optic nerve chasmal hypothalamic tumors.pdf

158

5

Optic Nerve, Chiasmal, and Hypothalamic Tumors

JOANN ATER, NANCY J. TARBELL, AND EDWARD LAWS, JR.

Gliomas are the most common tumors in the opticnerve, chiasmal, and hypothalamic regions of the cen-tral nervous system (CNS). As such, they are the fo-cus of this chapter. For completeness, the less com-mon tumors of these regionsmeningiomas andcraniopharyngiomasare also covered. Germ celltumors can also occur in this region but are discussedin Chapter 7.

GLIOMAS

Gliomas that affect the optic nerves, chiasm, and hy-pothalamus represent a unique type of tumor with avariable clinical course. Histologically, most othermidline astrocytomas of childhood are of the pilo-cytic subtype. These gliomas are among the neo-plasms of the nervous system whose tumor type andprognosis are age related. Except for infants, theprognosis for patients with these tumors is inverselyrelated to age at onset, with older individuals havinga poorer prognosis. In infancy, tumors affecting theoptic pathways can be malignant in their course, al-though the reasons for this are not known. Gliomasof the optic nerves and chiasm are strongly associ-ated with neurofibromatosis type 1. Several large se-ries report obvious signs of neurofibromatosis in asmany as 54% of affected children (Alvord and Lofton,1988; Hoyt and Baghdassarian, 1969; Listernick etal., 1988; Packer et al., 1983; Manera et al., 1994).Gliomas affecting the hypothalamus and anterior third

ventricle are also strongly associated with neurofi-bromatosis and may be found in tuberous sclerosis,another hereditary condition.

The pathology of optic pathway gliomas runs thegamut from very benign astrocytomas, considered bysome to be hamartomas, to tumors that are glioblas-toma multiforme. The typical histologic picture of aglioma of the optic nerve is one of dense arachnoidproliferation around an infiltrating pilocytic glioma,with thin hair-like tumor cells intermixed among thefibers of the optic nerve itself. The low-grade gliomasthat tend to affect the optic chiasm, anterior third ven-tricle, and hypothalamus frequently are characterizedas juvenile pilocytic astrocytomas, having few mitoses,no malignant features, or degenerative changes suchas Rosenthal fibers. Despite their relatively benignhistology, these tumors can progress and cause con-siderable morbidity in young children. Occasionally,anterior third ventricle tumors are discovered in con-junction with tuberous sclerosis; these tumors aregenerally noninfiltrating, relatively benign subependy-mal giant cell astrocytoma (see Chapter 3). Overall,approximately 4% to 5% of optic pathway tumors arefrankly malignant, and those usually have many of thecharacteristics typical of glioblastoma multiforme.The tumors with malignant histology occur mostcommonly in adolescents and older individuals.

In addition to patient age, anatomic distinctionsare extremely important in the evolution and prog-nosis of these tumors. Optic nerve gliomas can beconveniently grouped into two major categories: the

3601_e05_p158-170 2/15/02 4:31 PM Page 158

anterior optic nerve glioma, which primarily affectsthe optic nerve or nerves; and the posterior opticnerve glioma, usually centered in the optic chiasm.Obviously, tumors in both categories affect the visualsystem, but the two types differ in pace and progres-sion. Anterior optic nerve gliomas, which usually oc-cur in childhood, are ordinarily quite benign andprogress very slowly. Some of these tumors do notprogress at all or progress over many years. Poste-rior optic nerve gliomas, which occur in very youngchildren or older individuals, tend to form largermasses and present with more symptoms. These tu-mors may become large enough to affect the physi-ology of the hypothalamus and/or obstruct the ante-rior third ventricle, producing hydrocephalus. Ininfants who present with optic nerve or chiasmalgliomas, the spectrum ranges from indolent tumorsto aggressive tumors that can spread throughout theoptic pathway from the globe back to the occipitalcortex.

Tumors arising primarily in the hypothalamus oranterior third ventricle are less common and less of-ten associated with neurofibromatosis. Hamartoma-tous lesions also occur in the hypothalamus and inthe interpeduncular fossa. More typical juvenile as-trocytomas can occur in this region, along with stan-dard anaplastic astrocytomas and other malignantforms.

Clinical Presentation

Optic gliomas occur primarily in children, with morethan 71% diagnosed in patients younger than 10 yearsof age and 90% diagnosed during the first twodecades of life (Dutton, 1991). The tumors can rangefrom mild fusiform enlargement of the optic nerve ornerves within the orbit to very large, globular exo-phytic masses that extend from the chiasm and arevirtually indistinguishable from a primary hypothala-mic tumor.

In one series, more than 60% of optic pathway tu-mors involved the optic chiasm (Tenny et al., 1982).The signs and symptoms in children with optic path-way tumors who presented to The University of TexasM. D. Anderson Cancer Center between 1975 and1993 are listed in Table 51 (Manera et al., 1994).The clinical picture of a patient with a lesion affect-ing the optic nerves, chiasm, or hypothalamus is usu-ally one of progressive visual loss. In unilateral opticnerve tumors, this begins as a unilateral loss of op-

tic nerve function; in other tumors, mixed variants ofoptic nerve and chiasmal patterns of visual loss canoccur, with an asymmetric bitemporal hemianopsiabeing the most common finding in a chiasmal glioma.

In addition, behavioral changes, possibly relatedto elevated intracranial pressure or hypothalamic in-volvement, are prominent. Irritability, depression, so-cial withdrawal, somnolence, and aggressive behav-ior have been reported. Because of the frequentinvolvement of the suprasellar-hypothalamic region,children with optic nerve tumors of these areas canalso present with endocrine abnormalities. Althoughendocrine manifestations can occur with any of thesuprasellar lesions, such presentations are particu-larly common in lesions that arise in the hypothala-mus or floor of the third ventricle. The hypothalamicdysfunction produced by these lesions can range fromvarying forms and degrees of hypopituitarism to en-docrine-active syndromes produced by tumors thatsecrete hypothalamic-releasing factors. Tumors thataffect the physiology of the appropriate nuclei in thehypothalamus or of the pituitary stalk can result indiabetes insipidus. Finally, hypothalamic hamartomasthat present in the interpeduncular fossa are also as-sociated with precocious puberty. In a report of 33children with optic chiasmatic-hypothalamic tumors,5 (14%) of 33 presented with symptoms of endocrinedysfunction and 14 (56%) of 25 demonstrated en-docrine abnormalities on laboratory evaluation.Growth hormone deficiency was the most commonabnormality, followed by precocious puberty, delayedpuberty, and diabetes insipidus. In addition, 7 (21%)of 33 patients failed to thrive and had the diencephalicsyndrome (Rodriguez et al., 1990), which is charac-terized by severe emaciation and an inability to gainweight even when caloric intake is adequate (Russell,1951).

Evaluations of endocrine function in children withdiencephalic syndrome usually reveal normal thyroidfunction and elevated levels of cortisol and growth hor-mone. Usually the child is young at the time of diag-nosis and frequently has been subjected to extensivefailure-to-thrive evaluations before the diagnosis ismade. Because the only neurologic findings on exam-ination may be decreased visual acuity, visual field cuts,optic atrophy, or nystagmus, which are difficult to eval-uate in a child younger than 3 years, these signs maybe overlooked in a less than thorough examination.

The association of optic nerve gliomas with neu-rofibromatosis is well known. Optic nerve gliomas ac-

Optic Nerve, Chiasmal, and Hypothalamic Tumors 159

3601_e05_p158-170 2/15/02 4:31 PM Page 159

count for only 4% to 8% of all brain tumors in child-hood (Pollock, 1994), but as many as 70% of the op-tic nerve glioma cases are found in individuals withneurofibromatosis type 1 (Stern et al., 1979). In aprospective study of children referred to a neurofi-bromatosis clinic who had no specific ocular com-plaints, 15% were found to have optic nerve gliomas,30% unilateral, 30% bilateral, and 40% involving theoptic chiasm (Listernick et al., 1989). In addition, allchildren who had plexiform neurofibromas of theeyelid and glaucoma were found through compre-hensive neuroimaging to have optic nerve gliomas.Whether the natural history of these tumors in chil-dren with neurofibromatosis is the same or differentfrom the rest of the population remains controver-sial.

Prognosis and Natural History

The natural history of optic pathway tumors has beendebated for nearly a century, with some early inves-

tigators (Hoyt and Baghdassarian, 1969) believingthat these tumors are not neoplasms, but rather arehamartomas that do not grow continuously. From theliterature, however, it is clear that the clinical courseof optic pathway tumors can be quite variable, rang-ing from rare reports of spontaneous tumor regres-sion (Brzowski et al., 1992), to tumors that remainstable for life, as suggested by Hoyt and Baghdassar-ian (1969), to aggressive tumors that over time carryconsiderable risk of visual loss and death (Alvord andLofton, 1988). Several factors have now been identi-fied that at diagnosis predict favorable and poor out-comes (Kanamori et al., 1985). Table 52 summa-rizes these factors.

Most investigators have divided optic pathway tu-mors into two groups: anterior optic nerve gliomaswith isolated optic nerve enlargement and posterioroptic nerve gliomas with optic chiasmal involve-ment. Ten to 20 year survival rates are excellent (ap-proximately 90%) for patients with optic nerve tu-mors (Weiss et al., 1987) and more variable (40%

160 PRIMARY CENTRAL NERVOUS SYSTEM TUMORS

Table 51. Optic Pathway/Hypothalamic Tumors Referred to the Pediatric Brain TumorClinic at The University of Texas M. D. Anderson Cancer Center, 1980 to 1993*

No. %

Demographics

Total 60 100

Neurofibromatosis (NF) 31 54

Male 34 57

Female 26 43

Symptoms at diagnosis

Decreased visual acuity or blindness 28 47

Visual field deficit 12 20

Nausea/vomiting 17 46

Headache 19 32

Failure to thrive and diencephalic syndrome 6 10

Behavioral problems (irritability, social 12 20withdrawal, somnolence, aggressive behavior)

No symptoms with NF 12 7

Endocrine complaints 4 7

Radiographic findings

Multilobular suprasellar-optic chiasmal masses 35 58

Optic nerve and chiasmal swelling 17 32

Isolated optic nerve 6 10

Hydrocephalus 23 38*Median age at diagnosis was 5.2 (range, 0.75 to 14.3) years.

3601_e05_p158-170 2/15/02 4:31 PM Page 160

to 90%) for those with optic chiasmal tumors(Packer et al., 1983; Pierce et al., 1990; Horwichand Bloom, 1985; Tao et al., 1997). Upon carefulexamination, it can be observed that chiasmal in-volvement that is not extensive and not associatedwith a large exophytic mass may also carry an ex-cellent prognosis. The characteristics of tumors withthe worst prognosis include early onset in infancy,hypothalamic symptoms, signs of hydrocephalus,presence of diencephalic syndrome, third ventricu-lar involvement, and large chiasmal tumors extend-ing posteriorly. It is most difficult to determine thebest treatment for young children with these char-acteristics because aggressive treatment with sur-gery and irradiation do not necessarily lead to thebest survival rates or the best quality of life (Jan-noun and Bloom, 1990).

In evaluating the effects of the tumor itself and thetreatment of optic chiasmal gliomas, the series fromSan Francisco (Hoyt and Baghdassarian, 1969; Imesand Hoyt, 1986) is useful because of its long-termfollow-up period and its evaluation of the actualcauses of death. In the original 1969 report, 8 of 28patients were dead, and at follow up 15 years later 8more had died, leaving only 12 (46%) of 28 surviv-ing at a median follow up of 20 years. Nine of the 16deaths occurred in patients with neurofibromatosis;

only two of these patients had died of their chiasmaltumors. The remaining died of other malignantgliomas of the brain, neurofibrosarcomas of periph-eral nerves, or complications of management of cer-vical neurofibromas. Of the seven who died withoutneurofibromatosis, five died as a result of tumor andthree died of unrelated medical illnesses. Of those pa-tients treated with radiation, 4 of 14 patients died be-cause of their tumor, whereas only 5 of 14 who didnot receive radiation died, 1 from tumor and 4 fromother causes.

On the basis of these data, no benefit from radio-therapy (RT) could be demonstrated. In addition, therisk of death from tumor was greatest in the early follow-up period. However, most other investigatorshave concluded that RT does improve survival anddoes prolong the interval before disease progression(Alvord and Lofton, 1988; Pierce et al., 1990; Hor-wich and Bloom, 1985; Tao et al., 1997). For exam-ple, in a series of 26 children with chiasmal gliomastreated with RT at the Joint Center for Radiation Ther-apy, 60% had objective tumor shrinkage that oc-curred gradually over a period of 5 years. Vision ei-ther improved or stabilized in 72.7% of the children.The 15 year overall survival rate was 85.1% and free-dom from progression was 82.1%, with median fol-low up of 108 months (Tao et al., 1997).


Table 52. Classification of Optic Glioma by Factors Influencing Prognosis

FAVORABLE PROGNOSIS

Age at onset Early childhood to adolescence

Clinical features Visual loss with laterality

Slowly progressive or arrested course

Incidental finding in child with neurofibromatosis

No symptoms of endocrine dysfunction or hydrocephalus

Does not have diencephalic syndrome

Radiographic features Intrinsic optic nerve and/or chiasmal location

POOR PROGNOSIS

Age at onset Infancy to early childhood and adulthood

Clinical feature Hypothalamic symptoms and/or signs of increasedintracranial pressure

Severely affected vision in both eyes

Radiographic features Large exophytic chiasmal tumor with posterior extension

Extension into third ventricle

HydrocephalusAdapted from Kanamori et al. (1985) and Alvord and Lofton (1988).

3601_e05_p158-170 2/15/02 4:31 PM Page 161

Management of Optic Gliomas

The literature abounds with inconsistencies, contro-versy, and disparate conclusions about the progno-sis, natural history, and management of optic path-way gliomas. Although only 4% to 8% of childhoodbrain tumors originate from the optic tract, the po-tential morbidity of these tumors and their treatmentin the face of good survival rates has resulted in ex-tensive literature about the best forms of treatment tooptimize cure rates and minimize morbidity. How-ever, because of the relative infrequency of occur-rence of these tumors and their heterogeneous be-havior related to patient age and tumor location andsize, most series have not reported numbers adequateto allow definitive conclusions about this relativelyrare subgroup of gliomas, and randomized trialscould not be conducted with them. Furthermore, asSutton et al. (1995) aptly stated, It is unlikely thatany single modality (surgery, RT, or chemotherapy)will be the optimum treatment for all children withhypothalamic/chiasmatic astrocytoma. The challengefor the future is to determine the most appropriatetreatment for each patient, based on rate of tumorprogression, age, radiographic demonstration of ex-tension of tumor, prior therapy, and visual/endocrinestatus. It is therefore extremely difficult to arrive atany standard recommended treatment for these tu-mors. Recognizing that controversies exist, we haveadapted the following guidelines for the evaluation,treatment, and follow up of children with optic path-way gliomas.

Diagnosis

The evaluation of patients with optic pathway gliomasinvolves a thorough family history, an accurate as-sessment of visual status, evaluation for signs andsymptoms of increased intracranial pressure, and de-lineation of the patients endocrine status, looking for both hypopituitarism and endocrine-active syn-dromes. Physical examination should be directed to-ward these points, noting the presence of papilledemaor optic atrophy, deficiencies in visual acuity or vi-sual fields, and the general intellectual and neuro-logic state of the patient. Careful attention should bepaid to growth and development parameters and tothe presence of any lesions suggestive of neurofibro-matosis or tuberous sclerosis. For asymptomatic chil-dren with neurofibromatosis with no previous diag-nosis of optic glioma, routine screening with imagingor visual evoked potentials is not warranted, and tests

should be determined by findings on clinical exami-nation (Gutmann et al., 1997).

Diagnostic evaluation consists of appropriate lab-oratory testing, including measurement of pertinentpituitary hormones, a formal visual examination andmeasurement of acuity and visual fields, and imagingdiagnosis, which currently rests on magnetic reso-nance imaging (MRI) with gadolinium enhancementfor the most accurate delineation of the lesions in-volved.

In children with neurofibromatosis and optic nerveenlargement, characteristic findings on MRI or com-puted tomography (CT) scans are adequate to allowdiagnosis. Unless there is a history of acute visual lossor neurologic changes, these such patients can beevaluated and followed up for objective evidence ofprogression. The baseline and follow-up evaluationsshould include complete physical and neurologic ex-amination, careful ophthalmologic examination, in-cluding visual fields and MRI, or CT scan evaluations.The MRI scan is superior to the CT scan for detect-ing change, relationship of tumor to the optic chiasm,and tumor extension into adjacent brain. Visualevoked potentials can be useful if the baseline valueis normal and can be very sensitive in detecting dis-ease progression. Once vision is impaired, however,we have not found the visual evoked potentials to bevery useful, especially when visual field defects arepresent. Unfortunately, for young, uncooperative chil-dren the visual evoked potential studies were not asuseful as we had hoped. For very young children, themost useful evaluation of vision appears to be thatperformed by a child neurologist or pediatric oph-thalmologist.

At diagnosis, it is often unclear from the patientshistory how rapidly visual change is occurring; there-fore, for the first 6 months to 1 year, we perform radiographic and physical examinations every 3months. If no change is observed, evaluation inter-vals can be safely decreased to yearly. For childrenwithout neurofibromatosis, these guidelines can alsobe followed in cases of isolated optic nerve enlarge-ment. When tumor progression is identified, optionsfor further treatment include surgery, radiation, andchemotherapy. The pros and cons of these ap-proaches are discussed separately.

Surgery

When a suprasellar mass is present at diagnosis, sur-gical resection or biopsy is usually recommended.


3601_e05_p158-170 2/15/02 4:31 PM Page 162

The extent of resection depends on the extent and lo-cation of the tumor. A biopsy is necessary to confirmthe diagnosis in patients who present with a suprasel-lar-hypothalamic mass. Frequently, the origin of thetumor cannot be determined by radiography, andcraniopharyngiomas may be indistinguishable fromsuprasellar germinomas. Because management ofthese two entities differs somewhat, a definitive diag-nosis is needed.

In addition, careful surgical removal from the chi-asm of the portion of the tumor that is exophytic cansometimes improve vision by relieving external pres-sure on the adjacent optic nerve (Oakes, 1990).Sometimes surgical debulking can also relieve hy-drocephalus. These goals must, however, be balancedagainst the risks of increased visual loss and in-creased postoperative hypothalamic dysfunction,which can result in a disturbed sleepwake cycle, dis-torted appetite and thirst, hyperactivity, memory dys-function, and panhypopituitarism.

The indication for surgery varies with the type andlocation of the tumors affecting the optic pathways.For the typical unilateral optic nerve glioma locatedwithin the orbit, the indication for surgery is pro-gressive visual loss and progressive proptosis. Sur-gery is generally the treatment of choice when thereis loss of vision in an eye without extension of tumorinto the chiasm. Surgical excision of the lesion whenit has not reached the optic chiasm can be curative,but the eye remains blind. Current surgical techniquesallow for preservation of the globe and a good cos-metic result. Patients known to have optic nervegliomas with little proptosis and preserved functionalvision can be evaluated with periodic imaging stud-ies and visual assessments. If there is any evidence ofthe tumor extending toward the optic chiasm, treat-ment should be planned early. When surgery is indi-cated, the operation involves a frontal craniotomy andunroofing of the orbit, sectioning of the optic nerveat its junction with the globe, and removal of the op-tic nerve, including its intracanalicular segment up toits junction with the optic chiasm. Results are excel-lent provided that the tumor is totally excised and theremaining optic nerve is free of disease.

Astrocytomas that involve the optic chiasm cannotbe resected without causing significant visual impair-ment. Unfortunately, the characteristics of this type oftumor, as shown by neuroimaging scans, are still notspecific enough to allow a histologic diagnosis with-out biopsy. In these cases, the goal is to perform asafe but effective biopsy of the lesion without pro-

ducing additional visual impairment. This is ordinar-ily accomplished with a frontotemporal type of cran-iotomy using microsurgical techniques for the tumorbiopsy. Some tumors in this region are large enoughso that the exophytic component extending from thechiasm produces obstructive hydrocephalus; in suchcases, a tumor debulking that preserves the portioninvolving the optic chiasm can be accomplished torelieve the ventricular obstruction. This procedurecan be performed accurately and safely using carefulmicrosurgical techniques. There are reports of verysatisfactory results of removal of some hypothalamichamartomas using similar techniques, with reversalof some of the endocrine deficits, particularly preco-cious puberty.

Despite the risks, several neurosurgical groupshave advocated radical resection as primary treatmentfor children with hypothalamic gliomas. Wisoff et al.(1990) reported a series of 16 children with chias-matic-hypothalamic tumors treated with radical re-section, with 11 of 16 alive and well 4 months to4.5 years after surgery, most without other therapy.Infants were most likely to progress after surgery andrequire other therapy. It is evident that significant sur-gical judgment and skill are necessary to deal withthese difficult lesions, as the dysfunction produced by overzealous resections can have serious, life-threatening consequences, such as memory loss, in-appropriate thirst, and severe diabetes insipidus,which can ultimately result in an individuals requir-ing constant care. In addition, of 11 children with di-encephalic syndrome after surgical intervention in anM. D. Anderson Cancer Center series (Manera et al.,1994), 9 (82%) eventually became obese and suf-fered multiple endocrine deficits. The progression tomorbid obesity and endocrine deficits can also occurafter RT and during the natural course of tumor pro-gression, but the manifestation is usually not acute.

When the tumor is infiltrative, extensive, and diffi-cult to remove in bulk, hydrocephalus may be treatedwith a shunting procedure. Depending on the cir-cumstances, one can consider either a ventricu-loperitoneal or a ventriculocisternal (Torkildsen)type of shunt procedure.

In summary, we recommend a conservative surgi-cal approach primarily for diagnosis. Once the diag-nosis is made, children who exhibit favorable char-acteristics are followed up until signs of tumorprogression occur. For those who show unfavorablecharacteristics (Table 52), either RT or chemo-therapy is recommended for most, depending on the


3601_e05_p158-170 2/15/02 4:31 PM Page 163

age of the individual. For patients who have extensivetumors invading the hypothalamus, extending to thethird ventricular region, with massive infiltrationalong the optic tracts, or with clear-cut evidence ofrapidly progressive disease at the time of diagnosis,a delay in treatment is not recommended. However,in those few cases of tumors where surgical decom-pression and improvement of vision have occurred,especially in young children, very close follow upwithout intervention until objective signs of progres-sion occur is also an acceptable alternative.

Radiation

Most modern reports utilizing megavoltage RT docu-ment an advantage for patients who have progressivechiasmal gliomas (Pierce et al., 1990; Horwich andBloom, 1985; Wong et al., 1987; Tao et al., 1997).Radiation therapy is generally the treatment of choicefor symptomatic chiasmatic/hypothalamic gliomas inolder children. Many recent series report excellentsurvival after RTgenerally 90% at 10 years (Pierceet al., 1990; Horwich and Bloom, 1985; Tao et al.,1997). However, deaths can occur from disease pro-gression many years after treatment, and, thus, long-term follow up is critical in the management of thisdisease.

Outcome in terms of vision is an important mea-sure of treatment success for chiasmal/hypothalamicgliomas. Following RT, vision is improved in approx-imately one-third of patients, with most patients ex-periencing visual stabilization (Pierce et al., 1990;Tao et al., 1997). This success in maintaining or im-proving vision is possible only if treatment is initiatedbefore severe visual damage has occurred. Therefore,documented visual deterioration is a major indica-tion for the prompt initiation of therapy.

The overall survival rate for patients with optic sys-tem gliomas is excellent. However, conventional RThas been associated with significant morbidity. Mostradiation fields cover not only the tumor bed (tumorvolume) but also tissues thought to be at risk for mi-croscopic disease to allow for uncertainty in tumordefinition and for inconsistencies in the daily treat-ment set-up (target volume). The tolerance of thenormal brain parenchyma and its vascular and sup-porting structures becomes, therefore, the limitingparameter of external-beam therapy, and the risks of acute and long-term sequelae are major dose-limiting factors. Permanent radiation injury can in-

clude pituitary-hypothalamic dysfunction as well asmemory and intellectual deficits. Young children areat greater risk than adults (Glauser and Packer, 1991;Ellenberg et al., 1987; Ater et al., 1999). After irra-diation, 72% of children treated at the Joint Centerfor Radiation Therapy developed new onset of hy-popituitarism, most commonly growth hormone de-ficiency in 59%, with panhypopituitarism in 21% (Taoet al., 1997).

With conventional fractionation schedules (1.8Gy/day), total doses of 50 to 54 Gy are consideredstandard for the treatment of optic gliomas. Late ef-fects appear in a predictable manner in terms of ra-diation dose, volume, and fractionation. Fractionationexploits the differences in response to irradiation be-tween normal brain and tumor tissue; normal tissuestolerate multiple small doses of irradiation much bet-ter than they tolerate a single, large fraction.

Until recently, greater precision in the delivery ofconventional RT was limited by an incomplete diag-nostic definition of tumor volumes, unsophisticatedtreatment planning systems, and imprecise immobi-lization devices. Computed tomography and MRI nowprovide much improved delineation of CNS neo-plasms, and three-dimensional treatment planningsystems are currently available. These technologicaladvances allow for accurate focal administration of adose to the target area and have thus promoted thewidespread use of radiosurgery techniques.

Stereotactic Radiosurgery

Stereotactic radiosurgery is a highly accurate andprecise technique that utilizes stereotactically di-rected convergent beams of ionizing radiation to treata small and distinct volume of tissue with a single radiation dose. The multiple-beam approach of ra-diosurgery results in sharp dose fall-off beyond thetarget, thus sparing adjacent normal tissue. The tech-nique must, however, be reserved for select small le-sions because it ablates both normal and abnormaltissue within the treatment volume. Some investiga-tors have advocated using stereotactic radiosurgeryto reduce the treatment volumes of discrete, well-circumscribed lesions, although certain parameters,including the size and location of the target volume,are associated with complications from radiosurgery(Marks and Spencer, 1991; Loeffler and Alexander,1993; Tishler et al., 1993). Certain intracranial le-sions cannot be treated safely or effectively with ra-


3601_e05_p158-170 2/15/02 4:31 PM Page 164

diosurgery once the target volume becomes relativelylarge or is located near brain stem, retina, and theoptic pathways. For example, the maximum radiationtolerance of the optic nerve appears to be between 8and 10 Gy; if more than 1 cm of the eighth nerve istreated with radiosurgery, hearing loss is predictableeven with doses as low as 15 Gy (Tishler et al., 1993).Kjellberg and others have published isoeffect datapredicting the incidence rates of brain necrosis us-ing a proton facility (Kjellberg et al., 1983; Flickinger,1989). These isoeffect curves demonstrate the rela-tionship between tumor necrosis, radiation dose, andfield size and demonstrate the limitations of usinglarge single fractions for intracranial lesions that in-volve critical structures such as the optic system.Therefore, although stereotactic radiosurgery is precise in the administration of large single frac-tions, complications associated with larger volumes(greater than 3 cm) and with certain locations (brainstem, visual pathways) limit the use of this procedurein the primary management of pediatric tumors, par-ticularly in the management of patients with opticpathway tumors.

Stereotactic Radiotherapy

Fractionation of the daily dose of radiation combinedwith the precision of radiosurgical techniques may bethe optimal way to treat relatively small, symptomaticoptic tumors that do not show extensive involvementalong the optic tracts. Stereotactic RT is defined asthe use of stereotactic radiosurgery hardware andsoftware (stereotactic head frame and support sys-tem, small-field collimators, and three-dimensionalplaning) combined with radiation routine fractiona-tion (1.8 Gy/day) or some form of altered fractiona-tion such as hypofractionation (a few large fractionsof 4.0 to 8.0 Gy). Basic requirements necessary toadminister stereotactic RT include specially designedsoftware and reproducible repeat head fixation andlocalization systems.

Dose-optimization treatment using stereotactic RTor other conformal techniques has now become routine for lesions that are well controlled by con-ventional RT. These RT techniques may become thetreatment of choice for many diseases such as in-completely resected craniopharyngioma, pituitary ad-enoma, and small optic pathway tumors. The radia-tion dose to nearby nontarget volume structures vitalfor memory (mesial temporal lobe), for endocrine

function (hypothalamic-pituitary axis), and for nor-mal structural development (skull, mandible, and softtissues of the scalp) is markedly reduced with thesetechniques. This technique of dose optimization isparticularly important in the pediatric population. Formany pediatric intracranial tumors, focal RT tech-niques will largely replace conventional RT in orderto reduce the long-term side effects of therapy (Dun-bar et al., 1994; Loeffler et al., 1999).

In general, the use of stereotactic techniques asdefinitive treatment should be restricted to lesionsthat (1) are distinct on imaging scans, (2) are of rel-atively small volume, and (3) are noninvasive or non-infiltrating. Although stereotactic techniques do notreplace large-field RT in the treatment of widely in-filtrating or seeding tumors, it is clear that conven-tional RT is no longer acceptable for a large sub-group of patients who have more focal intracranialtumors.

Chemotherapy for Optic Chiasmal Tumors

The use of chemotherapy for low-grade astrocytomasin children, especially optic chiasmal-hypothalamictumors, has been investigated at several medical cen-ters. In 1977, Packer and the group at Childrens Hos-pital of Philadelphia started to treat patients youngerthan 6 years of age newly diagnosed with intracranialvisual pathway gliomas with combination chemother-apy. Their justification for this approach was that thebeneficial effects of radiation on vision could not beconfirmed in their patient population as only 1 of 21children demonstrated visual improvement after RT(Packer et al., 1983). In addition, these investigatorsfound that progressive neurologic deterioration andvisual loss did occur in patients who received radia-tion late in the course of their disease, usually 5 to10 years after diagnosis.

Between 1977 and 1988, 32 children younger than6 years of age were treated with vincristine and actin-omycin D chemotherapy as initial therapy after diag-nosis. At last report, 10 (31%) remained free of pro-gressive disease and had not required additionaltherapy (Janss et al., 1995). For those whose diseaseprogressed, the median time was 27 months after theinitiation of treatment. Ten year overall survival forthe entire group was 85% because of the success ofsalvage treatment.

Various chemotherapeutic agents, including lo-mustine; vincristine; a combination of procarbazine,


3601_e05_p158-170 2/15/02 4:31 PM Page 165

lomustine, and vincristine; and cisplatin-containingcombinations, have been somewhat effective in pa-tients with recurrent low-grade gliomas (Edwards etal., 1980) and have been utilized for optic chiasmaltumors. In an M. D. Anderson Cancer Center trial ofnitrogen mustard, vincristine, procarbazine, andprednisone (MOPP) given to children younger than3 years of age with low-grade astrocytomas; six chil-dren either had hypothalamic or brain stem lesions.With a median follow up of more than 7 years, all pa-tients survived with stable disease. However, five ofsix eventually received RT for tumor progression at amedian of 1 year after diagnosis (Ater et al., 1988).

Combination chemotherapy with 6-thioguanine,procarbazine, dibromodulcitol, lomustine, and vin-cristine has been substituted effectively for RT for chil-dren with chiasmal and hypothalamic astrocytomas.Investigators at the University of California at SanFrancisco (Petronio et al., 1991) initially reported re-sults for 19 infants and children (median age, 3.2years) with chiasmal and hypothalamic gliomas who received chemotherapy, 12 at diagnosis and 7 atthe time of tumor progression. Most received 6-thioguanine, procarbazine, dibromodulcitol, lomus-tine, and vincristine chemotherapy, and two receivedother combinations. Of the 18 patients with evaluabledisease initially managed with chemotherapy, tumorsin 15 (83%) either responded to therapy or stabi-lized. With a median follow-up period of 18 months,all are surviving; disease progressed in only 4 of 15and was successfully treated with radiation. Vision ini-tially improved or stabilized in 16 (88%) patients.This series was updated in 1997 and now includes agroup of 42 children with a mean age of 5 years. Themedian time to progression was 132 weeks, with a 5year survival rate of 78% (95% CI, 60% to 87%) (Pra-dos et al., 1997).

In low-grade hypothalamic and chiasmal gliomas,the criteria used to evaluate the usefulness of the che-motherapy are different from those in the usual phaseII studies that assess tumor response. For low-gradeastrocytomas, prolonged stable disease has been con-sidered a response by some investigators. Friedmanand the Pediatric Oncology Group (1992) studied theresponse of pediatric brain tumors to carboplatin.Based on results from 13 children with clearly pro-gressive, low-grade astrocytomas of the optic path-way, third ventricle, thalamus and suprasellar region,and temporal region, in which 73% achieved stabledisease and one had a partial response, Friedmans

group determined that carboplatin is active againstlow-grade astrocytomas. The duration of stable dis-ease in this subgroup of patients ranged from 3months to greater than 68 months (median, 40months).

A multi-institutional group studied the combina-tion of weekly low-dose carboplatin plus vincristinegiven for low-grade gliomas (Packer et al., 1993,1997). At the most recent report, 78 children withnewly diagnosed progressive low-grade gliomas witha median age of 3.1 years were treated with this reg-imen. Fifty-eight were chiasmatic-hypothalamic in lo-cation, and the remainder occurred in other loca-tions. Forty-five (56%) children had objective tumorresponse. Tumor response did not correlate withlength of disease control. The only significant factorpredictive of outcome was age, with a 2-year pro-gression-free survival rate for children younger than5 years at start of treatment of 81% compared with58% in older children ( p 0.01) (Packer et al.,1997).

Partly because of variability in prognostic factorssuch as age, the most effective regimen cannot begleaned from these single-arm studies. Therefore, anational randomized trial in the Childrens CancerGroup (CCG) is currently underway testing the effi-cacy of chemotherapy for progressive low-gradegliomas in children younger than 10 years old, com-paring the carboplatin-vincristine regimen to theCCNU-based regimen reported by Prados et al.(1997). Neuropsychological and endocrine outcomeof children treated with chemotherapy will also beevaluated in this trial.

Long-term Follow-up and Complications of Therapy

The use of chemotherapy for hypothalamic-chiasmalgliomas is gaining support not only because of thepreviously mentioned risks of extensive surgery butalso because of the consequences of conventional RT.For young patients, there is a risk of increased en-docrine deficits and intellectual impairment follow-ing RT delivered to the hypothalamic region (Mooreet al., 1992; Ater et al., 1997, 1999; Tao et al., 1997).Serial IQ scores before radiation showed no decline,but among those receiving radiation, IQ scores fell amedian of 12 points from baseline ( Janss et al.,1995). Also, several reports (Rajakulasingam et al.,1979; Mitchell et al., 1991) have recognized the risk


3601_e05_p158-170 2/15/02 4:31 PM Page 166

of radiation-induced moyamoya vascular change inthe suprasellar region, which can result in vasospasm,transient ischemic-type episodes, seizures, andstrokes. The actual incidence of moyamoya is notknown, but it may be higher than suspected becausethe symptoms may also be attributed to progressivedisease. The risk of moyamoya appears to be relatedto the patients age at radiation, and the condition hasbeen seen generally in children receiving radiationbefore 3 years of age or in association with neurofi-bromatosis type 1 (Poussaint et al., 1995).

When patients experience new symptoms that sug-gest progressive disease, especially many years aftertreatment, MRI or CT scans can provide essential in-formation, but definitively distinguishing progressivedisease from another cause can remain difficult. Attimes the MRI scan can be diagnostic, revealing hem-orrhage, stroke, or tumor growth. However, in a re-port by Epstein et al. (1992), three children with chi-asmatic-hypothalamic gliomas who had symptoms oftumor progression 9.5, 11.5, and 2 years after RTwere found to have misleading radiographic findings.Neuroradiographic studies including angiographyshowed large mass lesions. These were presumed tobe tumor recurrences and chemotherapy was initi-ated. However, on autopsy of two and biopsy of thethird, the bulk of the mass was found to consist ofnumerous vessels of variable size. The authors pro-posed that these lesions probably represented in-corporation of the rich vasculature in the chiasmalregion into the tumor, which underwent degenera-tion secondary to radiation therapy (Epstein et al.,1992). Further prospective evaluation of the vascularphenomenon associated with these tumors and theirtreatment is needed.

MENINGIOMAS


Meningiomas can occur anywhere within the craniumand are related to the arachnoid cap cells of the pac-chionian granulations, where spinal fluid is absorbedinto the venous sinuses. Meningiomas arise fromthese structures and are attached to the dura. Theyoccur most commonly in females, and several differ-ent subtypes of meningioma can specifically affect thevisual apparatus and the hypothalamus. Arising pe-ripherally, meningiomas may grow out of the optic

nerve sheath itself. These tumors tend to involve thedura of the optic nerve and ultimately strangulate thenerve; they may also occlude the blood supply to the ophthalmic artery. Direct surgery on these tumorsusually results in devascularization of the optic nerveand blindness. The indications for surgery are pro-gressive visual loss and proptosis, similar to the sce-nario with optic nerve gliomas but with a less favor-able prognosis, as meningiomas can extend readilyfrom the intraorbital segment of the optic nervesheath through the optic canal to involve the in-tracranial dura.

Meningiomas may also arise from the dura aroundthe optic foramen in which case these lesions maystrangulate the optic nerve and affect the ophthalmicartery. Both optic nerve sheath meningiomas andmeningiomas of the optic foramen can be bilateral.This is most commonly seen in optic foramen menin-giomas where tumor cells may bridge from one op-tic foramen to the other or may arise as two sepa-rate, nearly symmetric, lesions around the opticforamen. More common are meningiomas that arisefrom the dura of the planum sphenoidale or the tu-berculum sellae. The former cause optic nerve-typevisual loss, compressing the optic nerves from aboveand pushing them inferiorly. The latter tend to besuprasellar tumors and may affect either the opticnerves or the optic chiasm or both. Some menin-giomas arise from the dura of the diaphragma sellaeand also act like suprasellar tumors, producing chi-asmal-type visual loss and sometimes compressingthe pituitary stalk, causing distortions of pituitary-hypothalamic function. Tumors arising from the duraof the inner third of the sphenoid wing commonly in-volve both the cavernous sinus (and the nerves withinit) and the optic nerve on the same side. Patients af-flicted with these tumors may present with double vi-sion, ptosis, pupillary abnormalities, and optic nerve-type visual loss.

Management

Diagnosis

As with gliomas affecting the optic nerves and chiasm,patients with meningiomas in similar regions needcareful documentation of their visual function and vi-sual fields. Basic laboratory tests that include hor-monal evaluations are important. The diagnosticimaging method of choice is an MRI scan with


3601_e05_p158-170 2/15/02 4:31 PM Page 167

gadolinium contrast, which clearly shows the menin-gioma and frequently its areas of origin from the dura.Such scans accurately reveal the effects of the tumoron the surrounding anatomy and help guide the sur-geon in devising a safe and effective approach.

Surgery

Indications for surgery of meningiomas generally arethose of progressive enlargement of the tumor alongwith the progressive visual and neurologic signs thatmay accompany such growth. A number of menin-giomas reach a certain size and stop growing, so anargument can be made for careful follow up in somecases.

Basic surgical principles include the planning of acraniotomy that provides excellent exposure of the le-sion with the ability to protect and preserve normalneurologic structures. Adjuncts such as intraopera-tive corticosteroids and mannitol to shrink the braintemporarily are most helpful, and the surgery is car-ried out with precise microsurgical techniques. Oc-casionally, a laser or ultrasonic surgical aspirator al-lows the surgeon to manage difficult tumors that mayhave a very firm consistency.

Because the vast majority of meningiomas are be-nign, surgery may not be indicated for patients whosevision is preserved without it when curative surgerycould produce blindness or other forms of neuro-logic deficit. In such instances, RT has been benefi-cial in a reasonable number of patients.

Radiation Therapy

The largest use of radiation for menigiomas has beenwith conventional RT (Goldsmith et al., 1994). Post-operative RT is indicated for malignant menigiomas,subtotally resected tumors, tumors with atypical his-tologies, or multiply recurrent meningiomas. Focusedradiosurgery also has its role in the management ofsmall meningiomas, and many promising results havebeen reported (Hakim et al., 1998).

CRANIOPHARYNGIOMA


Craniopharyngiomas are developmental lesionsthought to arise from squamous remnants of Rathkespouch. Although these tumors tend to appear as tu-

mors of childhood, they can actually occur at any age;there are three basic types of clinical presentation thatare age related. In childhood, craniopharyngiomastend to be large, cystic suprasellar lesions that pre-sent as failure of growth and development, which arerelated to the tumors effects on the hypothalamus.Craniopharyngiomas may also present with progres-sive visual loss of the chiasmal type along with ob-structive hydrocephalus in large lesions that affect thethird ventricle. In young adulthood, craniopharyn-giomas tend to present in a fashion similar to pituitary adenomas. In women, the amenorrhea-galactorrhea syndrome is a common presentation andmay or may not be associated with progressive chi-asmal-type visual loss. Men may develop hypopitu-itarism and impotence along with visual symptoms.In the elderly, these tumors usually present with men-tal function changes, but may also produce increasedintracranial pressure and visual loss.

Management

Diagnosis

Medical evaluation for craniopharyngiomas shouldinclude a careful history and physical examination,paying particular attention in children to their growthand development, including secondary sexual char-acteristics, and to sexual function in older patients.Careful evaluation of the visual system, including vi-sual acuity and visual field determinations, should becarried out. Laboratory evaluation should include acareful review of pituitary hormone status. Becausecraniopharyngiomas frequently arise from the pitu-itary stalk, some patients, particularly children, pre-sent with diabetes insipidus; appropriate laboratorytests should be ordered if this is one of the featuresof clinical presentation.

Patients with increased intracranial pressure usu-ally have headaches and may have papilledema. Theimaging procedure used for diagnosis is an MRI scanwith gadolinium contrast. This modality usually isfairly diagnostic for craniopharyngioma. Becausemany of these lesions are calcified, a CT scan or evena plain skull X-ray may show the presence and posi-tion of calcified portions of the tumor. For the eval-uation of postoperative residual disease, CT and MRIscans are often complementary, with CT demonstrat-ing residual calcification (not easily seen on MRI)and MRI most often demonstrating possible residualcystic or solid craniopharnygioma.


3601_e05_p158-170 2/15/02 4:31 PM Page 168

Surgery

The surgical principles utilized in the management ofcraniopharyngioma are a subject of some contro-versy. It is clear that a proportion of these lesions,particularly cystic lesions in children, can be totallyexcised. Ordinarily this is accomplished using a cran-iotomy for those lesions that are suprasellar. Thecraniotomy procedure utilized to attack a cranio-pharyngioma can be tailored to the position and ex-tent of the lesion. Subfrontal, frontotemporal (pteri-onal), and a variety of skull base approaches can beutilized to approach and effectively remove these le-sions. A scrupulous microsurgical technique is es-sential and can provide good results in both extent oftumor removal and preservation or restoration of vi-sion. If a craniopharyngioma is associated with sig-nificant enlargement of the sella, then the tumor mayhave had its origin below the diaphragma sella andmay be amenable to total removal using thetranssphenoidal approach. For these lesions, the sizeof the sella, whether the tumor is primarily cystic orprimarily solid, and whether calcifications are pres-ent can be important factors in determining the ex-tent of debulking. Many suprasellar craniopharyn-giomas, particularly in older patients, are intimatelyassociated with the floor of the third ventricle, the hy-pothalamus, and the optic chiasm. In such cases, at-tempts at total removal can produce significant neu-rologic damage; thus the surgeon must use goodjudgment in attempting complete tumor removal. Of-ten, it is better to remove the bulk of the tumor andto treat the small remnants adherent to vital struc-tures with postoperative irradiation.

Radiotherapy

Conventional RT has been effective for craniopharyn-giomas (Hetelekidis et al., 1993). For the reasonsstated previously, however, conventional RT is notrecommended for the immature brain (generally,children 3 years of age or younger). Stereotactic tech-niques include radiosurgery, stereotactic RT, and di-rect colloid instillation into cystic craniopharyn-giomas. Radiosurgery has been utilized for adjunctivemanagement of craniopharyngiomas. However, be-cause the chiasm is often in close proximity, the sameconstraints as discussed earlier apply for cranio-pharyngiomas (Tarbell et al., 1994).

Radiosurgery should only be considered whenthere is a very small area (less than 2 cm) of resid-

ual/recurrent tumor that is away from the optic chi-asm. Direct instillation of colloidal radioisotopes intothe cysts of primarily cystic tumors appears effectivewhen appropriately applied. This technique has beenwidely used in Europe with limited experience in theUnited States. Stereotactic radiation or conformal ra-diation treatments using conventional fractionationmay be the safest mode of treatment for patients witha solid component of residual disease.

REFERENCES

Alvord EC Jr, Lofton S. 1988. Gliomas of the optic nerve orchiasm: outcome by patients age, tumor site, and treat-ment. J Neurosurg 68:8598.

Ater JL, Moore BD, Slopis J, Copeland D. 1999. Neuropsy-chological effects of focal cranial radiation therapy on chil-dren treated for brain tumors. ASCO Proc 18:149A.

Ater JL, van Eys J, Woo SY, et al. 1997. MOPP chemotherapywithout irradiation as primary postsurgical therapy forbrain tumors in infants and young children. J Neurooncol32:243252.

Ater JL, Woo SY, van Eys J. 1988. Update on MOPP chemo-therapy as primary therapy for infant brain tumors. PediatrNeurosci 14:153.

Brzowski AE, Bazan C 3d, Mumma JV, Ryan SG. 1992. Spon-taneous regression of optic glioma in a patient with neu-rofibromatosis. Neurology 42:679681.

Dunbar SF, Tarbell NJ, Kooy HM, et al. 1994. Stereotactic ra-diotherapy for pediatric and adult brain tumors: prelimi-nary report. Int J Radiat Oncol Biol Phys 30:531539.

Dutton JJ. 1991. Optic nerve gliomas and meningiomas. Neu-rol Clin 9:163177.

Edwards MS, Levin VA, Wilson CB. 1980. Brain tumor che-motherapy: an evaluation of agents in current use for phaseII and III trials. Cancer Treat Rep 64:11791205.

Ellenberg L, McComb JG, Siegel SE, Stowe S. 1987. Factors af-fecting intellectual outcome in pediatric brain tumor pa-tients. Neurosurgery 21:638644.

Epstein MA, Packer RJ, Rorke LB, et al. 1992. Vascular mal-formation with radiation vasculopathy after treatment of chi-asmatic/hypothalamic glioma. Cancer 70:887893.

Flickinger JC. 1989. An integrated logistic formula for predic-tion of complications from radiosurgery. Int J Radiat On-col Biol Phys 17:879885.

Friedman HS, Krischer JP, Burger P, et al. 1992. Treatment ofchildren with progressive of recurrent brain tumors withcarboplatin or iproplatin: a Pediatric Oncology Group Ran-domized phase II study. J Clin Oncol 10:249256.

Glauser TA, Packer RJ. 1991. Cognitive deficits in long-termsurvivors of childhood brain tumors. Childs Nerv Syst7:212.

Goldsmith BJ, Wara WM, Wilson CB, Larson DA. 1994. Post-operative irradiation for subtotally resected meningiomas:a retrospective analysis of 140 patients treated from 1987to 1990. J Neurosurg 80:195201.

Gutmann DH, Aylsworth A, Carey JC, et al. 1997. The diag-


3601_e05_p158-170 2/15/02 4:31 PM Page 169

nostic evaluation and multidisciplinary management of neu-rofibromatosis 1 and neurofibromatosis 2. JAMA 278:5157.

Hakim R, Alexander E 3rd, Loeffler JS, et al. 1998. Results oflinear accelerator-based radiosurgery for intracranialmeningiomas. Neurosurgery 42:446453.

Hetelekidis S, Barnes PD, Tao ML, et al. 1993. 20-year expe-rience in childhood craniopharyngioma. Int J Radiat OncolBiol Phys 27:189195.

Horwich A, Bloom HJ. 1985. Optic gliomas: radiation therapyand prognosis. Int J Radiat Oncol Biol Phys. 11:10671079.

Hoyt WF, Baghdassarian SA. 1969. Optic glioma of childhood.Natural history and rationale for conservative management.Br J Ophthalmol 53:793798.

Imes RK, Hoyt WF. 1986. Childhood chiasmal gliomas: updateon the fate of patients in the 1969 San Francisco study. BrJ Ophthalmol 70:179182.

Jannoun L, Bloom HJ. 1990. Long-term psychological effectsin children treated for intracranial tumors. Int J Radiat On-col Biol Phys 18:747753.

Janss AJ, Grundy R, Cnaan A, et al. 1995. Optic pathway andhypothalamic/chiasmatic gliomas in children younger than5 years with a 6-year follow-up. Cancer 75:10511059.

Kanamori M, Shibuya M, Yoshida J, Takayasu M, Kageyama N.1985. Long-term follow-up of patients with optic glioma.Childs Nerv Syst 1:272278.

Kjellberg RN, Hanamura T, Davis KR, Lyons SL, Adams RD.1983. Bragg-peak proton-beam therapy for arteriovenousmalformations of the brain. N Engl J Med 309:269274.

Listernick R , Charrow J, Greenwald MJ, Esterly NB. 1989. Op-tic gliomas in children with neurofibromatosis type 1. J Pe-diatr 114:788792.

Loeffler JS, Alexander E III. 1993. Radiosurgery for the treat-ment of intracranial metastases. In: Alexander E III, Loef-fler JS, Lunsford LD (eds), Stereotactic Radiosurgery. NewYork: McGraw-Hill, pp 197206.

Loeffler JS, Kooy HM, Terbell NJ. 1999. The emergence of con-formal radiotherapy: special implications for pediatricneuro-oncology. Int J Radiat Oncol Biol Phys 44:237238.

Manera RB, Ater JL, Leeds N, et al. 1994. Treatment outcomeand neurological, neuroendocrine and neurobehavioralprofile of children with supratentorial midline brain tu-mors. Pediatr Neurosurg 21:265.

Marks LB, Spencer DP. 1991. The influence of volume on thetolerance of the brain to radiosurgery. J Neurosurg75:177180.

Mitchell WG, Fishman LS, Miller JH, et al. 1991. Stroke as alate sequela of cranial irradiation for childhood brain tu-mors. J Child Neurol 6:128133.

Moore BD 3rd, Ater JL, Copeland DR. 1992. Improved neu-ropsychological outcome in children with brain tumors di-agnosed during infancy and treated without cranial irradi-ation. J Child Neurol 7:81290.

Oakes WJ. 1990. Recent experience with the resection of pi-locytic astrocytomas of the hypothalamus. In: AE Marlin(ed), Concepts in Pediatric Neurosurgery, vol 10. Basel:Karger, pp 108117.

Packer RJ, Ater J, Allen J, et al. 1997. Carboplatin and vin-cristine chemotherapy for children with newly diagnosedprogressive low-grade gliomas. J Neurosurg 86:747754.

Packer RJ, Lange B, Ater J, et al. 1993. Carboplatin and vin-cristine for recurrent and newly diagnosed low-gradegliomas of childhood. J Clin Oncol 11:850856.

Packer RJ, Savino PJ, Bilaniuk LT, et al. 1983. Chiasmaticgliomas of childhood. A reappraisal of natural history andeffectiveness of cranial irradiation. Childs Brain 10:393403.

Petronio J, Edwards MS, Prados M, et al. 1991. Managementof chiasmal and hypothalamic gliomas of infancy and child-hood with chemotherapy. J Neurosurg 74:701708.

Pierce SM, Barnes PD, Loeffler JS, McGinn C, Tarbell NJ. 1990.Definitive radiation therapy in the management of sympto-matic patients with optic glioma. Cancer 65:4552.

Pollock IF. 1994. Brain tumors in children. N Engl J Med331:15001507.

Poussaint TY, Siffert J, Barnes PD, et al. 1995. Hemorraghicvasculopathy after the treatment of central nervous systemneoplasia in childhood: diagnosis and follow-up. Am J Neu-roradiol 16:693699.

Prados MD, Edwards MS, Rabbitt J, Lamborn K, Davis RL, LevinVA. 1997. Treatment of pediatric low-grade gliomas with a nitrosourea-based multiagent chemotherapy regimen. J Neurooncol 32:235241.

Rajakulasingam K, Cerullo LJ, Raimondi AJ. 1979. Childhoodmoyamoya syndrome. Postradiation pathogenesis. ChildsBrain 5:467475.

Rodriguez LA, Edwards MS, Levin VA. 1990. Management ofhypothalamic gliomas in children: an analysis of 33 cases.Neurosurgery 26:242246.

Russell A. 1951. A diencephalic syndrome of emaciation in in-fancy and childhood. Arch Dis Child 26:274275.

Stern J, DiGiacinto GV, Housespian EM. 1979. Neurofibro-matosis and optic glioma: clinical and morphological cor-relations. Neurosurgery 4:524528.

Sutton LN, Molloy PT, Sernyak H, et al. 1995. Long-term out-come of hypothalamic/chiasmatic astrocytomas in childrentreated with conservative surgery. J Neurosurg 83:583589.

Tao ML, Barnes PD, Billett AL, et al. 1997. Childhood opticchiasm gliomas: radiographic response following radio-therapy and long-term clinical outcome. Int J Radiat OncolBiol Phys 39:579587.

Tarbell NJ, Barnes P, Scott RM, et al. 1994. Advances in radi-ation therapy for craniopharyngiomas. Pediatr Neurosurg21(suppl 1):101107.

Tenny RT, Laws ER Jr, Younge BR, Rush JA. 1982. The neu-rosurgical management of optic glioma. Results in 104 pa-tients. J Neurosurg 57:452458.

Tishler RB, Loeffler JS, Lunsford LD, et al. 1993. Tolerance ofcranial nerves of the cavernous sinus to radiosurgery. Int JRadiat Oncol Biol Phys 27:215221.

Weiss L, Sagerman RH, King GA, Chung CT, Dubowy RL. 1987.Controversy in the management of optic nerve glioma. Can-cer 59:10001004.

Wisoff JH, Abbott R, Epstein F. 1990. Surgical management ofexophytic chiasmatic-hypothalamic tumors of childhood. J Neurosurg 73:661667.

Wong JY, Uhl V, Wara WM, Sheline GE. 1987. Optic gliomas.A reanalysis of the University of California, San Franciscoexperience. Cancer, 60:18471855.


3601_e05_p158-170 2/15/02 4:31 PM Page 170

optic nerve chasmal hypothalamic tumors.pdf

Documents