Stroke & Vascular Disorders
Oct. 26, 2023
At vero eos et accusamus et iusto odio dignissimos ducimus qui blanditiis praesentium voluptatum deleniti atque corrupti quos dolores et quas.
Optic pathway gliomas may involve the optic nerve, chiasm, and optic radiations and tracks. Isolated optic nerve tumors, especially in patients with neurofibromatosis type 1, can often be observed without any specific intervention for many years. For those patients with isolated optic nerve tumors and complete loss of vision, surgical resection is possible. For other clinical situations, chemotherapy and, less frequently, radiotherapy may be initial treatment. In this article, the author reviews the treatment approaches for these tumors and the efficacy of treatment. He also focuses on the therapeutic implications of new insights into the biology of these tumors.
• Visual pathway gliomas may occur sporadically, but 50% are associated with neurofibromatosis type 1 (NF1).
• Most visual pathway gliomas are pilocytic astrocytomas.
• Mutations in BRAF, most commonly an activating fusion, occurs in the majority of patients with non-NF1-related visual pathway pilocytic tumors.
• Chemotherapy is frequently effective in delaying, if not obviating, the need for radiotherapy in “non-operable” lesions.
• Molecular-targeted therapy is playing a larger role in management of progressive lesions.
Chiasmatic gliomas arising in childhood have been the subject of many reports, with various recommendations for treatment (46; 24; 82). The first descriptions of the various types of orbital tumors were by Byers (17). Hudson was the first to suggest that optic pathway tumors were hamartomatous rather than neoplastic (48), a concept that was debated heatedly over the next several decades. The surgical accessibility of these lesions was initially described by Martin and Cushing (69).
At times, lesions of the chiasm have been discussed as a singular entity, and at other times, as a subgroup of diencephalic tumors or visual pathway tumors. Although chiasmatic gliomas may remain relatively limited to the chiasm, there is often contiguous involvement of other brain stem structures. In children with neurofibromatosis, chiasmatic gliomas usually extend to 1 or both optic nerves or posteriorly to the optic tracts or optic radiations (78). Patients without neurofibromatosis may have isolated globular chiasmatic lesions, but also may have tumors that extend into other regions of the diencephalon, such as the hypothalamus and thalamus. It is unclear whether separation of chiasmatic lesions from other diencephalic or visual pathway gliomas is of prognostic significance. Prior to the advent of magnetic resonance imaging, chiasmatic tumors were often arbitrarily classified as being anterior, involving the chiasm and optic nerve, or posterior, involving the chiasm plus hypothalamus or other parts of the visual pathway (73). This distinction was used for prognostic reasons, on the belief that patients with anterior lesions had a more favorable prognosis. Since the introduction of magnetic resonance, such arbitrary distinctions have diminished importance, as the extent of tumor can be more clearly delineated.
Another change concerning concepts of chiasmatic gliomas has been the identification, primarily through screening programs for children with neurofibromatosis, of lesions in asymptomatic patients or children with minimal, apparently, static symptomatology. Although a proportion of patients with neurofibromatosis and asymptomatic, radiographically confirmed lesions will develop progressive disease, the majority remain stable for months or years. This has given more credence to the contention that some chiasmatic gliomas are not truly tumors, but are rather hamartomatous masses without growth potential. Longitudinal studies following patients over many years or decades are needed to clarify the growth potential of chiasmatic lesion in children with neurofibromatosis. However, it is well-documented that isolated optic nerve lesions in children with neurofibromatosis may remain quiescent for many years (46).
Predominantly arising in a chiasm, a pilomyxoid variant of pilocytic astrocytoma has been identified, demonstrating monomorphous pilomyxoid features. It is believed to have a poorer prognosis and a higher tendency to disseminate throughout the nervous system than the classical pilocytic astrocytoma (58). The 2017 WHO classification does not identify it as a distinct entity, although molecular studies have demonstrated that it has a different molecular profile than classic pilocytic tumors and a tendency to “mature” into the more classical tumor (66; 67).
Anaplastic astrocytoma with piloid features is still another entity. This clinically more aggressive lesion is seen in both patients with and without neurofibromatosis 1 and is associated with other concomitant genetic alterations in tumor tissue (90).
The clinical symptoms and signs of patients with chiasmatic glioma are dependent on involvement of contiguous brain structures. As stated previously, children with neurofibromatosis may have no clinical deficits referable to the chiasm at the time of diagnosis (78). Children with contiguous optic nerve involvement most commonly present with unilateral or, less commonly, bilateral proptosis and associated visual loss (46; 36).
The visual loss usually consists of decreased vision in the proptotic eye with a variable degree of visual field loss in the contralateral eye. The proptosis found is usually mild, nonpulsatile and irreducible. There may be associated nasal and downward displacement of the globe and restricted motility. Evaluation of the optic nerve usually discloses optic atrophy without other signs of papilledema. Visual acuity is usually less than 20/200 out of the proptotic eye.
Children with chiasmatic lesions without intraorbital disease will not have proptosis. Older patients, especially those 5 years of age or greater, will often have headaches and complaints of decreased vision. The visual loss may be unilateral or bilateral and the pattern of loss is variable. Central scotomas or measurable depression of central field vision occur in approximately 70% of patients. Peripheral field defects are common, but are also variable and include quadrantic or hemianopic fields (36). Bitemporal hemianopic visual loss occurs in less than one half of patients.
Funduscopic examination in children with chiasmatic lesions tends to disclose unilateral or bilateral optic pallor or atrophy; in young children, optic nerve hypoplasia may also be present. Ocular motor difficulties may occur and include abduction deficit and a rapid, fine nystagmus. The etiology of this fine nystagmus is unclear and may be, at least partially, due to decreased visual acuity.
Hypothalamic dysfunction may occur with resultant precocious puberty or diabetes insipidus. The diencephalic syndrome, manifest by emaciation, failure to thrive, and a seeming over-alertness, is a relatively common presenting symptom complex, especially in children less than 2 years of age at diagnosis.
The symptoms and signs of intracranial involvement in children less than 5 years of age at time of diagnosis, especially in young children less than 2 years of age, are more variable (36; 82). There is usually a greater delay between the onset of symptoms and diagnosis. Irritability and failure to thrive, often associated with macrocephaly, is a relatively common presentation. Although children may have significant visual loss at the time of diagnosis, the most common early symptom is strabismus. It is not uncommon for children to undergo eye patching to treat amblyopia, and even strabismus surgery, before the correct diagnosis is made. The extraocular movement disorders in children with chiasmatic gliomas are variable, but most commonly consist of a unilateral or bilateral pendular nystagmus. Infants may have a shimmering unilateral, or less commonly bilateral, nystagmus, making distinctions with spasmus nutans (a condition effecting children usually between 4 and 8 months of age wherein there is an acquired uniocular, or at times biocular, nystagmus and head nodding) difficult (02; 63). Although some have suggested that clinical findings can be used to separate children with spasmus nutans from those with chiasmatic gliomas, because of the overlap and difficulty in evaluating young children, neuroradiographic studies are usually required to reliably separate patients. Children less than 5 years of age usually demonstrate unilateral or bilateral optic atrophy and, at times, optic nerve hypoplasia. Although hydrocephalus may be present at diagnosis, papilledema is uncommon.
The overall prognosis for patients with chiasmatic gliomas is relatively good. As stated previously, some patients, especially those with neurofibromatosis, may have static lesions for many years (82; 65). Some may apparently never develop progressive disease, although this has not been proven for patients with isolated chiasmatic or intracranial lesions. Certainly, some children with isolated orbital lesions or orbital lesions with associated chiasmatic involvement may do well without any specific intervention for many years (47).
Malignant optic gliomas have been reported as a cause of rapidly progressive visual loss in adults; these tumors tend to occur in middle-aged or elderly individuals who present with the rapid onset of blindness (100; 101; 95). Subsequently, tumor invasion into adjacent brain becomes apparent, leading to death within a few months of presentation (105). These tumors' behavior appears clinically distinct from that of their pediatric counterparts, resembling more that of glioblastoma multiforme.
The reasons for death in patients with chiasmatic gliomas are poorly characterized and are infrequent. Progressive disease may be noted, although the tumor tends to remain low grade. Malignant transformation does occur but is more frequent in older patients (90). Such secondary tumors may also occur in previously irradiated patients (110). Patients may also die of secondary tumors or due to treatment-related sequelae.
A 9-month-old child was brought to the neurologist with a 3-month history of lethargy and poor feeding. On examination, the child’s head circumference was 48 cm, which was above the 98th percentile for the child's age. The child was somnolent but arousable. General physical examination disclosed no abnormal skin lesions. Cranial nerve examination found the nerves reactive and equal. There was a relative afferent pupillary defect on the right, and the child would not fix or follow. There was a shimmering type of nystagmus on lateral gaze that also had a somewhat pendular component. Extraocular movements were intact other than the nystagmus. The child had normal facial movements, and the remainder of the cranial nerves were normal.
On motor and coordination testing, the child could not sit. Overall truncal tone was decreased. There were no specific areas of weakness, although the child was not terribly interested in reaching for objects. The child could not bear weight. Reflexes were 3 and symmetric.
On CT scan, a large, regular mass was found filling the suprasellar region. The mass clearly involved the chiasm and extended into the hypothalamus. There were also some greater extensions off to the left. The optic nerves were not clearly enlarged. On MRI, the mass was of decreased signal intensity on the T1-weighted image, and of increased signal intensity on the T2-weighted image. There was extension or streaking along the optic pathways.
The molecular understanding of optic pathway gliomas has greatly expanded. In patients with neurofibromatosis type 1, the lack of the gene product of the NF1 gene due to bi-allelic inactivation of the NF1-gene results in increased intracellular signaling and activation of RAS-MAPK pathway signaling, as well as possible signaling through the phosphatidylinositol 3-kinase (P13K) pathway (94). In patients without neurofibromatosis type 1, BRAF oncogenic activating mutations due to a BRAF-KIAA fusion protein have been frequently noted, also causing increased RAS-MAPK signaling (86; 49; 93). Less frequently, BRAF point mutations have also been noted (86). The exact proportion of patients with optic pathway gliomas who have BRAF fusion proteins is unclear; it is now believed approximately 40% of visual pathway gliomas in children without neurofibromatosis type 1 have BRAF-KIAA fusions, and 40% have v600E mutations (52; 97; 53). Fusion between KIAA 1549 and BRAF results in a fusion protein that lacks the BRAF regulatory domain. Other activating gene fusions and mutations, such as mutations in FGFR1, PDFR1, PTPN11, and MYB have been documented (52). Patients with neurofibromatosis type 1 do not have concomitant BRAF mutations (42). Both patients with neurofibromatosis 1 and BRAF mutations may develop more clinically aggressive tumors, including anaplastic astrocytoma with piloid features. This latter entity is marked by concomitant additional genetic abnormalities, usually CDKN2A/B deletions or mutations or loss of ATRX expression (90).
Etiology for the vast majority of patients, except for children with neurofibromatosis, is unknown. There is a higher incidence of chiasmatic gliomas, especially with optic nerve or posterior visual pathway involvement, in children with neurofibromatosis.
Although some chiasmatic gliomas and intraorbital lesions have been considered hamartomas, when gliomas are placed in tissue culture, numerous piloid astrocytes with bulk abundant intracellular fibers develop (70; 13). Degenerative changes of these fibers probably represent early formation of Rosenthal fibers. A distinctive feature of some optic nerve gliomas is the ability of their long, thin cellular processes to form fiber tangles in tissue cultures. Pilomyxoid astrocytomas, which are most likely to occur in the hypothalamic/chiasmatic region, are a possible low-grade variant with a more aggressive natural history (107; 16; 04).
Involvement of the subarachnoid space may also occur, accompanied by arachnoid hyperplasia of intermixed glial and leptomeningeal connective tissue; involvement of the orbital soft tissue or bone does not occur. In adults, a histologically more malignant form of visual pathway or chiasmatic glioma is more common. These tumors tend to be marked by regions of pleomorphic and mitotically active astrocytes and tend to invade adjacent structures.
The vast majority of childhood chiasmatic gliomas are histologically benign, pilocytic astrocytomas. These lesions are formed by thin, elongated (pilocytic) and stellate astrocytes (13; 96; 28). Scattered oligodendroglial cells are commonly found but rarely comprise a significant portion of the tumor. In some areas, a microcystic pattern may be present, and Rosenthal fibers can be found in densely fibrillary areas. Mitotic figures in areas of tumor necrosis are rarely present. In intraorbital lesions, connective tissue separation of tumor into glial bundles is commonly seen, but does not occur in intracranial masses. Otherwise, intraorbital and intracranial lesions tend to be histologically identical. Although the overlying dura matter is usually intact, microscopic infiltration into the overlying nerve sheet does occur. The polymyxoid variant, primarily reported in the diencephalic region, may act somewhat more aggressively (19). Increased microvessel density has been associated with a higher rate of progression (12).
Chiasmatic gliomas constitute approximately 3% to 5% of all primary childhood central nervous system tumors, occurring at an incidence of less than 1 per 100,000 children at risk per year. The incidence is higher in children under 2 years of age, as chiasmatic gliomas and associated diencephalic or visual pathway gliomas may comprise as high as 20% of all intracranial tumors. Approximately 20% of children with chiasmatic gliomas will have stigmata of neurofibromatosis. The association between optic nerve gliomas and neurofibromatosis is even stronger, as neurofibromatosis is present in up to 50% of children with intraorbital gliomas. Some have suggested that an optic nerve glioma in a child under 10 years of age is diagnostic of neurofibromatosis, with or without other clinical manifestations of the disease. The reported incidence of optic nerve gliomas in patients with neurofibromatosis is approximately 20%, although the reported incidence is probably increasing with the wider utilization of magnetic resonance imaging screening programs, at diagnosis, in asymptomatic children with neurofibromatosis.
Chiasmatic gliomas may occur at any time during childhood. However, more than 75% of isolated optic nerve gliomas occur in the first decade of life, with the majority being diagnosed by 5 years of age. In some series, the peak age of chiasmatic gliomas is as early as the second year of life. The male-to-female ratio is nearly equal.
The risk factors, other than the association with neurofibromatosis, for children with chiasmatic gliomas are unknown. There are no preventative measures that will avert the development of optic pathway tumors. Because of the higher incidence of these tumors in patients with neurofibromatosis type 1, routine screening of the visual axis is indicated at the time of initial presentation in these patients (64; 27).
Distinction between chiasmatic gliomas and other tumors that may involve the chiasm is usually not difficult. As stated previously, separation between chiasmatic gliomas and other lesions that may infiltrate the chiasm, such as hypothalamic gliomas and thalamic lesions, is usually arbitrary. Some chiasmatic lesions, especially when there is associated optic nerve involvement or when the patient has neurofibromatosis, may act benignly; in these cases it is impossible to determine whether the lesion is a true glioma or rather a hamartomatous abnormality. Two other tumor types commonly arise in the chiasmatic region in childhood, craniopharyngioma and suprasellar germinoma. The craniopharyngioma is usually radiographically distinguishable. Craniopharyngiomas are more common than chiasmatic gliomas and usually present, early in the course of illness, with nonspecific signs and symptoms of increased intracranial pressure due to obstruction of the third ventricle or foramen of Monro by superior extension of the tumor. Visual field deficits are common in craniopharyngiomas and occur in 50% to 90% of patients; homonomous hemianopsias and bitemporal hemianopsias are the most frequent abnormalities found. In addition, neuroendocrine deficits are present in as many as 90% of patients at diagnosis. Suprasellar germinomas, despite their histological aggressivity, often present insidiously. These lesions may cause endocrinologic dysfunction for many months prior to diagnosis, such as isolated growth hormone insufficiency and diabetes insipidus. Histiocytosis X may also primarily involve the hypothalamus with involvement of the chiasm. In these cases, diabetes insipidus is the most common presenting symptom, although visual impairment may occur. Higher-grade gliomas may arise in the chiasmatic region, but are more common in the thalamus and are much more common in older patients. Other tumors that may arise in the chiasmatic area include primitive neuroectodermal tumors and mixed germ cell tumors. Nontumorous lesions of the chiasm are relatively infrequent. Inflammatory masses may arise in the chiasm, but are uncommon without involvement of the basilar meninges. These include fungal granulomas and tuberculomas. Retro-orbital optochiasmatic arachnoiditis is clinically difficult to separate from chiasmatic tumors.
The diagnostic evaluation of children with chiasmatic gliomas centers on neuroimaging. Prior to the advent of CT and MRI, skull x-rays were often used to detect the presence of intraorbital disease. Skull x-rays disclose enlargement of the optic foramina with an intact margin of cortical bone (23). Criteria used to diagnose optic gliomas on skull x-rays included an optic foramen greater than 7 mm in diameter or a greater than 1.5 mm difference between foramina. Pressure of the chiasmatic tumor against the anterior clinoid process at the lateral aspect of the sphenoid bone produces the gourd shape, or so-called "j-shaped" sella. The anterior clinoid process tends to be thin, especially inferomedially, with preservations of a well-corticated margin. This type of cellular deformity is found in as high as one half of patients with intracranially lesions that involve the chiasm. Before the advent of CT scanning, pneumoencephalography was often needed to confirm the presence of an intracranial mass.
On CT, intraorbital tumors are usually seen as fusiform, isodense masses that contrast homogeneously after intravenous contrast (82; 33). Intracranial involvement, on CT, is usually demonstrated as an isodense, or less frequently, a hypodense mass, filling the suprachiasmatic cistern and often involving large portions of the diencephalon. Enhancement is variable, but there may be marked enhancement, even in low-grade lesions.
Areas of calcification are infrequently seen. Children with neurofibromatosis uncommonly have isolated chiasmatic involvement. More commonly, there is involvement of either 1 or both optic nerves or enhancement along the optic radiations, tracts, or into the occipital lobes. It is unclear whether these areas represent neoplastic tissue, or are dysplastic or hamartomatous regions. In children without neurofibromatosis, there is often involvement of other contiguous diencephalic sites, such as the thalamus and hypothalamus. The tumor may have a globular appearance, and there also may be associated cystic components.
MRI often shows a larger and more infiltrative tumor than appreciated on CT (83).
On T2-weighted images, gliomas are seen as high signal intensity areas involving the visual pathway. Coronal and sagittal images usually demonstrate anatomic relationships well.
As is the case on CT, the enhancement pattern on MRI is variable, but portions of the tumor usually enhance. At times, chiasmatic gliomas may be disseminated throughout the nervous system at diagnosis; this occurs primarily in young children with or without the pilomyxoid variant (04).
Even in young patients, every attempt should be made to test visual acuity and fields in patients with chiasmatic gliomas. Formal visual field evaluation procedures, such as Goldmann perimetry, are needed to determine the extent of visual field loss. Visual evoked responses are frequently abnormal in patients with chiasmatic gliomas, especially with orbital involvement. However, they have not been shown to be useful in following disease. Optical coherence tomography, used to detect retinal nerve fiber layer thinning, may be a useful adjunct in the sequential follow-up of those involved (20; 84; 07; 08).
Due to the frequent coexistence of hypothalamic dysfunction in children with chiasmatic gliomas, baseline endocrinologic studies are indicated in children at the time of diagnosis. In children without delayed growth, this includes measurements of growth hormone production. In pubertal children, sexual hormones should be measured. All patients require evaluation of thyroid function, as well as evaluation for diabetes insipidus and inappropriate secretion of antidiuretic hormone.
The most appropriate management for children with chiasmatic gliomas remains somewhat controversial. It is relatively well accepted that a period of observation is indicated, prior to the initiation of treatment, for children with chiasmatic lesions and neurofibromatosis (83). This is especially true for patients with primarily orbital lesions with minimal chiasmatic involvement. For patients with isolated or bilateral optic nerve gliomas, treatments with surgery or radiotherapy have been utilized (28). Radiotherapy has primarily been employed in patients with visual impairment, but still useful vision, to prevent progressive visual loss. Surgery has been primarily utilized in patients with isolated intraorbital lesions for cosmetic reasons. Although various types of surgical techniques have been utilized, most have attempted to spare the globe, if possible (96; 28). Surgery has also been recommended for patients with isolated orbital lesions to prevent the development of life-threatening, intracranial disease. However, it is unclear how frequent isolated intraorbital lesions will extend into the chiasm and it is also unclear how well surgical removal (or for that matter, radiotherapy) is able to prevent such intracranial spread.
Similarly, the management of chiasmatic lesions is unsettled. A period of observation is probably indicated for patients with neurofibromatosis and possibly for patients without neurofibromatosis with questionably progressive disease before instituting any specific treatment. Spontaneous partial regressions have been noted (85). Delay is often difficult in patients with histories consistent with relatively rapid or steady progression prior to the time of diagnosis. Surgery has been utilized for both the diagnosis and initial management of children with chiasmatic lesions. In children with neurofibromatosis, because the relative incidence of chiasmatic gliomas occurring in this region of brain is much higher than any other tumor type, it does not seem that surgery is warranted for diagnosis. However, in unusually aggressive lesions or those recalcitrant to initial therapy, a biopsy to determine if there are histologic or molecular genetic findings, suggesting a more malignant phenotype, may be indicated (90). On the other hand, in children without neurofibromatosis, most have recommended surgical confirmation before institution on other means of management. The majority of chiasmatic lesions are solid and not easily amenable to extensive surgical resections. However, there are occasional globular, partially cystic lesions; these can be subtotally resected (11). Such resections may cause increased neurologic and visual morbidity, especially in young children (112). However, at times, especially when the tumor is bulky and laterally exophytic, resection is more feasible. Resection may relieve obstructive hydrocephalus and occasionally, removal of a significant portion of the tumor results in disease stabilization, delaying the need for other types of treatment. The presence of different molecular subtypes of diencephalic gliomas increases the rationale for biopsy in non-NF1 patients. This is of significant benefit in young patients with large, progressive tumors.
The role of radiation therapy for patients with chiasmatic gliomas has not been unequivocally established. There is fairly good evidence that radiation therapy, in doses ranging from 5000 to 5400 cGy, results in tumor shrinkage with apparent stabilization of disease (106; 44; 96; 104; 68; 74; 108). Five-year survival rates seem higher in patients who have undergone radiotherapy as compared to patients who have received no treatment. However, at 10 and 20 years, it is difficult to demonstrate a difference in survival (03). There have also been some reports of improved visual function after radiation, although this has not been a universal finding. In 1 report of 41 children treated with radiotherapy, the majority had stable vision at the mean follow-up of 5 years from treatment; however, 11.5% had visual decline from baseline (01). Ten point six percent of subjects had visual improvement. Visual change usually occurred within 2 years of radiotherapy and was not associated with radiographic change (01). In another retrospective review, “early” radiotherapy, defined as initial or first-line salvage therapy, was found to result in better blindness-free survival than later radiotherapy (41). Children with neurofibromatosis type 1 are at higher risk of developing secondary tumors after radiotherapy, compared to those with neurofibromatosis type 1 who do not receive irradiation (98; 108). Others have reported reasonable results after the use of brachytherapy with temporary Iodine-125 seeds (59).
The majority of patients, after treatment with radiotherapy, will have growth failure as well as other endocrinologic deficits (82; 50). The detrimental effects of extensive cranial radiotherapy on intellectual function in children are well documented (109). Malignant gliomas have been noted to develop in previously irradiated fields (45), and Moyamoya disease is a known sequelae of radiotherapy for a chiasmatic tumor, especially in younger patients and in patients with neurofibromatosis (89; 76; 57; 110; 108). This is especially a concern in young children with large lesions, as local radiotherapy would result in irradiation of large areas of developing cortex. Another issue concerning radiation therapy is the determination of the most appropriate radiation portal. Children with neurofibromatosis often have scattered areas within the brain of abnormal signal on MRI, which may or may not represent gliomatous areas (83). These lesions often arise close to, but are not contiguous with, the chiasmatic mass seen. Inclusion of these areas within the radiation portal will result in essentially whole brain irradiation being given to many children with neurofibromatosis. On the other hand, exclusion of these areas from the radiation portal may make it impossible to deliver adequate radiation in the future if 1 of the noncontiguous lesions demonstrates enlargement, due to concerns of field overlap. Because the median age at diagnosis of children with visual pathway gliomas is usually below 5 years, and because as high as 30% of patients are under 2 years of age at the time of diagnosis, patients can be expected to have a high incidence of damage after local radiotherapy. In a series, greater than 50% of children with chiasmatic gliomas had significant cognitive dysfunction after treatment with radiation (82; 72).
Because of the risk of neurologic damage due to cerebral vasculopathy and the risk of neuropsychological sequelae of radiation therapy, especially in young patients with large intracranial lesions and the risk of endocrinologic dysfunction, alternative means of treatment have been employed (92; 83; 34; 35; 87; 88). Treatment with chemotherapy using actinomycin D and vincristine has resulted in disease stabilization or regression of disease in 80% of children with newly diagnosed chiasmatic gliomas less than 5 years of age (83; 51). This treatment delayed the need for radiotherapy for a median of 3 years in these patients. Other chemotherapeutic regimens are presently under study. The widest experience has been with the combination of carboplatinum and vincristine (80). In a series, over 60% of children under 5 years of age with chiasmatic gliomas had a radiographic objective response to treatment. Approximately 70% of children with visual pathway gliomas were free of progressive disease 3 years following treatment (77). The long-term benefit of chemotherapy is unproven, and visual outcome is often unsatisfactory. However, the potential benefit of conservative treatment, with observation and the use of chemotherapy at time of progression instead of immediate aggressive resection, has been supported by the experience of Sutton and colleagues. In this series of 33 patients, such an approach resulted in 85% of patients surviving over 10 years after diagnosis, the majority of whom had functional vision and performed well in school (103). However, visual outcome may be impaired, especially in children without neurofibromatosis type 1 (18). Various drug regimens are presently under study for children with chiasmatic gliomas (71). In a study, 85 children were treated with alternating procarbazine plus carboplatin, etoposide plus cisplatin, and vincristine plus cyclophosphamide. This resulted in an objective response rate of 42% and a relapse-free rate of 61% (60). In 1 study, 85 children were treated with alternating procarbazine plus carboplatin, etoposide plus cisplatin and vincristine plus cyclophosphamide. This resulted in an objective response rate of 42% and a relapse-free rate of 61% (60). A randomized prospective study comparing the 2 regimens in children without neurofibromatosis type 1 found no difference in event-free survival in children younger than 10 years of age, with the 5-year event-free survival being 39% ± 4% for carboplatin/vincristine and 52% ± 5% for the 4-drug regimen. The choice of agents is also dependent on potential side effects, and the CCNU/procarbazine regimen was not used in children with neurofibromatosis type 1, due to concerns of mutagenesis (06). Children with neurofibromatosis treated with carboplatin and vincristine did better than those without NF1 (05). In non NF1 children carboplatin, vincristine and temozolomide was also effective (21). Vinblastine has been shown to be effective in children with newly-diagnosed disease or those who have failed treatment with first-line chemotherapy (14; 62). However, visual outcome after chemotherapy may be suboptimal (99; 32). Given the potential benefits of chemotherapy, some have even suggested using chemotherapy as the treatment of choice in adolescents (22).
The combination of bevacizumab and irinotecan has also shown efficacy in the recurrent setting and may result in improved vision (79; 09; 54; 38). Another approach actively being pursued is the use of molecular-targeted therapy blocking growth pathways believed active in tumors (43), such as the Ras-MAP Kinase pathway (MEK inhibitors and v600e inhibitors) and mTOR in both patients with and without neurofibromatosis type 1 (25; 26; 40; 39; 10; 81). The MEK inhibitors, in those with neurofibromatosis 1 and BRAF mutations, have likely shown the best efficacy (10; 31). In a phase II study of selumetinib, a MEK-inhibitor, of 25 children without neurofibromatosis type 1 with recurrent low-grade gliomas after treatment with carboplatin containing chemotherapy, 9 (36%) achieved a radiographic partial response (31). The majority of the remaining either had less marked tumor shrinkage or stable disease, and the 2-year progression-free survival was 70% ± 9%. All had tumors either with BRAF fusions or V600E mutations, and responses were seen in both molecular tumor types. In a separate arm of the same study, 25 children (18 with pilocytic astrocytomas, 1 with a ganglioglioma, and 6 without tissue confirmation) with radiographically-diagnosed optic pathway and hypothalamic low-grade gliomas were treated (30). Only 6 had molecular characterization, 3 were BRAF-fusion positive, 2 were negative, and 1 had inadequate tissue. Six of 25 had partial responses, and 14 had stable disease (an 80% benefit rate); visual acuity improved in 4 out of 19 (21%) of evaluable patients and worsened in 2 out of 19 (11%). Five of 19 had visual field improvement. The 2-year progression-free survival was 78 ± 8.5% (30). These results have led to a prospective phase III randomized trial in children with newly-diagnosed, non-totally resected, low-grade gliomas, including those of the visual pathway, comparing treatment with selumetinib to the combination of carboplatin and vincristine.
In the same phase III study of selumetinib for recurrent low-grade gliomas, a cohort of 25 children with neurofibromatosis type 1 and progressive low-grade gliomas were studied, and an even better response rate and 2-year progression-free survival was seen (96±4%), also leading to a newly-opened, similar randomized prospective study comparing the same agents (31). Other MEK inhibitors are in active phase II study for progressive low-grade gliomas (15; 91).
For those children with recurrent V600E BRAF mutated low-grade gliomas, responses have also been seen after treatment with V600E BRAF inhibitors (56; 61; 75), and a prospective randomized trial shows final analysis comparing one such drug, vemurafenib plus dabrafenib (an MEK-inhibitor), to carboplatin and vincristine chemotherapy for newly-diagnosed patients. The use of one V600E BRAF inhibitor, sorafenib, resulted in paradoxic tumor activation in BRAF-fusion low-grade gliomas, and such treatment is contraindicated (55). The toxicity profile of MEK inhibitors is different from most chemotherapies, and rash is a frequent side effect. The long-term sequelae of such biological therapies are not well characterized.
Another molecular-targeted agent, the mTOR inhibitor, everolimus, has been used in children (both in those with and without NF1) with progressive/recurrent low-grade gliomas, with some success. Partial responses are relatively infrequent (less than 20%), but at least stable disease on treatment is seen in over 80% of children with NF1 and approximately 40% of those without NF1 two years from initiation of therapy (111; 114).
Children with chiasmatic lesions, especially those without neurofibromatosis or those with large globular lesions that involve contiguous areas of brain or streak along the visual pathways, will frequently develop progressive disease. After treatment with chemotherapy, disease progression can be halted in the majority of patients; 5-year event-free survival is approximately 40% in those without neurofibromatosis type 1 (44; 96; 82; 06; 37). Ten-year event-free survival rates are poorer (03; 06; 37). Five-year progressive-free survival after molecular-targeted therapy is unknown. Chiasmatic tumors have been shown to have poorer prognoses than other hemispheric low-grade tumors, likely due to their unresectability, although other molecular differences have not been fully explored (102; 113; 37). Overall survival at 5 to 10 years remains good, at 80% to 90% (37). After treatment, there is usually no improvement or minimal improvement in visual and associated neurologic function. Up to one third of patients will have deterioration of vision despite radiographic stability (32).
Following irradiation, vascular abnormalities have been reported to occur in the chiasm, including the formation of hemorrhagic vascular masses that mimic arterial venous malformations, with resultant repeated hemorrhages and deaths (29; 110).
Approximately one third to one half of patients with chiasmatic gliomas will have increased intracranial pressure at the time of diagnosis. Anesthesia techniques for increased intracranial pressure must be utilized.
All contributors' financial relationships have been reviewed and mitigated to ensure that this and every other article is free from commercial bias.
Roger J Packer MD
Dr. Packer of Children’s National Medical Center and George Washington University has no relevant financial relationships to disclose.See Profile
Nearly 3,000 illustrations, including video clips of neurologic disorders.
Every article is reviewed by our esteemed Editorial Board for accuracy and currency.
Full spectrum of neurology in 1,200 comprehensive articles.
Listen to MedLink on the go with Audio versions of each article.
Stroke & Vascular Disorders
Oct. 26, 2023
Oct. 26, 2023
Oct. 17, 2023
Oct. 17, 2023
Oct. 17, 2023
Oct. 17, 2023
Oct. 17, 2023
Oct. 17, 2023