomas, medulloblastomas, and meningiomas (14). Neurofibrosarcomas, which are malignant neoplasms arising from the nerve sheets of the peripheral nerves, although not primary central nervous system tumors, deserve mention because these tumors may arise from the cranial nerves and affect central nervous system brain structures. Neurofibrosarcomas occur in approximately 2% to 5% of patients with neurofibromatosis 1, but occurrence is more common in adults than in children.

Prevention

There is no known prevention for the development of intracranial tumors in patients with neurofibromatosis 1. However, given the hypothesis that these tumors arise from a loss of the neurofibromatosis 1 allele as a somatic event in patients who already have one abnormal allele due to neurofibromatosis 1, limitation of exposure to potentially damaging environmental factors (such as radiation) seems reasonable.

Diagnostic workup

The diagnostic workup of a patient with neurofibromatosis 1 and a presumed intracranial neoplasm usually begins and ends with an MR scan (09; 48). Computed tomographic scanning may demonstrate the presence of an abnormality, but is not as sensitive as MR.

In patients with abnormal regions of brain on MR scans performed without contrast enhancement, contrast agents such as gadolinium should be given. Gadolinium enhancement is often useful in separating hamartomatous regions from true neoplasms, as hamartomatous areas will rarely, if ever, show gadolinium contrast enhancement. However, some true low-grade neoplasms and, rarely, high-grade glial neoplasms, will not enhance with gadolinium. The use of screening MR studies with neurofibromatosis 1 is controversial, although their use in young children may lead to improved visual outcomes (57).

Some have suggested that the use of evoked potentials, especially visual evoked potentials, may be helpful in diagnosing and following patients with neurofibromatosis 1 and visual pathway gliomas. The ability of evoked responses to sensitively follow the course of disease has never been well substantiated. Ocular coherence tomography is a newer technique that shows promise in following patients with visual pathway gliomas (03).

Following neuroimaging confirmation of the presence of the tumor, every attempt should be made to test visual acuity and fields, even in very young patients. Formal visual field evaluation procedures, such as Goldmann perimetry, are needed to determine the exact extent of visual field loss. Optical coherence tomography detects retinal nerve fiber layer thinning in subjects with optic nerve gliomas and may be a useful adjunct in patient management (10; 54; 03; 04).

Due to the frequent coexistence of hypothalamic dysfunction in children with visual pathway (diencephalic) gliomas, baseline endocrinologic studies are indicated in all children at the time of diagnosis. In children with evidence of delayed growth, this also includes measurement of growth hormone production. In pubertal children, sexual hormones should also be evaluated. All patients require evaluation of thyroid function, as well as screening for diabetes insipidus and inappropriate secretion of antidiuretic hormone.

Management

The management of children with neurofibromatosis 1 and visual pathway gliomas remains unsettled. Due to the frequent detection of such lesions at the time when the patient is asymptomatic or has static visual deficits, neurologic deficits, or both, most observers have recommended no specific intervention at the time of initial diagnosis. As stated previously, in multiple follow-up series, less than 20% of patients with visual pathway, presumed low-grade tumors and neurofibromatosis 1, have progressed over variable periods of observation ranging between five and 20 years. For this reason, it is relatively well accepted that a period of observation is indicated, prior to the initiation of treatment, for children with isolated optic nerve tumors and probably for children with chiasmatic tumors and neurofibromatosis 1 (16).

For patients with isolated or bilateral optic nerve gliomas that progress, treatment with surgery, radiotherapy, or both has been utilized. Due to concerns over mutagenic potential, radiotherapy is rarely used in children with neurofibromatosis 1 (43). Surgery has primarily been utilized in patients with isolated intraorbital lesion for cosmetic reasons (65; 19). Although various types of surgical techniques have been utilized, most have attempted to spare the globe, if possible. Surgery has also been recommended for isolated orbital lesions to prevent the development of secondary life-threatening intracranial tumor spread. However, it is unclear how frequent isolated intraorbital lesions will extend into the chiasm and how well surgical removal (or for that matter, radiotherapy) is able to prevent such spread.

Similarly, the management of chiasmatic gliomas is unsettled. As stated previously, a period of observation is probably indicated for all patients with neurofibromatosis 1 unless there is clear evidence of progressive visual loss at the time of diagnosis. Surgery has been utilized for both diagnosis and the initial management of children with chiasmatic lesions (06; 79). Because the relative incidence of low-grade chiasmatic gliomas occurring in this region is much higher than that of any other tumor, it does not seem that surgery is needed for diagnosis in the majority of children with neurofibromatosis 1. However, in older patients, especially in those with tumors that act unusually aggressive or are poorly responsive to therapy, biopsy may be warranted to rule out the preserve of other tumor specific genetic changes (60). The majority of chiasmatic tumors in children with neurofibromatosis 1 are solid and not easily amenable to extensive surgical resection. However, occasionally, globular, partially cystic lesions can be subtotally resected. Such resections may result in increased neurologic and visual morbidity. At times, especially when the tumor is bulky and laterally exophytic, resection may relieve obstructive hydrocephalus and occasionally results in disease stabilization for prolonged periods of time. This type of delay in disease progression is of significant benefit in very young children with large progressive tumors.

The role of radiation therapy for patients with neurofibromatosis 1 and chiasmatic gliomas remains controversial. There is fairly good evidence that radiation therapy, in doses ranging between 5000 cGy and 5400 cGy, results in tumor shrinkage with stabilization of disease in over 80% of treated patients (12; 69). However, a significant number of patients with neurofibromatosis 1 and chiasmatic involvement will have apparent stable disease for years before requiring any therapy. In addition, the theoretic risk of the radiotherapy causing a second genetic abnormality in radiated cells and resulting in the transformation of a lower-grade tumor to a higher-grade tumor, or in the development of a malignancy in nontumor cells, has led many to question the role of radiation therapy for patients with neurofibromatosis 1 and visual pathway gliomas, unless absolutely mandatory (11; 66). The difficulty in separating tumor from abnormal regions of signal intensity within the brain, which may also be seen on MR in patients with neurofibromatosis 1, has made it difficult to determine the leading edge of tumor. This often requires large radiation portals to be used to cover all areas of abnormal signal intensity on MR in patients with presumed gliomas and neurofibromatosis 1, and probably a higher incidence of neurocognitive sequelae. Possibly as importantly, radiotherapy may predispose these patients to secondary malignant tumors and vasculopathy (71).

The theoretic concerns over the use of radiotherapy in children with neurofibromatosis 1 and the documented concern that very young patients with large lesions who receive local radiotherapy often have significant intellectual dysfunction has led to the routine employment of chemotherapy for patients with progressive visual pathway gliomas and neurofibromatosis 1 (63; 53; 51). Treatment with actinomycin D and vincristine has resulted in disease stabilization or regression of disease in as high as 80% of children with newly diagnosed chiasmatic tumors, including some children with neurofibromatosis 1 (53). The largest experience has been with the combination of carboplatinum and vincristine (51). In a published series, which included 14 children with neurofibromatosis, the combination of carboplatinum and vincristine was able to show disease stabilization in over 90% of patients (47). In addition, nearly two-thirds of treated patients had objective shrinkage of tumor while receiving chemotherapy. A randomized international trial confirmed the efficacy of carboplatin and vincristine in 127 children with neurofibromatosis 1, and progressive gliomas 5-year event-free survival was nearly 70% in this latter study (02; 35). Other drug regimens are presently under study for children with chiasmatic gliomas (58; 46; 39). Vinblastine has been shown to also be effective (08). Most are attempting to obviate the need for radiation. Children with neurofibromatosis 1 and visual pathway gliomas often have associated cognitive deficits (34). However, despite apparent excellent tumor control, visual outcomes have been poorly documented, and visual deterioration has been noted despite stable MRIs (42; 68; 23; 17). In one series, 47% of eyes had visual deterioration after treatment with standard chemotherapy (68). In another, visual deterioration was noted in approximately 30% (23). Overall, after treatment, over 40% of patients have significant impairment of vision (59).

Given the molecular genetic defect in neurofibromatosis, biological-based therapy is an attractive option. The experience with a mTor inhibitor, rapamycin, demonstrates that biological agents can result in regression of astrocytomas in patients with another genetic syndrome, tuberous sclerosis, who have a defect in a different cellular signaling pathway. In a preclinical mouse model of neurofibromatosis 1 optic glioma, pharmacological mTOR inhibition with an mTOR inhibitor, rapamycin, resulted in decreased tumor cell proliferation in a dose-dependent fashion and decreased tumor volume (15; 29). mTORC2 inhibitors may be even more effective (30). This has connotations for patients with neurofibromatosis 1, and trials with everolimus for children with neurofibromatosis 1 and recurrent low-grade gliomas have demonstrated prolonged stable disease in the majority, but few radiographic responses (72).

Bevacizumab with or without irinotecan is another option for children with progressive lesions failing front-line chemotherapy and has resulted not only in tumor shrinkage but also in visual improvement (50; 33).

Drugs that inhibit the RAS-MAPK pathway, primarily MEK inhibitors, are being increasingly used (26; 25; 78; 05; 52). Of all the biological agents employed, they seem to have the greatest efficacy (05). In a phase II study of selumetinib, a MEK inhibitor, for 25 children with neurofibromatosis 1 and progressive low-grade gliomas following initial chemotherapy, 40% had greater than a 50% reduction in tumor size, and the remaining 60% had either a smaller, but documentable, reduction in size or stable disease (21). No patient progressed radiographically or had clinical deterioration while on active treatment; at least two had visual improvement. The 2-year progression-free survival rate at 2 years was 96%. Patients were treated for a maximum of 2 years and at least one-half remained stable 2 years off treatment. Similar studies are underway with other MEK inhibitors. As of January 2020, a prospective randomized trial comparing selumetinib to carboplatin and vincristine in children with progressive low-grade gliomas who have not been treated is underway through the Children’s Oncology Group (49). MEK inhibitors, by and large, have been well tolerated but can cause side effects not usually seen with conventional chemotherapies such as severe acneiform rash and gastrointestinal upset. Rarely, they are associated with hypotonia (although asymptomatic elevations of muscle enzymes is common), retinal venous occlusion, or cardiomyopathy; their long-term safety is unknown (05; 21).

The management of other primary central nervous system tumors in children with neurofibromatosis 1 is the same as for some tumors in patients without neurofibromatosis 1; except for those with brainstem gliomas. In children with neurofibromatosis 1 and brainstem gliomas a period of observation is probably indicated prior to any specific intervention (56; 73). Once again, there have been serious concerns raised over the use of radiotherapy for children with neurofibromatosis 1 and low-grade malignancies of the cortex and other regions of brain outside the diencephalon. Most patients with low-grade gliomas should be observed prior to the initiation of any specific form of treatment, especially if extensive surgical resection can be undertaken at the time of diagnosis. The utility of chemotherapy is also increasingly utilized in children with neurofibromatosis 1 and low-grade gliomas outside the visual pathway (02).

Special considerations

Anesthesia

The same precautions are taken as for any patient with possible increased intracranial pressure.