Principles & Practice of Pediatric Oncology
5th Edition

Tumors of the Central Nervous System
Susan M. Blaney
Larry E. Kun
Jill Hunter
Lucy B. Rorke-Adams
Ching Lau
Douglas Strother
Ian F. Pollack
Tumors of the central nervous system (CNS) constitute the second most common pediatric cancer diagnosed in the United States each year. Depending on the upper age chosen, the number of children, adolescents, and young adults who received diagnoses of a CNS tumor in 1999 ranged between 1,700 (for ages 0 to 14 years) and 2,200 (for ages 0 to 20 years).1 If one includes such histopathologically “benign” diagnoses as craniopharyngioma and choroid plexus papilloma (CPP) the number is even higher.2,3 Figure 27.1 shows the approximate incidence of the common pediatric CNS tumors.
Unfortunately the morbidity with CNS tumors and their therapy may be significant in terms of physical deficits as well as neuropsychological and neuroendocrine sequelae. Although not quantifiable, the long-term morbidity of childhood CNS tumors likely exceeds that associated with other pediatric malignancies. Deaths caused by CNS tumors are the highest among pediatric cancers.4 Nevertheless, there has been a 15% survival rate increase for children with CNS tumors between 1975–1979 and 1995–2000 such that the overall survival is now approximately 70%.5 Despite these modest increases in survival, the optimal treatment of childhood CNS tumors continues to pose a tremendous challenge that requires a multidisciplinary approach involving many pediatric specialists and subspecialists including neurosurgeons, neuropathologists, neurooncologists, neuroradiologists, radiation oncologists, neurologists, ophthalmologists, and physiatrists. In addition, the contributions of molecular biologists, pharmacologists, nurses, neuropsychologists, social workers, audiologists, nutritional experts, child-life specialists, and physical, occupational, and speech therapists are invaluable.
In this chapter, we review the current understanding of the biology of brain tumors and the principles associated with each of the diagnostic and treatment modalities. We then provide an overview of the clinical management and associated long-term sequelae of the more frequently encountered CNS tumors.
In the early 1990s, there appeared to be an increase in the incidence of CNS tumors from 2.7 cases per 100,000 children during the years 1977 to 1981 to 3.3 cases per 100,000 children from 1990–1994.6 This higher incidence was primarily attributed to the greater utilization of magnetic resonance imaging (MRI) for evaluating children with neurologic conditions, although this has not been definitively proven.7 Another contributing factor to the apparent increased incidence of pediatric brain tumors was the increasingly widespread use of stereotactic biopsies to document tumor histologies at nonbrainstem sites in tumors that previously would not have been subject to biopsy. In addition, during this time the World Health Organization (WHO) classification of malignant gliomas changed, resulting in a shift of some diagnoses from benign to malignant. These combined factors affected the detection and reporting of brain tumors.7 Ongoing observation is required to determine whether there was truly a rise in the incidence of childhood CNS tumors or whether the changes in the detection and reporting of brain tumors resulted in the observed increase.
The incidence of brain tumors peaks in the first decade of life, then decreases until a second peak in older adulthood. This second peak historically occurred in the seventh decade of life, but recent Surveillance, Epidemiology, and End Results (SEER) data demonstrated a shift in this peak to the eighth decade.8 The first peak is characterized by a predominance of males and by equal incidence rates for whites and blacks, except for the first 2 to 3 years of life, when a greater percentage of whites than nonwhites are affected.4 The male predominance is primarily explained by a disproportionate incidence of both medulloblastoma (MB) and ependymoma in males. For other tumor types, the genders are equally affected. During the first 2 years of life, supratentorial tumors predominate, whereas infratentorial lesions are more common through the rest of the first decade. Supratentorial tumors again predominate during late adolescence and through adulthood. Tumors

of embryonal histology such as MB, supratentorial primitive neuroectodermal tumors (sPNETs), and pineoblastomas occur almost exclusively in children and young adults and primarily occur during the first decade. High-grade gliomas, including glioblastoma multiforme, are much less common in children than in adults.
Figure 27.1 Approximate incidence of common central nervous system tumors in children.
Syndrome Gene(s) CNS Tumor Type(s) Non-CNS Tumors Reference
Cowden PTEN Dysplastic gangliocytoma of the cerebellum (Lhermitte-Duclos)   33
Hereditary retinoblastoma Rb Pineoblastoma, glioma, meningioma Retinoblastoma, osteosarcoma, malignant melanoma 20
Li-Fraumeni TP53 Multiple brain tumor types, most commonly supratentorial PNET, medulloblastomas and astrocytoma Sarcoma, adrenocortical tumor, acute leukemia, premenopausal breast cancer 23, 24, 940
Neurofibromatosis Type 1 NF-1 Neurofibroma, optic nerve glioma, astrocytoma   25, 26
Neurofibromatosis Type 2 NF-2 Acoustic and peripheral schwannoma, meningioma, spinal ependymoma   29, 30
Nevoid basal cell carcinoma (Gorlin syndrome) PTCH Medulloblastoma, meningiomas Basal cell carcinomas 17, 18, 20
Rubenstein-Taybi CBP Medulloblastoma, oligodendroglioma, meningioma   20
Tuberous sclerosis TSC1, TSC2 Subependymal giant-cell astrocytoma   31, 32
Turcot APC Medulloblastoma (most common) Colorectal adenomas 22, 941
  hMLH1, hPMS2 Astrocytoma and ependymoma (less common) Colorectal adenocarcinoma  
von Hippel-Lindau VHL Hemangioblastoma   34
CNS, central nervous system; PNET, primitive neuroectodermal tumor.
Only two factors are consistently noted to place a child at increased risk for a CNS malignancy: various genetic disorders (Table 27.1) and exposure to ionizing radiation.9,10
Fewer than 10% of children with brain tumors have a genetic disorder that places them at increased risk for developing a brain tumor. Although rare, these syndromes (Table 27.1) place affected children at a markedly higher risk for developing other tumors as well.11 All of the currently known syndromes associated with a predisposition for developing brain tumors have an autosomal dominant pattern of inheritance, and somatic mutations have been demonstrated in specific genes for each (Table 27.1).

Fewer than 5% of children with MBs have inherited disorders.12 The most common of these are the nevoid basal cell carcinoma (Gorlin) syndrome, Turcot syndrome, and the Li Fraumeni syndrome.13,14,15
Gorlin syndrome has been linked to germline mutations of the Sonic hedgehog receptor PTCH.16,17,18 Affected children are born with multiple skeletal anomalies and macrocephaly. They have a 3% to 5% incidence of MB,13,19 primarily of the desmoplastic subtype,19 which occurs at an earlier age than in children without Gorlin syndrome. Thus, in any child younger than 2 years who presents with a desmoplastic MB an underlying diagnosis of Gorlin syndrome should be strongly considered, particularly because these children are predisposed to basal cell carcinomas, a risk that is substantially increased in the fields of radiation used to treat MB.
Children with Turcot syndrome have an autosomal dominant disorder in which a patient with a primary brain tumor also develops colorectal adenomas and/or colorectal adenocarcinoma.20 Two subgroups of Turcot syndrome appear to exist. In the first, which is due to mutation of the adenomatosis polyposis coli (APC) gene, patients have an increased risk of MB, extensive colorectal adenomas, and extracolonic manifestations such as osteomas, desmoid tumors, jaw cysts, and supernumerary teeth.20,21 Patients in the second subgroup, which is associated with mutations in DNA mismatch repair genes (hPMS2 and hMLH1), develop gliomas and colorectal adenocarcinoma.20,22
Children with Li-Fraumeni syndrome, which is caused by germline mutations in the TP53 gene, may develop multiple cancer types.23,24 The protein encoded by this gene is multifunctional, having a role in cell cycle control, in ensuring DNA integrity and repair and, in some circumstances, in inducing apoptotic cell death. Children with inherited mutations of the TP53 gene most commonly develop low- or high-grade gliomas that may be multifocal. They may also develop MBs, primitive neuroectodermal tumors (PNETs), or choroid plexus tumors. They also develop tumors outside the CNS with an increased incidence including sarcomas, leukemias, and adrenocortical carcinomas.
Children with neurofibromatosis type 1 (NF-1), due to mutation of the NF-1 gene, are at risk for developing dermal and plexiform neurofibromas and have a markedly increased risk for astrocytomas.25,26,27,28 The astrocytomas are typically low-grade neoplasms that occur within the optic pathway involving both optic nerves, the chiasm, and the optic radiations. Low-grade gliomas (LGAs) may also occur within the cerebral hemispheres, the brainstem, or the cerebellum. Gliomas and plexiform neurofibromas may undergo malignant transformation. Other cancers that occur in association with NF-1 include myelogenous leukemia, rhabdomyosarcoma, and pheochromocytoma.
Neurofibromatosis type 2, resulting from mutations in the NF-2 gene, is associated with meningiomas within the brain and spine and schwannomas of the cranial nerves, spinal nerves, and peripheral nervous system.29,30 Bilateral acoustic nerve schwannomas are highly associated with neurofibromatosis type 2. Gliomas and ependymomas also occur with increased frequency and tend to be located in the spinal cord.
Finally, several rare tumor types occur most frequently in association with specific inherited disorders. Subependymal giant-cell astrocytomas, which occur in the anteromedial aspect of the brain near the foramina of Monro, most often occur in children with tuberous sclerosis.31,32 Cerebellar gangliocytoma (Lhermitte-Duclos) occurs in the context of Cowden syndrome, owing to mutation of the PTEN gene that encodes a dual specificity phosphatase.33 Hemangioblastomas, typically in the cerebellum, spinal cord, or retinas, occur in association with von Hippel-Lindau syndrome, which arises from mutation of the VHL gene that appears to have a role in DNA replication.34,35,36
Ionizing Radiation
Exposure to ionizing radiation is a well-documented cause of brain tumors. Children treated with radiation therapy for tinea capitis during the 1940s and 1950s were found to have increased risk for the development of gliomas, meningiomas, and nerve sheath tumors 22 to 34 years later.37 More recently, brain tumors after cranial irradiation for acute lymphoblastic leukemia have been reported.38,39,40,41
Other Cancers
Brain tumors may be seen in association with other cancers or as a result of their treatment. Patients with retinoblastoma may also develop a pineoblastoma, the so-called trilateral retinoblastoma syndrome.42 Brain tumors may also be seen in a minority of patients with malignant rhabdoid tumors of the kidney.43 Pituitary tumors occur in patients with various forms of the multiple endocrine adenomatosis syndrome (see Chapter 37). Finally, brain tumors, particularly high-grade astrocytomas or meningiomas, may occur in patients who previously received CNS radiation therapy.
CNS lymphomas occur with increased frequency in patients with a variety of underlying primary or secondary disorders of the immune system, including the Wiskott-Aldrich syndrome, ataxia-telangiectasia, acquired immunodeficiency syndrome, and after solid-organ transplantation.44,45,46
Familial Conditions
Data are inconclusive regarding less completely understood familial factors outside of known Li-Fraumeni families. Some studies show no influence of family history on the occurrence of brain tumors, whereas others report an increased risk of brain tumors with a family history of bone cancer, leukemia, and lymphoma. The children or siblings of persons with brain tumors may be at higher risk for developing brain tumors themselves.47,48,49,50,51 Reports of familial clustering of embryonal tumors, gliomas, and CPPs also exist.52,53,54,55,56,57
Environmental Exposures
The effect of environmental exposures, including diet, on the occurrence of childhood brain tumors has been studied by

numerous investigators without conclusive evidence for an association.9,58,59 Studies of the effect of cellular telephones on the occurrence of childhood brain tumors have not been performed. Although studies with regard to mobile phone use and brain tumor development have been performed in adults, they are associated with a variety of limitations including small subject number, lack of appropriate controls, and short follow-up duration. The results of these studies are variable; some suggest there is no conclusive evidence for association, some are inconclusive, and others suggest a slight to significant increase risk of CNS tumor development.60,61,62,63,64 An association of polyomavirus (e.g., JC and SV-40 viruses) with certain types of pediatric brain tumors such as MB, ependymoma, and choroid plexus tumors has been reported;65 however, there is no definitive evidence that such viruses are directly involved in the pathogenesis of these tumors.66,67,68
Several factors confound the epidemiologic study of pediatric brain tumors. First, until recently, etiologic studies considered pediatric cancer a single entity, and brain tumors were not examined separately. Second, the etiology of brain tumors is most likely multifactorial, and these factors may influence distinct histologic types of tumors to variable degrees. Finally, pediatric brain tumors are rare, and this rarity affects research methodology. Nearly all studies of pediatric brain tumors are case-control studies in which individuals with and without brain tumors are compared with respect to past exposures. Inaccuracies and disparities in patient or parent recall may limit observations of disease and their associations.9 Should a link with environmental exposures truly exist, it may be difficult to establish.
Cancers arise as a result of mutations in genes that regulate cell proliferation and death. Gene mutations may originate within the germline or may occur as somatic mutations exclusively within tumor cells. As previously noted, only a small fraction of children with brain tumors have germline mutations either acquired from their parents (giving them an inherited predisposition to cancer) or as new mutations. Although the causes of the somatic mutations underlying the vast majority of all brain tumors are unknown, the ongoing prospective characterization of genetic abnormalities typically associated with childhood brain tumors is important as it may have implications for tailoring treatment.
Fluorescence-activated cell sorting and direct chromosomal preparations from biopsy specimens have demonstrated that LGAs, meningiomas, and pituitary adenomas almost universally possess a unimodal diploid DNA content.69 In contrast, direct biopsy preparations from more aggressive and malignant tumors, such as anaplastic astrocytoma and glioblastoma multiforme, frequently show bizarre chromosomal aberrations and a dominance of triploid and tetraploid cells in the cell lines derived from these tumors. These findings are in agreement with the genetic analyses of cancers arising in other sites, which generally have demonstrated increasingly abnormal genetic content as cancers evolve into high-grade malignancies. Conversely, fluorescence-activated cell sorting and karyotyping performed on freshly prepared tumor cell populations show near-diploid genetic content, possibly as a result of tumor contamination with normal diploid stromal cells. Alternatively, small diploid or near-diploid anaplastic cells may harbor specific mutations that render them both drug- and radiation-resistant, such that their malignant characteristics are not caused by overt loss or gain of chromatin.69,70,71,72,73,74,75,76
Chromosomal anomalies have been defined further through molecular genetic analyses, including studies of loss of heterozygosity, comparative genomic hybridization (CGH), spectral karyotyping, and fluorescence in situ hybridization (FISH). The most consistent genetic abnormalities have been found in MBs and atypical teratoid-rhabdoid tumors (AT/RTs). Deletions of the short arm of chromosome 17 (17p) distal to the TP53 locus occur in 30% to 50% of MBs.77,78 Loss of 17p frequently occurs in association with duplication of 17q, which is characteristic of isochromosome 17q, the most common cytogenetic abnormality of MBs.77,79,80 Trisomy 7 is the second most common cytogenetic abnormality.77,80 Deletions or mutations of chromosome 9q involving the Sonic hedgehog receptor PTCH have been found in association with approximately 10% to 15% of sporadic MBs.81,82,83,84 A variety of other chromosomal deletions or additions have been found less consistently.85,86
The nonrandom occurrence of monosomy 22 is common in several neural tumors, including MBs, ependymomas, meningiomas, acoustic neuromas, and AT/RTs.74,87 Mutations or deletions of chromosome 22q11.2 involving the hSNF5/INI1 gene are present in nearly all AT/RTs and in rhabdoid tumors outside the CNS.86,88 The function of the protein encoded by this gene is incompletely understood. It forms part of the SWI/SNF complex that apparently alters chromatin structure to allow gene transcription.89 In addition to monosomy 22, deletions of 6q in ependymomas have been described by several investigators.90,91
A variety of genetic events have been associated with the progression of LGAs to high-grade glioblastomas in adults: nonrandom loss of chromosome 10 or of portions of 9p, 17p, and others; gene amplifications of EGFR or MDM2; and, mutations of genes such as PTEN and TP53.69,92,93,94 These observations suggest that, as in colon carcinoma, malignant progression of CNS tumors may be a multistep process involving the accumulation of genetic abnormalities that promote activation of multiple dominant oncogenes and inactivation of recessive tumor suppressor genes.78,95,96,97,98,99,100,101 However, there are apparent differences between pediatric and adult malignant gliomas in their patterns of genetic alterations, specifically a much lower incidence of epidermal growth factor receptor (EGFR) amplification in glioblastoma,102 a paucity of TP53 mutations in malignant gliomas arising in infants,103 and an association between TP53 mutations and outcome in older children.104 These observations call attention to the need for studies specifically targeted to childhood gliomas to address issues of genetic progression and age-specific molecular contributors to tumor growth.

Classification of CNS tumors has been a challenge for more than 100 years. Almost from the beginning, several different histologic classifications have coexisted, reflecting a lack of consensus among neuropathologists. Historical details of this problem have been reviewed in detail by Zulch.105 All classification schemes are basically arbitrary, hypothetical statements that may or may not reflect the natural order of the things being classified. Ideally, however, they change with progress in knowledge, are clinically relevant, and provide a structure within which a subject may be arranged. These goals may not necessarily be achieved by the same scheme.
Classification of CNS tumors became important clinically after successful establishment of neurosurgery as a specialty and received major impetus from Harvey Cushing during the 1920s and 1930s. The classification system proposed by Cushing’s student, Percival Bailey, in 1926 has served as the prototype system for CNS tumors. That scheme, which was based on the cell-of-origin notion introduced by German pathologists, suggested that tumors developed from cells arrested at various stages of development; for each putative developmental stage of a cell, a corresponding tumor was identified. Figure 27.2 outlines Bailey and Cushing’s basic schema, denoting the different cell types, the developmental stages through which they were said to pass before reaching maturity, and the corresponding tumor types presumed to arise from them (noted in parentheses in the figure). This classification was embraced immediately because it reflected some aspects of clinical behavior and prognosis, corresponding to Cushing’s experience, and because of the esteem that Cushing enjoyed.
Figure 27.2 Bailey and Cushing schema of normal developing cells and neuroepithelial tumors derived from them.
Methods of Classification
Morphologic and Histogenetic Classification
In addition to their cell-of-origin concept, Bailey and Cushing recognized that tumors were composed of heterogeneous cells. They decided to classify tumors on the basis of the morphology and presumed histogenesis of the predominant cell type. Hence, if the majority of cells resembled astrocytes, the tumor was called an astrocytoma, even though a small number of other cells (e.g., oligodendrocytes) also were present. Most classifications in current use are based on this concept.
Approximately 60 years ago, Kernohan et al.106 introduced a grading system based on the concept that glial cells of all types become progressively more anaplastic over time. Criteria were advanced for grading glial tumors on a scale from 1 (most benign) to 4 (most malignant). The scheme was to be applied to astrocytomas, oligodendrogliomas, and ependymomas. Based on histologic features, it has been readily adopted and is used commonly for astrocytomas for which clinical correlations exist. For oligodendrogliomas and ependymomas, the scheme has been an awkward fit and, as a result, never has been used consistently for them. Revision of the Kernohan grading system for astrocytomas has been proposed by Daumas-Duport et al.107 An international panel of neuropathologists, working under the aegis of the WHO, has expanded the concept of grading all CNS tumors, using the 1-to-4 scale to indicate biologic malignancy.108
Although it is generally accepted that CNS tumors in children differ in many respects from those in adults, there had been no separate classification system for pediatric tumors until an adaptation of the 1979 WHO scheme was published in 1985.109 This was later modified and is presented in Table 27.2. This scheme recognized not only morphologic

entities of neoplasia, but tumor location as well. Site designations are listed Table 27.3. Consideration of location acknowledges the importance of site of origin as a factor in determining clinical outcome. An obvious example is the pilocytic astrocytoma; a child with a lesion in the cerebellum generally has a better prognosis than one whose lesion is in the diencephalon.
  1. Tumors of neuroepithelial tissue
    1. Glial tumors
      1. Astrocytic tumors
        1. Astrocytoma (fibrillary, protoplasmic, gemistocytic, pilocytic, gigantocellular)
        2. Anaplastic astrocytoma
      2. Oligodendroglial tumors
        1. Oligodendroglioma
        2. Anaplastic oligodendroglioma
      3. Ependymal tumors
        1. Ependymoma
          1. Myxopapillary
        2. Anaplastic ependymoma
      4. Choroid plexus tumors
        1. Choroid plexus papilloma
        2. Choroid plexus adenoma
        3. Choroid plexus carcinoma
      5. Mixed gliomas
        1. Oligoastrocytoma
          1. Anaplastic oligoastrocytoma
        2. Ependymoastrocytoma
          1. Anaplastic ependymoastrocytoma
        3. Oligoastroependymoma
          1. Anaplastic oligoastroependymoma
        4. Gliofibroma
      6. Glioblastomatous tumors
        1. Glioblastoma multiforme
        2. Giant-cell glioblastoma
        3. Gliosarcoma
      7. Gliomatosis cerebri
    2. Mixed glial-neuronal tumors
      1. Ganglioglioma
        1. Dysembryoplastic neuroepithelial tumor
      2. Superficial cerebral astrocytoma–desmoplastic infantile ganglioglioma
      3. Pleomorphic xanthoastrocytoma
      4. Subependymal giant-cell tumor (of tuberous sclerosis)
      5. Anaplastic ganglioglioma
        1. Anaplastic superficial cerebral astrocytoma–desmoplastic infantile ganglioglioma
        2. Anaplastic pleomorphic xanthoastrocytoma
    3. C.Neural tumors
      1. Gangliocytoma
        1. Anaplastic gangliocytoma
      2. Neurocytoma
        1. Anaplastic neurocytoma
    4. D.Embryonal tumors
      1. Primitive neuroectodermal tumor (PNET)
        1. PNET, not otherwise specified
        2. PNET with glial differentiation
        3. PNET with ependymal differentiation
        4. PNET with neuronal differentiation
        5. PNET with retinal differentiation
        6. PNET with mesenchymal differentiation
        7. PNET with melanocytic differentiation
        8. PNET with differentiation along multiple lines
      2. Medulloepithelioma
        1. Medulloepithelioma with differentiation along divergent lines as above (1.b–h)
      3. Atypical teratoid-rhabdoid tumor
    5. E.Pineal cell tumors
      1. Pineocytoma
From Rorke LB, Gilles FH, Davis RL, et al. Revision of the World Health Organization classification of brain tumors for childhood brain tumors. Cancer 1985;56:1869–1886, with permission.
Site Designation
Central nervous system parenchyma I
Hemispheres a
Thalamus and/or basal ganglia b
Hypothalamus/chiasm c
Anterior optic pathways d
Intraventricular e
Pineal region f
Brainstem g
Cerebellum h
Brainstem and cerebellum (including cerebellopontine angle and middle cerebellar peduncle) i
Spinal cord (level and site, if known) j
Intradural, extramedullary jl
Meninges II
   Intradural a
   Extradural b
Parasellar region III
Skull and/or vertebral column IV
Orbit (eye) V
Peripheral nervous system VI
Pitfalls of Morphologic and Histogenetic Classification
It is important to recognize that the histogenetic concepts underlying the morphologic classification schemes are not fully tenable in light of current knowledge.110,111 Cancer is a genetic disease in which the genotypic instability of neoplastic cells may change the histologic features consequent to both time and treatment, although a remarkably large proportion of tumors exhibit relatively characteristic features, allowing easy diagnosis to the trained eye. It is no longer appropriate to seriously consider the cell-of-origin hypothesis advanced by Bailey and Cushing, because it has been established that cell populations displaying similar, or even identical patterns of differentiation may not have a common embryogenesis. Hence it is not possible to determine either ancestry or progeny of a tumor cell or cells.112 In addition, such theorizing is largely irrelevant because the behavior of the tumor is dependent on the intrinsic nature of the component cells and other factors.

Phenotypic Classification
An alternative to classification based on histogenetic concepts is the phenotypic approach. Essentially, this involves evaluation of the tumor by identification of cell types comprising it.113,114,115 Beyond examination of sections stained by routine hematoxylin and eosin (H&E), other special stains, immunohistochemistry, ultrastructural study, and cytogenetics can now be used to determine the cell types comprising the tumor with greater precision than was possible a decade ago. Use of monoclonal antibodies to identify specific antigens, such as cytoskeletal and membrane proteins, hormonal polypeptides, and neurotransmitter substances, has been especially useful in classifying, on routine light microscopy, tumors with unusual morphologic features that previously were relegated to the “unknown” category. In particular, it has allowed greater understanding of the biology of embryonal tumors, as the differentiating potential of the primitive cells can be identified.116,117
Use of this phenotypic approach, coupled with increasing information of the cytogenetics and microarray analyses of CNS tumors, has forced changes in the historic classification schemes.
Application of the newer techniques, for example, led to identification and separation of AT/RTs from the larger group of embryonal tumors, that is, PNETs or MB, into which diagnostic niche they had formerly been placed. This distinction carries significant clinical implications because the prognosis of AT/RT is quite generally quite different from that for either PNET or meulloblastoma.118
In addition, use of immunoperoxidase techniques with various antibodies has also forced reconsideration of the nosology of certain tumors long regarded as astrocytomas, specifically subependymal giant-cell astrocytoma, pleomorphic xanthoastrocytoma, and superficial cerebral astrocytoma. In 1980, Bender and Yunis119 called attention to the presence in subependymal giant-cell astrocytomas of cells that expressed neurofilament protein, an observation confirmed by Nakamura and Becker120 and by Bonnin et al.121 In fact, the majority of cells composing these tumors expresses vimentin, whereas only a few express glial fibrillary acidic protein (GFAP), and some tumor cells actually coexpress both neurofilament protein (NFP) and GFAP.120 The expression of either of the neuronal markers would rarely, if ever, be seen in pure astrocytomas.
Powell et al.122 documented expression of neurofilament protein in a variable number of cells in pleomorphic xanthoastrocytoma and noted that some of these tumors also contain foci that resemble frank ganglioglioma.122,123
It has become apparent that the tumor described by Taratuto et al.124 in 1984 as a dural astrocytoma and the desmoplastic infantile ganglioglioma (DIG) defined in 1987 by VandenBerg et al.125 really belong in the same category, because both express GFAP and NFP and exhibit a prominent desmoplastic component.126 These observations form the basis of Rorke’s removal of these three tumors from the astrocytoma category and their placement within the glial-neuronal neoplasms (Table 27.2).
Yet another example of a change in diagnostic category brought about by application of a nonroutine diagnostic tool, namely the electron microscope, is the central neurocytoma. This tumor commonly arises in the region of the lateral and third ventricles. Until Hassoun et al.127 studied the ultrastructural features of two such tumors, they were diagnosed either as ependymomas or oligodendrogliomas. However, it is now clear that the tumors were neuronal not glial neoplasms and their place in the classification scheme was changed accordingly.
The foregoing brief review of tumor nomenclature based on morphology, with or without cytogenetic information, emphasizes the reality of a continuously changing classification system that reflects the current state of knowledge and the folly of clinging to historical schemes that were accurate in their time but which are no longer tenable.
Unsolved Issues in Classification of Embryonal Tumors of Neuroepithelial Origin
From the time that the prototypic CNS embryonal tumor—MB—was described by Bailey and Cushing, controversy has swirled around its nature, origin, and name.128 The essential issue at the heart of the controversy is whether the MB is a tumor unique to the cerebellum. If tumors of their histology and biologic behavior occurred only in the cerebellum, no problem would exist. In reality, however, histologically and biologically identical tumors may arise in the cerebrum, the pineal and suprasellar regions, the spinal cord, and the brainstem. Such tumors were well known to Bailey and Cushing who, however, remained ambivalent in regard to what name should be given to the extracerebellar tumors resembling MBs.129 As a consequence, a large number of diagnostic terms have been applied to MB-like tumors that arise outside the cerebellum.
A proposal to resolve the problem in 1983 seemed, instead, to stimulate further controversy.130 Specifically, it was proposed that tumors composed primarily of apparently undifferentiated neuroepithelial cells be considered a unique diagnostic group regardless of site of origin in the CNS, recognizing that the majority of such tumors do, in fact, arise in the cerebellum. Introduced by Hart and Earle, the diagnostic term PNET was suggested as appropriate for this group, with the addition of modifiers specifying differentiation along one or more lines (e.g., neuronal, glial) if such were identified by use of phenotypic markers (i.e., antibodies for NFP, GFAP, etc.).131 This proposal was criticized as a “simplistic” approach that would foster intellectual languor among pathologists who would use the diagnosis of PNET as a catchall term for all difficult-to-diagnose tumors primarily composed of poorly differentiated neuroepithelial cells.131 In addition, many were reluctant to abandon the diagnostic term medulloblastoma because of its familiarity and association with an enormous body of medical literature.
A major issue relative to nomenclature was the question of whether the MB is a uniquely cerebellar tumor. Although tumors resembling MB are found in other CNS sites, they do not always exhibit the histologic variations that are common to those arising in the cerebellum. For example, it is exceptionally rare to find a classical desmoplastic or nodular variant in a noncerebellar location.
Studies yielding conflicting data relative to the question of whether the MB was a tumor specific to the cerebellum accounted for some of the controversy. More than 60 years ago, it was claimed that the tumor arose from embryonic cells of the granular layer.132,133,134 This assertion was challenged by

later investigators,135,136,137 but a considerable number of recent studies provide evidence that at least some MBs likely result from malignant transformation of external granular cells and, hence by necessity, can only be found in the cerebellum.138,139,140,141 The best candidate in this category is the desmoplastic MB.142,143,144,145 This histologic subtype is almost never seen in PNETs that occur in other CNS sites, although such tumors elsewhere may exhibit other patterns of desmoplasia.
It also appears that the cytogenetics of MBs differ from PNETs elsewhere in the CNS. In particular, an i(17)q abnormality has been identified in 30% to 50% of MBs but, to date, has been found in only one cerebral PNET.144,146,147 In addition, other cytogenetic abnormalities that occur in MBs have not been found in PNETs elsewhere and vice versa.144
Studies of cell signaling systems associated with normal cerebellar development and possible dysregulation as a factor in pathogenesis of MBs also contribute to emerging information regarding the MB.142,144
Given information currently available, it appears that desmoplastic MBs are indeed unique to the cerebellum, but the much larger group of embryonal tumors exhibiting the so-called classical histologic features and the nodular “neuroblastic” type, which occur outside of the cerebellum, are biologically different and possibly related to their site of origin in the CNS.
Figure 27.3 Classification of childhood central nervous system (CNS) tumors by gene expression profiles. This figure shows signal-to-noise rankings of genes comparing each tumor type to all other types combined. For each gene, red would indicate a high level of expression relative to the mean; blue would indicate a low level of expression relative to the mean. MD, medulloblastoma; Mglio, malignant glioma; Rhab, rhabdoid; Ncer, normal cerebella; PNET, supratentorial primitive neuroectodermal tumor; σ, standard deviation from the mean.
Neurodevelopmental studies in fact offer evidence that primitive neuroepithelial cells, although histologically similar, are genetically programmed differently depending on location.148,149,150,151,152,153 Hence, it is conceivable that the behavior of transformed cells in different locations in the CNS would reflect biologic heterogeneity as well. Such a concept is in keeping with observations, already noted previously, that site of origin of histologically similar tumors is of significance in governing biology and prognosis and lends credence to the suggestion by Rorke154 that tumor location should be indicated, along with histologic diagnosis.
Until further insight is gained relative to these complex issues, it seems prudent to base a classification scheme on phenotypic features of a tumor or group of tumors. These characteristics may be determined by utilization of immunohistochemical techniques, ultrastructural evaluation (in some circumstances), cytogenetic study, and available molecular biology techniques (Fig. 27.3).
For example, using DNA microarray technology to screen the expression level of a large number of genes simultaneously, Pomeroy et al.142 were able to distinguish MBs, sPNETs, AT/RTs, and malignant gliomas based on a limited number of differentially expressed genes. Although considerable progress has been made along this path,11 a great deal of work remains to be done.
Marker Tumor Types Containing Positive Cells
Glial fibrillary acidic protein Astrocytoma, ependymoma, mixed glioma, gliosarcoma, ganglioglioma, glioblastoma multiforme, gliofibroma; positive cells occasionally may be found in oligodendroglioma, capillary hemangioblastoma, choroid plexus papilloma, PNET, AT/RT
Neurofilament Ganglioglioma, gangliocytoma, PNET, neurocytoma, subependymal giant-cell tumor, AT/RT
Vimentin Mesenchymal tumors, meningiomas, sarcoma, melanoma, lymphoma, ependymoma, astrocytoma, gliofibroma, chordoma, schwannoma, hemangioblastoma, carcinoma, PNETs, AT/RT
S100 and neuron-specific enolase Positive in a variety of normal and neoplastic cells of neural and nonneural origin; of questionable utility for diagnostic purposes
Desmin Tumors containing muscle (rhabdomyosarcoma, teratoma, etc.), PNET
Cytokeratin Chordoma, choroid plexus tumors, meningioma, certain anaplastic gliomas, nongerminomatous germ-cell tumors, PNET, AT/RT
Epithelial membrane antigen Meningioma, ependymoma, epithelial areas of teratomas, AT/RT
Synaptophysin PNET, ganglioglioma, gangliocytoma, central neurocytoma, neuroendocrine tumors
Smooth muscle actin Tumors containing muscle, AT/RT
Retinal S-antigen Pineal parenchymal tumors, PNETs, retinoblastoma
Alpha-fetoprotein Embryonal carcinoma, endodermal sinus (yolk sac) tumor
Human chorionic gonadotrophin Germinoma, choriocarcinoma
Placental alkaline phosphatase Germ cell tumors
AT/RT, atypical teratoid-rhabdoid tumor; PNET, primitive neuroectodermal tumor.

Technical Handling of Tissues
For optimal study, tumor tissues removed at biopsy and postmortem require special handling. If a tumor is in an accessible location, the surgeon should be encouraged to remove and submit as much tissue as is safely possible; postmortem, generally no limitations apply. At any rate, a fresh sample should be sent for cytogenetic or other studies, a piece should be frozen, and tiny fragments should be fixed in glutaraldehyde in the event that ultrastructural study is required. Some of the specimen should be placed in formalin and the remainder in Bouin’s fixative. Fixation in Bouin’s creates three disadvantages, including the following: (a) cytogenetic studies cannot be performed, (b) synaptophysin antigen cannot be retrieved, and (c) the tissues cannot be tested for B-cell markers in the event that a lymphoma is suspected; hence, formalin-fixed material obviates these difficulties.
High-quality technical preparations are of utmost importance to establish a diagnosis. Use of the microwave enhancement technique for certain antigens is recommended.155,156 Even under the most optimal circumstances, however, classification based on phenotypes may be problematic.
Table 27.4 contains a listing of the widely available markers that are used most commonly and their utility in the differential diagnosis of tumors arising in the CNS. These tumors include primary neuroepithelial tumors, those arising from meningeal covering, and germ cell tumors.
No single clinical finding is pathognomonic for the diagnosis of a childhood brain tumor. At the onset of illness, the nature of neurologic and systemic dysfunction is varied. Signs and symptoms may be a direct result of tumor infiltration into adjacent brain and/or spinal cord or a consequence of CSF flow obstruction with resultant increased intracranial pressure (ICP). The clinical presentation primarily reflects the site of tumor origin, the age and developmental level of the affected child, and, occasionally, the tumor type. Clinical prodromes may include features of increased ICP, symptoms and signs of a localizing nature, or symptoms and signs without a localizing quality.
Increased Intracranial Pressure
Brain tumors cause increased ICP directly by infiltrating or compressing normal CNS structures or indirectly by causing obstruction of cerebrospinal fluid (CSF) pathways, resulting in noncommunicating hydrocephalus. Initial features of elevated ICP are typically insidious, nonspecific, and nonlocalizing. Among school-aged children, declining academic performance, fatigue, personality changes, and vague intermittent headaches are common. Over time, morning headaches, vomiting, and lethargy ensue. Papilledema may develop if the pressure is longstanding. Rapid progression of symptoms secondary to increased ICP is infrequent. However, when such occurs, a quickly growing midline or posterior fossa tumor requiring immediate intervention should be suspected.
Headaches resulting from brain tumors may have ominous features distinct from tension headaches or migraines.157,158 When children with a tumor are recumbent, increased ICP may worsen, resulting in a headache that wakens them at night or is present on waking in the morning. On arising, vomiting may occur along with some relief of pain. Once such patients are upright, the headache diminishes over the course of the morning. Over time, headaches gradually increase in severity

and frequency and clearly differ from any previous pain. The pain, which usually is frontal or occipital rather than temporal, may be further exacerbated with Valsalva maneuvers. The clinical suspicion for tumor should be greatest in those children with recent and continuing complaints of headache, and such should prompt a careful history and evaluation for related symptoms and signs. In fact, by 6 months from headache onset, nearly 100% of children have associated neurologic signs, such as papilledema, strabismus, ataxia, or weakness.157
Signs and symptoms of elevated ICP in infants and young children, whose skulls may more easily accommodate the growth of a mass lesion, may be quite different and may include irritability, anorexia, failure to thrive, and developmental delay or regression. Chronically increased pressure may lead to macrocephaly and separation of the cranial sutures. Infants may develop a tense or bulging anterior fontanelle associated with a shrill, neurogenic cry. Funduscopic evaluation of these patients may reveal only optic pallor but no evidence of papilledema. The setting-sun sign, a seemingly forced downward deviation of the eyes and part of Parinaud’s syndrome, may also be seen.
Parinaud’s syndrome is a collection of ophthalmologic findings stemming from increased ICP at the dorsal midbrain. In addition to the impaired upward gaze seen in infants, older children also display large pupils with impaired reflex constriction to light but not with accommodation. Convergence of gaze may evoke repetitive, bilateral, adducting nystagmus with retraction of the globes in the orbit. Cranial nerve IV palsy, with the affected eye deviated upward and laterally, may also occur. Affected children often compensate for the trochlear nerve palsy by tilting their heads toward the shoulder of the unaffected eye.
A head tilt may occur also with increased ICP because of a stiff neck and cervical root irritation from incipient cerebellar herniation of a posterior fossa mass. Other signs of increased ICP include listlessness and horizontal diplopia from pressure on the long, free intracranial course of the abducens nerve.
Localizing Symptoms and Signs
Children with supratentorial tumors (i.e., of the cerebrum, basal ganglia, thalamus, hypothalamus, and optic chiasm) may demonstrate various localizing symptoms and signs that precede those of increased ICP. The most common of these signs and symptoms include hemiparesis, hemisensory loss, hyperreflexia, seizures, and visual complaints.
Vision loss may localize to any location in the optic pathway. Complaints occasionally start insidiously with such events as a failed school eye examination or a need for eyeglasses. Tumors confined to the optic nerve produce monocular vision loss. Chiasmatic tumors present often with a complex visual field loss and decrement in acuity, whereas lesions located more posteriorly, in the optic tract, lateral geniculate nucleus, optic radiations, or occipital cortex, demonstrate some aspect of hemianopsia.159 A paradoxical increase in pupillary size to light when moving the source from one eye to the other indicates a relative afferent pupillary defect (the Marcus-Gunn pupil), a potential sign of tumor at the optic nerve or chiasm. Among infants, chiasmatic tumors may result in unilateral or bilateral pendular nystagmus, with head nodding and head tilt, a triad known as spasmus nutans.160
In contrast to the experience in adults, seizures are seldom the sine qua non of a supratentorial mass in children. Nevertheless, all/simple and complex partial (i.e., focal) seizures and most unexplained generalized (grand mal) seizures mandate computed tomography (CT) or MRI of the brain. After a first seizure and subsequent CT, fewer than 1% of patients are given diagnoses of a tumor.161 Seizure features that are associated with an increased risk of a tumor include a change in the character of preexisting seizures, status epilepticus as the first seizure, prolonged postictal paralysis, resistance to medical control, and the presence of focal symptoms or deficits.162,163,164 An initially normal CT scan in patients with any of these seizure characteristics or with persistent epilepsy does not rule out the possibility of a tumor, and repeat imaging with MRI may be indicated.
Other localizing signs of a supratentorial tumor may be more subtle. Children with frontal lobe tumors may have a long history of behavioral problems. Likewise, hypothalamic tumors in infants may cause failure to thrive and emaciation with a paradoxical euphoric mood and increased appetite, the so-called diencephalic syndrome, rather than motor or visual symptoms.165,166
For infratentorial tumors—those arising from the cerebellum and brainstem—localizing features include ataxia, long-tract signs, or cranial neuropathies. Initial cerebellar dysfunction may be insidious, with clumsiness, worsening handwriting, difficulty with hopping or running, or slow or halting speech. Tumors arising in the cerebellar hemispheres more commonly cause lateralizing signs, such as appendicular dysmetria and nystagmus, whereas midline cerebellar masses lead to truncal unsteadiness or increased ICP.
Cranial neuropathies often suggest brainstem pathology. Diplopia, with images seen side by side, is common from invasion of the abducens nerve within the pons. Inability to abduct one or both eyes (abducens palsy), however, can be a false localizing sign, because it may result also from increased ICP trapping the abducens nerve against the edge of the tentorium. Inability to deviate both eyes conjugately (gaze palsy) or the inability to adduct one eye properly on attempted lateral gaze implies an intrinsic brainstem disorder. These latter findings alone or, more likely, in combination with deficits of the trigeminal, facial, or auditory nerve strongly suggest tumor involving the brainstem. Indeed, masses involving the cerebellopontine angle may result in facial weakness, absent corneal reflex, and hearing loss. Weakness of an entire half of the face (peripheral seventh nerve palsy) suggests a posterior fossa tumor; weakness of the lower face on one side, with spared eyelid closure and forehead movement (central seventh nerve palsy), suggests involvement anywhere superior to the pons. Drooling and swallowing difficulties may arise from involvement of the medulla. A partial Horner’s syndrome (ipsilateral ptosis, miosis, and anhidrosis) may be present also in some patients with hypothalamic, brainstem, or upper cervical cord disease, as a result of compromise of the descending sympathetic tracts.
Nonlocalizing Symptoms and Signs
Some symptoms are characteristic of a brain tumor but not specifically localizing. Affected children may display changes in affect, energy level, motivation, or behavior. They may

exhibit weight gain or loss with anorexia. Sexual precocity or delayed puberty, growth failure, somnolence, or symptoms of an autonomic nature may suggest hypothalamic or pituitary dysfunction or may be nonspecific. Vomiting can occur with irritation of the area postrema in the floor of the fourth ventricle from a generalized increase in ICP or from direct irritation by a mass.
Figure 27.4 Left: Axial T2-weighted image reveals a large well-defined hyperintense pineal region mass. Center: Sagittal T1-weighted SE image demonstrating a hypointense mass in the pineal region (arrow). Right: Sagittal T1-weighted magnetization transfer contrast (MTC) post-Gd image showing uniform enhancement of the pineal region mass, which was proven to be a germ cell tumor.
As many as 15% of primary CNS tumors, particularly MB, germ cell tumors, ependymoma, and high-grade gliomas, have disseminated to other CNS sites by the time of diagnosis.167 Although such dissemination usually is asymptomatic, neurologic dysfunction from such lesions sometimes overshadows the symptoms of the primary tumor, confusing the localization of tumor origin. For example, spinal cord and cauda equina involvement may cause back or radicular pain, bowel or bladder dysfunction, or long-tract symptoms. Thus, examination at the time of diagnosis should include a search for local tenderness of the spine, focal extremity weakness, or sensory loss.
Syndromes Specific to Tumor Types
Although a pathologic brain tumor diagnosis requires tissue biopsy, certain patterns of symptoms and signs are particularly suggestive of specific tumor histologies. In the suprasellar region, pilocytic astrocytomas of the optic pathway and hypothalamus may cause visual field loss, nystagmus, and diencephalic syndrome.168 Craniopharyngiomas also occur in the suprasellar and sellar regions, but these neoplasms present more often with both visual deficits and endocrinopathies, particularly short stature and diabetes insipidus.169,170 Endocrinopathies may be obscured, however, if increased ICP and hydrocephalus from obstruction of the third ventricle and foramen of Monro are present.
Germ cell tumors may occur in the anterior hypothalamus or the pineal region (Fig. 27.4). Hypothalamic tumors frequently cause endocrinologic abnormalities such as growth failure and diabetes insipidus that precede the diagnosis by several years. Emotional and behavioral disturbances also can occur.
Pineal region tumors, including germ cell tumors and pineal parenchymal tumors, pineoblastoma and pineocytoma, are apt to be associated with Parinaud’s syndrome. Focal motor deficits appear more commonly with infiltrating glial tumors in the pineal region.171
In the posterior fossa, brainstem glioma, MB, ependymoma, and pilocytic astrocytoma form the oncologic differential diagnosis. MB and ependymoma often compress the fourth ventricle, leading to signs and symptoms of increased ICP. Vomiting may be extreme with ependymoma because of invasion of the area postrema, an emetic chemoreceptor on the dorsal medulla that protrudes into the fourth ventricle. The classic brainstem glioma, a diffusely infiltrative pontine glioma, presents with a prodrome of less than 6 months consisting of a triad of long-tract signs, ataxia, and cranial neuropathies, particularly an abducens palsy.172,173 The atypical, focal brainstem glioma presents with a longer prodrome, often without abducens palsy. Cerebellar pilocytic astrocytomas frequently present first with vague symptoms and then with ataxia of long duration, usually a period of 18 months.174 In the rare cerebellar hemangioblastoma, an elevated hemoglobin level may be noted, secondary to extramedullary hematopoiesis.175
Although a single seizure seldom is the presenting symptom for histologically malignant cerebral tumors, long-standing epilepsy may be associated with low-grade neoplasms.162,163,164,176,177 In children with long-standing epilepsy found to harbor a tumor, the most common diagnoses are ganglioglioma, dysembryoplastic neuroepithelial tumor (DNET), oligodendroglioma, and LGAs.178,179 Tumors are found in 12% to 33% of children who undergo surgery for intractable seizures.180
Among infants with brain tumors, seizures may occur in conjunction with macrocephaly as the harbinger of DIG, a massive, cystic, and malignant-appearing tumor with a favorable prognosis.181 CPP presents during infancy with hydrocephalus in nearly all cases. In congenital brain tumors, the most common diagnoses are malignant astrocytoma, teratoma, embryonal tumors, and CPP.182 For those tumors diagnosed within 2 months of birth, the mass occupies more than one third of the intracranial volume in 75% of patients.
Magnetic Resonance Imaging
Preoperative assessment of tumor type and extent, by imaging, is based on the combination of anatomic location, tissue characterization, and enhancement pattern, taken in conjunction with the clinical history. Since its introduction into clinical practice, MRI has superseded CT as the diagnostic tool of

choice for pediatric brain and spinal cord tumors. Advantages of MRI include the ease of imaging in three orthogonal planes without the need to move the patient, imaging without the use of x-irradiation, and improved soft tissue contrast. Nevertheless, the clinical presentation of children with brain tumors most frequently leads to initial evaluation by unenhanced CT. Whenever MRI is readily available in a timely fashion, CT with iodinated contrast is not recommended because of its inferiority in delineating tumor extent as compared with gadolinium-enhanced MRI.
Routine MRI sequences include T1-weighted imaging (T1WI) before and after gadolinium and T2- and proton density–weighted imaging. Postgadolinium imaging is usually performed with magnetization transfer suppression, which amplifies contrast enhancement by suppressing the signal intensity of normal background brain tissue. As a result, the detection of contrast enhancement is increased by a factor of two to three.183 This can be useful in demonstrating enhancement within a tumor, extension of the tumor along white matter pathways, and the subarachnoid spread of tumor. Notable is that contrast enhancement is a reflection not of vascularity but of breakdown of the blood–brain barrier (BBB) and, given this factor, neither CT nor MRI defines the true extent of tumor spread.
MRI offers other advantages over CT scanning such as fluid-attenuated inversion recovery (FLAIR) sequences, fast-echo planar imaging, and the capability to obtain volumetric measurements. The penumbra of edema surrounding a tumor, which may contain metastatic foci, can be delineated with a FLAIR sequence. This sequence may be useful to the radiation oncologist for targeting focal therapy, although it tends to overestimate the extent of tumor. Fast-echo planar imaging has enabled the development of diffusion, perfusion, and gradient echo (GE) techniques, which are discussed later in this section. Finally, with the advent of frameless stereotaxy, MRI has superseded CT for preoperative planning by virtue of the capability to acquire three-dimensional volumetric data that can be reformatted in any plane in the operating room, which allows for tumor localization in relation to markers placed on the skin.
Because it has been replaced by magnetic resonance angiography, conventional angiography rarely is performed in pediatric CNS tumors. Digital subtraction angiography still may be indicated, however, in those cases displaying a mass with blood and a differential diagnosis of vascular malformation versus hemorrhagic tumor. In addition, if a highly vascular tumor is suspected, diagnostic angiography may be performed as part of a neurointerventional procedure before resection to minimize blood loss.
Assessment of pediatric brain tumors has historically focused on morphology. However, with the introduction of higher field strengths, faster gradients, parallel imaging, and new sequence design, together with new contrast agents, the ability to combine parameters of function with anatomy may provide meaningful insights into tumor physiology. For example, T1-weighted as well as T2*-weighted (T2*-w), dynamic gadolinium enhanced imaging can be used to assess vascularity, permeability, and microcirculation of brain neoplasms. Characterization of cerebral blood flow is possible using dynamic magnetic resonance (MR) angiography, and activation functional MRI (fMRI) enables visualization of alterations in cortical blood flow secondary to selective cortical stimulation. Diffusion-weighted imaging (DWI) can be used to better delineate and even differentiate tumors. Spectroscopic chemical shift imaging allows for metabolic mapping both within and around tumors and helps to differentiate tumor recurrence from radiation necrosis. The combined use of these new techniques enables us to learn more about the pathophysiology of CNS lesions in vivo and may ultimately lead to improvements in the planning of therapy as well as prognostication.
It has been shown that repeated follow-up at appropriate time intervals is the best method for detecting early recurrence. To this end, techniques for re-localizing whole brain acquisitions, so that they are more strictly comparable over time for the radiologist to read and compare may improve the accuracy of interpretation and allow for earlier intervention, when required. Perhaps the single most important and highest impact factor will be the establishment and implementation of rigorous imaging protocols that will result in strictly comparable images with the exact same order and timing of image acquisition.
Magnetic Resonance Spectroscopy
Magnetic resonance spectroscopy (MRS) provides measurement of metabolites within the tissue under investigation. For example, proton MRS (HMRS) determines in both a qualitative and a quantitative fashion the chemical environment of the hydrogen nuclei within the tissues targeted. Frequency-domain spectra, which reflect the distribution of resonance frequency of the particular nuclei in the sample, form the data for analysis. Spectra are represented by a series of peaks with positions expressed in parts per million (ppm); the result can be considered a histogram of nuclei with different precession frequencies.
Spectra can be acquired using a single- or multivoxel technique, with short (10 to 30 millisecond) or long (135 to 280 millisecond) echo times. In using a short echo time, more peaks are captured, but the spectrum is superimposed by a complicated baseline, and its analysis is more difficult. With longer echo times, fewer peaks are captured, but the measurement precision is improved. In pediatric brain tumors, the three most important metabolic peaks (reading from right to left) are N-acetyl aspartate (NAA), 2.02 ppm; creatine-phosphocreatine (Cr/PCr), 3.02 ppm; and choline (Cho), 3.22 ppm (Fig. 27.5).
NAA is a marker of neuronal and axonal integrity. Cr is a marker for energy metabolism. Cho is a marker for cell membrane turnover and, as such, is elevated in tumors, demyelination, and inflammation; it is decreased in liver disease. The relative ratios of different metabolites vary, depending on the location of the voxel in the brain and on age during the first 5 years of life, when myelination of the immature brain increases. It is, therefore, imperative to have normal age-matched control data from the same brain region in interpreting the spectra from young children. As a general rule, the NAA increases over time, especially during the first 18 months of life, whereas the Cho slightly decreases over the same period. Creatine-phosphocreatine tends to remain rather stable over time; for this reason, it has historically been used as an internal control when metabolic data are expressed as ratios.
Figure 27.5 Normal long-echo single-voxel spectrum from the cerebellum of an age-matched control for a patient with medulloblastoma (see Figure 27.8). Reading from right to left, note normal peaks of N-acetyl aspartate (NAA), creatine-phosphocreatine (Cr/PCr), and choline (Cho).

The finding of a lipid-lactate peak usually indicates the presence of ischemia or necrosis. Phosphorus 31 MRS has been used in head and neck tumors, especially in the diagnosis and management of lymphoma. However, it is more time-consuming and a generally less available technique. Fluorine MRS has been used to research tumor drug pharmacokinetics, such as those of 5-fluorouracil. The uptake of drug in the tissue of interest and its subsequent metabolism can be followed.
Single-voxel MRS can be used to interrogate new tumors having a volume greater than 1 mL. However, a multivoxel technique, such as two-dimensional chemical shift imaging, in which several subcentimeter voxels can be examined simultaneously, can be very helpful in distinguishing recurrent tumor from radiation necrosis. From a compilation of two-dimensional chemical shift imaging spectra, one can obtain a three-dimensional MRS data set. For the future, whole-brain single-metabolite techniques are being developed,184 which, in addition to double-quantum spectroscopy, will further and more accurately separate out the metabolic peaks.
Single-voxel long TE MRS data obtained by plotting Cr:Cho against NAA:Cho ratios has been used with benefit in separating posterior fossa MB from juvenile pilocytic astrocytoma (JPA) and ependymoma and in predicting disease progression.185 It has also been used in the presurgical diagnosis of large suprasellar tumors, successfully separating craniopharyngioma from pituitary adenoma and hypothalamic region astrocytoma using single-voxel short or long TE stimulated acquisition mode (STEAM) or point resolved spectroscopy (PRESS) acquisitions.186 In other series, MRS findings have been shown to have prognostic information for supratentorial tumors as evidence by low NAA:Cho and Cr:Cho ratios in patients who died versus high ratios in survivors.187 MRS also appears to be useful in monitoring the response of histologically proven pediatric glioma to adjuvant chemotherapy or radiation therapy as demonstrated by correlations between the ratio of tumor Cho to brain Cho versus tumor volume or clinical response.188 In patients with recurrent brain tumors, the Cho:NAA ratio also appears to have prognostic significance; children with a maximum Cho:NAA ratio of less than or equal to 4.5 had a projected survival of more than 50% at 63 weeks.
Another application of MRS is in distinguishing recurrent tumor from radiation necrosis, which can be the great mimicker, having both mass effect and enhancing following Gd administration.189
Gradient Echo Imaging
T2* GE imaging has been used with good effect in detecting the presence of altered blood and blood products within tumors. There is some evidence for its lack of sensitivity and it may be replaced in the future by susceptibility-weighted imaging (SWI), a form of “super” GE sequence. High-resolution three-dimensional GE imaging with long echo times is used to create high contrast phase images. A phase mask is used to enhance the tissue contrast in the magnitude images making it possible to differentiate tissues with small differences in iron content, to separate fat and water, and to better visualize clot and microhemorrhage. Although spectral information is usually obtained by collecting a time series of data, it is possible to extract spectral imaging from phase alone. Phase images contain direct information about the background magnetic field and chemical shift of tissues. It is claimed that SWI can be used (a) to separate veins from arteries, (b) to image vessel walls, (c) to image microhemorrhage and brain iron, and (d) to enhance T1 contrast in gray/white matter.190,191,192 SWI has already proven itself superior to conventional GE imaging in the detection of hemorrhage in traumatic brain injury and all the previously mentioned properties lend themselves to a better understanding of the physiology of tumors, an example of which as shown in Figure 27.6.
Diffusion-Weighted Imaging
DWI describes Brownian, or random, motion of water molecules. However, in the intracellular and extracellular spaces of the brain, macromolecular proteins, intracellular organelles, cell walls, and myelin sheaths restrict or slow diffusion in certain directions. This restriction results in a new directionality of diffusion within given regions of the brain. This directionality, or anisotropy, is most noticeable in the white matter tracts in which diffusion tends to be much faster in the direction parallel to myelinated axons. Images can be made for individual directions, usually corresponding to the three orthogonal imaging planes: axial, coronal, and sagittal. Alternatively, the individual directional images can be averaged to produce a relatively direction-independent, or isotropic, diffusion image called a trace image. From this, a mean apparent diffusion coefficient (ADC) map can be calculated.
Magnetic Resonance Perfusion Imaging
In addition to DWI, the development of fast-echo planar imaging has made possible an assessment of the vascularity of tumors using a gadolinium first-pass bolus technique. This relies on changes in the T2 signal of gadolinium-laden blood as it passes through the region of interest.193 Resulting data, reflected in maps of relative cerebral blood volume, provide

some semiquantitative analysis of the blood flow to a particular region. Early work suggested a positive correlation between relative cerebral blood volume and tumor grade.194 Perfusion imaging may be helpful in targeting a lesion for biopsy. Application of perfusion imaging may be particularly useful for the study of neovascularization and angiogenesis inhibition.
Figure 27.6 Left: T1-weighted magnetic resonance imaging (MRI) with contrast. Right: Susceptibility-weighted image (SWI) showing hemorrhage in a glioma (arrow). (Courtesy of Mark Haacke, PhD, The Magnetic Resonance Imaging Institute for Biomedical Research, Detroit, Michigan.)
Measurements of tumor blood flow (TBF) are important for understanding tumor physiology and can be valuable both in selecting and evaluating therapies. For example, it has been shown that brain tumors typically demonstrate reduced blood flow and a slower transit time, when compared to normal brain tissue. Because radiation therapy preferentially kills those cells that contain substantial amounts of dissolved oxygen (i.e., high blood flow), the anoxia and hypoxia that results from cells in a microenviroment with reduced blood flow predisposes them radioresistant. Likewise, TBF measurements may theoretically facilitate the evaluation of therapeutic effectiveness for a particular antiangiogenic agent used in the treatment of CNS tumors. MR perfusion imaging is performed either by utilizing a first-pass Gd bolus technique or by using arterial spin tag labeling. The Gd technique takes advantage of the T2* decrease in signal effect of the first pass of a bolus of contrast while running a GE sequence sensitive to T2* effects. This enables a time versus signal intensity curve to be drawn from which a variety of parameters such as mean transit time (MTT), time-to-peak, area under the curve, a measure of relative cerebral blood volume (rCBV), and relative cerebral blood flow (rCBF) can be calculated. Although this technique has been applied to a range of tumors over the past decade, it is associated with a variety of problems, not the least of which lies in the fact that tumors have a complex vascular supply including the promotion of angiogenesis as well as the presence of areas of necrosis.195,196 The enhancement that is seen in any tumor on T1WI following the administration of contrast does not merely reflect the vascularity of the lesion but also demonstrates evidence of breakdown of the BBB. This leads to the so-called leaky tumor, which can only be accurately solved mathematically using a multicompartmental model and which will be different for each individual tumor and vary at different stages in the natural history of each tumor.
Activation Functional Magnetic Resonance Imaging
After the use of a stimulation paradigm or task designed to activate a specific functional area of the brain, the target can be located anatomically by an increase in blood flow to that area. Functional MRI uses the BOLD (b lood oxygen level–dependent) technique to generate differential signals based on the relative hemoglobin-oxyhemoglobin content of the blood flowing away from activated brain during an appropriate stimulus. Repetition of the task improves the robustness of the data, and subtraction of rest from activity reduces background signal. Data are presented on maps that outline the activated area of interest in relation to the lesion, and these may be useful in preoperative planning. An example involving finger-thumb opposition and the motor cortex of the hand is shown in Figure 27.7. Different activation tasks can be designed to stimulate other eloquent areas of the brain for vision, hearing, and language.
Single-Photon Emission Computed Tomography and Positron Emission Tomography
A limited number of reports address the role of thallium 201 single-photon emission CT (SPECT) in pediatric brain

tumors. In a heterogeneous and somewhat limited series, although representative of the pediatric population, Rollins et al.197 failed to establish any clinically useful indication for the test. Tl SPECT was less sensitive and less specific than was gadolinium-enhanced MRI and did not correlate with histologic grade, biologic aggressiveness, or tumor type. In children with brainstem gliomas, however, Tl SPECT was able to distinguish persistent or recurrent tumor from radiation necrosis.198
Figure 27.7 Preoperative functional magnetic resonance imaging in an 11-year-old boy presenting with left-sided focal motor seizures, performed for surgical planning. The bright pixels (black arrow) represent the statistical parametric map produced by using a left-sided finger-thumb opposition paradigm. The bright pixels reflect the increase in oxygenated blood flowing away from the right-sided sensorimotor cortex. The white arrow indicates the solid tumor component within a cystic mass displacing the sensorimotor cortex anteriorly. Histopathology demonstrated a cystic dysplastic ganglioglioma that later was removed successfully via a posterior approach without damage to the child.
Figure 27.8 Left: Sagittal T1-weighted through the posterior fossa demonstrating slightly hypointense tumor mass within the posterior fossa (arrow). Right: Axial apparent diffusion coefficient (apparent diffusion coefficient through the posterior fossa demonstrating slightly returning a pattern of restricted diffusion consistent with a medulloblastoma (arrow).
Positron emission tomography (PET) adds another dimension to brain imaging. Using the appropriate tracer or combination of tracers, subtle metabolic changes can be measured with a relatively high spatial resolution approximating 3 to 4 mm. Registration of PET with MRI is an issue that will have to be resolved if meaningful application of this technique using fluorodeoxyglucose and other isotopes is to occur in the clinical arena.199
Imaging Characteristics of Central Nervous System Tumors
Posterior Fossa Tumors
The differential diagnosis of cerebellar tumors in children consists in large part of MB, or PNET, JPA, ependymoma, hemangioblastoma, and exophytic brainstem glioma. Although MB tends to be more solid and less homogeneously enhancing and JPA more commonly cystic and typically strongly enhancing, distinguishing one from the other still can be difficult with CT or conventional MRI alone. However, DWI may help in the differential diagnosis preoperatively. Because of the high, tightly packed cellular content of MB, a pattern of “restricted” diffusion appears to occur in these tumors, similar to that seen in stroke (Fig. 27.8).200 This is reflected by a reduction in the ADC. The ADC reflects physical factors, such as temperature and viscosity, in addition to the relative ease or restriction of the motion of molecules through tissues and membranes. The ADC of regions of tumors appears to correlate with cellularity; there is a tendency for lower ADC values in high-grade gliomas and higher values in LGAs. In comparing tumor tissue to normal brain, ADC may be important in distinguishing regions of edema from nonenhancing tumor as well.201
MRS may help also in the preoperative differential diagnosis of posterior fossa tumors. Using a discriminant analysis, single-voxel MRS of pediatric cerebellar tumors has been proven capable of separating out MB (or PNET), JPA, and ependymoma from normal cerebellar tissue based on a plot of Cr:Cho ratios against NAA:Cho ratios (Fig. 27.9).185
Using CT, MRI, and MRS characteristics, the following can be said of posterior fossa tumors: The solid component of JPA is hypodense on CT, hypointense to gray matter on T1WI,

and hyperintense on T2WI, demonstrating marked enhancement after administration of gadolinium. DWI usually shows evidence of unrestricted diffusion, especially in the presence of a cystic component. MRS tends to show moderate reduction in NAA and modest elevation in Cho (Figs. 27.10 and 27.11). MB is hyperdense on CT and uniformly hypointense to gray matter on T1WI and tends to be isointense to gray matter on T2WI with restricted DWI, most likely secondary to the tumor’s dense cellularity and high cellular nucleus-to-cytoplasm ratio. MRS tends to show very low NAA with very high Cho (Fig. 27.12). Ependymoma may be distinguished by its different anatomic location and pattern of spread through the foramina of Luschka and Magendie, and its characteristic appearance of speckled calcification on CT, but MRS can give additional support to the diagnosis, especially in those cases in which the size and extent of the mass render difficult the identification of its point of origin. Solitary cerebellar hemangioblastoma usually will demonstrate serpiginous flow voids, reflecting the vascularity of the lesion, in addition to intense enhancement after intravenous contrast.202 Intrinsic hemangioblastoma may be seen also in the spinal cord.
Figure 27.9 N-acetyl aspartate–choline (NAA:Cho) versus creatine-choline (Cr:Cho) scattergram for astrocytoma, ependymoma, medulloblastoma, and normal cerebellar tissue. The straight lines are boundaries between the three tumor types found by discriminant analysis. PNET, primitive neuroectodermal tumor.
Figure 27.10 Sagittal T1-weighted midline image demonstrating posterior fossa tumor occupying most of the fourth ventricle extending inferiorly to the obex, with mild prominence of the supratentorial ventricular system. The histopathologic diagnosis was a juvenile pilocytic astrocytoma.
Figure 27.11 Long echo time (TE) magnetic resonance spectroscopy (135 milliseconds) from the same patient mentioned in Figure 27.10. Note the moderate decrease in N-acetyl aspartate (NAA) and moderate elevation in choline (Cho), a typical pattern for low-grade glioma.
Figure 27.12 Composite image demonstrating the set-up of a single voxel of interest in all three planes for long echo time (TE) magnetic resonance spectroscopy in a posterior fossa tumor. Note almost complete absence of N-acetyl aspartate at 2.02 ppm, loss of creatine-phosphocreatine peak at 3.02 ppm, and gross elevation of choline at 3.22 ppm. Histopathology confirmed a medulloblastoma (primitive neuroectodermal tumor).

Other rare tumors seen in the posterior fossa include AT/RT that are typically rapidly progressive, with surrounding edema and enhancement from breakdown of the BBB.
Brainstem Tumors
MRI is the procedure of choice for identifying brainstem tumors. Some work has shown a correlation between 201Tl SPECT and gadolinium enhancement suggest that, like gadolinium, thallium uptake requires breakdown of the BBB.198 The presence of absence of calcification, an unusual manifestation of brainstem tumors, is not evident from MRI but may be observed on CT. Calcification may occur rarely with oligodendroglioma of the brainstem and occasionally with extension of fibrillary astrocytoma of the cervical cord into the medulla or in association with vascular lesions. Dystrophic calcification secondary to radiation therapy may also occur. However, the major role of imaging is to distinguish focal, potentially resectable exophytic tumors, from diffuse intrinsic tumors, for which the diagnosis can be established by imaging alone. The latter tumors characteristically show T2 signal abnormality extending along the entire anteroposterior dimension of the brainstem. A recent retrospective review of the radiologic and pathologic findings of gadolinium-enhancing brainstem lesions on MRI suggests that the majority of enhancing tumors are low grade and potentially amenable to surgical resection.203
Supratentorial Tumors
Intraaxial tumors, which usually are LGAs, are imaged in much the same way, as are posterior fossa tumors. Published pediatric data concerning MRS of these lesions are few, but the peak-area ratios or NAA:Cho and Cr:Cho may be prognostic.187 Experience in this area of imaging is evolving.
Figure 27.13 Left: Axial T2-weighted image demonstrating a well-defined mixed signal intensity mass with marked surrounding edema. Center: Axial apparent diffusion coefficient map demonstrates a medial focus of restricted diffusion suspicious for an aggressive tumor. Right: Axial T1-weighted postgadolinium magnetization transfer contrast (MTC), image through the tumor demonstrates marked rim enhancement. Histopathology revealed a glioblastoma multiforme.
Ganglioglioma, DNET, and hypothalamic hamartoma have fairly specific imaging characteristics. Ganglioglioma and DNET tend to be superficial lesions with a predilection for the temporal lobe. Gangliogliomas may be cystic and demonstrate a variable pattern of relatively poor enhancement. DNET classically causes remodeling of the overlying inner table of the skull, a reflection of their natural history of slow growth. The anatomic location, lack of enhancement, and commonly associated history of gelastic seizures, precocious puberty, or other hormonal imbalance in a young child leads to a diagnosis of hypothalamic hamartoma.
High-grade glial tumors (Fig. 27.13) demonstrate a pattern of marked enhancement with evidence for restricted diffusion on ADC maps. MRS of these tumors demonstrates a marked elevation in choline with a reduction in NAA, not dissimilar to the pattern seen with MB.
Extraaxial sellar and pineal region tumors, together with meningiomas, lack a BBB and therefore enhance avidly after contrast administration. Single-voxel MRS has been used to differentiate large cystic sellar and parasellar masses on the basis of their metabolites.186 Hypothalamic gliomas have a metabolic profile similar to that of JPAs and other LGAs, with a moderate reduction in NAA along with a moderate elevation in Cho. This is different from pituitary adenomas, which demonstrate no NAA because they contain no neuronal tissue, a moderate elevation of Cho as a marker of cell membrane turnover, and variable amounts of lipid-lactate, depending on the degree of tissue necrosis. The “crankcase” oily contents of cystic craniopharyngiomas lead to a large lipid peak with little else.
Spinal Tumors
Intrinsic tumors of the spinal cord, such as primary gliomas and ependymomas, are best imaged with MRI. It is also the imaging modality of choice for metastatic disease in the subarachnoid space (e.g., from MB). Ideally, as time and patient condition permit, the spine is scanned preoperatively when MB or ependymoma is suspected. When such a scan cannot be performed, a baseline study should be performed approximately 3 weeks into the postoperative period to avoid the problems of interpreting the appearances of postsurgical blood and arachnoiditis. Most treatment protocols direct repeat imaging for studies showing and not showing disease. However, it may be reasonable in the nonprotocol setting, when initial spine study results are negative, to repeat the spine MRI after completion of therapy only if new symptoms develop. CSF cytology is complementary to spinal imaging in detecting subarachnoid disease.
Imaging Characteristics of Neurofibromatosis Type 1 and Associated Tumors
NF-1 frequently manifests as focal areas of signal intensity (FASI) on T2WI. Typical imaging features in affected young children include T2 bright signal returned from the basal ganglia and dentate nuclei, which gradually disappears in

young adulthood. Limited histopathology suggests that these lesions may represent spongiform dysplasia.204 However, it is recognized that some of these FASI grow and start to enhance and are diagnosed ultimately as gliomas, usually of low grade. Tumors identified in patients with NF-1 most commonly are located in the brainstem, optic chiasm, and hypothalamus and, less commonly, below the tentorium. Anomalies of the corpus callosum, including LGAs, have been identified also in children with NF-1.205 Three-dimensional multivoxel proton MRS has been used in children with NF-1 to interrogate these areas of focal signal abnormality. Proton MRS indicated that FASI (a) are characterized by significantly elevated Cho, reduced Cr, a Cho:Cr ratio greater than 1.3, and near-normal NAA levels; (b) are different from tumors that exhibit Cho:Cr greater than 2 and no NAA; (c) have no lipid or lactate signal; and (d) correlate in spatial extent but are more extensive than indicated by conventional MRI sequences.206 Use of three-dimensional CSI, as opposed to single-voxel imaging spectroscopy, to delineate metabolically the extent and volume of the lesion, may be useful for following the effects of treatment (Fig. 27.14).
Figure 27.14 Composite image demonstrating a set-up for three-dimensional chemical shift imaging (stack of two-dimensional slabs), in a patient with neurofibromatosis type 1 complicated by a brainstem glioma. The tumor is outlined from the spectroscopy grid by the extent of the abnormal N-acetyl aspartate–choline pattern.
Screening neuroimaging of asymptomatic children with NF-1 has not been shown to improve clinical outcome. Serial ophthalmologic examinations of affected children are critical. Those with unexplained ophthalmologic abnormalities should undergo MRI examination of the head and orbits.207 Optic nerve lesions seen in association with NF-1 generally are JPAs; they demonstrate enhancement on postgadolinium imaging. Unless directed otherwise by clinical protocol, follow-up imaging of these tumors should be considered only on evidence of deterioration of vision.
Tuberous Sclerosis
In tuberous sclerosis, subependymal giant-cell astrocytoma classically occurs at the level of the foramen of Monro. These lesions often are bilateral, may be calcified, and usually demonstrate marked enhancement with intravenous contrast. They require continued follow-up and repeat imaging, on an annual basis, to rule out the development of hydrocephalus. Tubers typically demonstrate elevated myoinositol on MRS.
Radiation Necrosis
Radiation necrosis reflects local damage to tissue following radiation therapy. It is a great mimicker and may demonstrate mass effect, enhance after intravenous contrast, and is often indistinguishable from recurrent or progressive tumor. Fluorodeoxyglucose PET has been considered the gold standard in distinguishing tumor progression from radiation-induced necrosis. However, an incidence of false-negative and false-positive results now is recognized.208 Two-dimensional CSI/MRS can be used to interrogate the region of interest on a subcentimeter voxel-by-voxel basis. A reduction or absence of both NAA and Cho, in the presence of lipid and lactate, is found in radiation necrosis. In the presence of recurrent tumor, elevation of Cho occurs.209
Imaging and Radiation Treatment Planning
Many of the techniques described in this section allow for more accurate localization of tumor and tumor extent than was historically obtained with CT imaging. MRI should delineate more clearly between tissue planes and improve image fusion needed for planning radiation therapy. More experience is needed to determine whether this capability will result in improved patient outcome.
Future Developments
Currently under development is the technique of whole-brain spectroscopy.184 This modality is performed in conjunction with three-dimensional volumetric imaging of the brain and allows for a quantitative expression of individual metabolites, such as NAA and Cho, expressed per gram of brain tissue. Other advances in MRS include the development of two-dimensional double-quantum spectroscopy, which will further and more accurately separate out the metabolic peaks.210,211,212
Diffusion tensor imaging is a technique performed using six to nine planes of imaging as opposed to the three that typically are employed for DWI in standard MRI. The data acquired allow the mapping of white matter tracts and detection of interruptions in them due to infiltrating tumor.213 The same data can be used to measure fractional anisotropy, and those results may allow further discrimination of differential tissue characteristics.
Finally, the development of pulsed and continuous arterial spin tag labeling as a technique to study perfusion of brain tissue should permit quantitative or semiquantitative analyses of blood flow.214 This procedure might have application to the assessment of angiogenesis in tumors and effects of antiangiogenic therapy.
For most CNS tumors, surgical intervention forms the initial step in the treatment plan by providing tissue to establish the

histologic diagnosis and, when possible, reducing the tumor burden. Notable exceptions to this rule are certain unresectable tumor types, such as diffuse infiltrative brainstem gliomas215 and globular chiasmatic gliomas in children with NF-1. Each of these tumors has an almost diagnostic MRI appearance in association with consistent histologic features. As surgery has been shown not to improve prognosis in patients with these tumors and the histologic diagnosis rarely is in question, operative intervention is not required. In addition, for certain deep-seated tumors, the risks imposed by a conventional surgical approach to them may be high and, for establishing a histologic diagnosis, stereotactic biopsy techniques provide a reasonable alternative to open tumor debulking. CT- or MRI-guided stereotactic procedures are highly accurate in targeting such deep-seated lesions and are associated with a morbidity of fewer than 5% and a mortality of fewer than 1% of patients.216,217
For the majority of pediatric brain tumors, however, open operations are preferred; the goal in such procedures is to remove as much tumor as is safely possible. Although a truly complete tumor resection is feasible only for well-circumscribed benign tumors, such as pilocytic astrocytomas and craniopharyngiomas, an extensive, near-total resection can be achieved with many parenchymal tumors. The limitation to complete resection in such tumors is the imperceptible blending of the neoplasm into the surrounding brain; scattered tumor cells may infiltrate past the margins of the resection into the normal parenchyma. Unlike the situation with solid, non-CNS tumors, it is rarely feasible to resect the tumor with surrounding margins of normal tissue because of the unacceptable risks of producing irreversible neurologic deficits. An outline of the neurosurgical approach and considerations for the most common types of brain tumors is provided in Table 27.5.
Preoperative and Perioperative Considerations
In occasional affected children who present with obtundation from a large mass lesion, resection is performed urgently. In children who are awake and alert but nonetheless harbor a large lesion that is producing substantial mass effect, the tumor resection is performed on the next operating day. Smaller lesions without significant mass effect can be managed more electively.
Because peritumoral edema commonly contributes to the neurologic impairment produced by the tumor, moderate doses of corticosteroids generally are administered preoperatively. For example, 0.1 to 0.5 mg per kg dexamethasone may be given every 6 hours. This dosing often will lead to a dramatic improvement in the patient’s symptoms and signs, avoiding the need for emergency surgery in the vast majority of cases. Corticosteroids are typically continued intraoperatively and during the early postoperative period. If a significant reduction in tumor volume has been achieved at the time of operation, corticosteroid therapy is tapered and then discontinued within several days of surgery.
Another factor that commonly contributes to increased ICP in children with brain tumors is the presence of obstructive hydrocephalus, observed most commonly in tumors arising near the aqueduct, such as with pineal region tumors, or the fourth ventricle, as seen with cerebellar vermian lesions. Although resection of the tumor often opens the CSF pathways and leads to resolution of the hydrocephalus, the resection frequently is rendered safer if the elevated pressure is relieved as an initial step. The use of preoperative shunting carries a small risk of upward herniation through the tentorial hiatus. In some institutions, preoperative relief of hydrocephalus is accomplished by endoscopic third ventriculostomy;218,219 however, a more common approach is to place an external ventricular drain immediately before the craniotomy for tumor resection, which has the advantage of allowing drainage of bloody spinal fluid and debris in the early postoperative period. If the operative procedure opens the CSF pathways and the patient’s absorptive pathways remain patent, the external ventricular drain often can be removed within several days of surgery.220 If the hydrocephalus persists, a third ventriculostomy can be performed or, if the hydrocephalus appears to be of a communicating type, a ventriculoperitoneal shunt can be inserted. Although in the past a great concern was that shunting would provide a route for systemic dissemination of tumor, more recent studies have failed to show an increased risk of tumor spread via the shunt.220 Overall, the use of shunts in children with posterior fossa tumors has diminished substantially.
The endocrinopathies commonly manifested by hypothalamic tumors can be exacerbated by tumor resection. Patients undergoing this procedure typically require stress doses of hydroxycorticosteroids before, during, and after surgical intervention. Although thyroid hormone replacement is occasionally instituted preoperatively, it is most commonly initiated postoperatively. Patients in whom the posterior pituitary stalk is sectioned or injured during surgery often manifest a triphasic response of impaired fluid regulation, characterized by an initial period of transient diabetes insipidus lasting 1 to 2 days, a subsequent period of inappropriate antidiuretic hormone release lasting several days, and a final phase of persistent diabetes insipidus. In view of the rapid changes in vasopressin levels during the first several days postoperatively, careful attention to fluid replacement and cautious administration of synthetic vasopressin, where indicated, are essential to avoid potentially deleterious swings in electrolyte levels and fluid balance.
Children with cerebral cortical tumors and those in whom cortical retraction is required in the approach to a deep-seated lesion may be at risk for seizures during the perioperative period. Preoperatively, such patients often are started on an anticonvulsant medication (e.g., phenytoin) that is continued during the postoperative period, even if they have not experienced previous seizures. Patients generally are maintained on anticonvulsants for at least 1 week postoperatively; the decision to continue such therapy in patients without documented seizures is of uncertain benefit.221 In patients who experience a preoperative seizure disorder from the tumor and have been rendered seizure-free by tumor resection, anticonvulsants often can be stopped within several months after surgery.
Intraoperative Considerations and Surgical Technique
Recent studies have indicated that the extent of surgical resection has a major impact on the likelihood of long-term survival


for children with many types of pediatric brain tumors, particularly ependymomas,222,223,224,225,226,227,228 high-grade gliomas,229,230,231,232 MBs,233,234,235,236,237,238 LGAs,239 and choroid plexus tumors.240,241,242 Accordingly, with the exceptions noted earlier, in which surgical resection is not indicated or in which stereotactic biopsy may be a preferable initial step, extensive resection is the goal for many types of pediatric brain tumors.
Tumor Type Surgical Interventionb Common Morbidity Comments
Supratentorial Hemispheric      
Low-grade glioma GTR if possible Varies with site (e.g., hemiparesis, hemi-sensory deficits, hemianopsia, seizures) Outcome improved by extensive resection
High-grade glioma GTR if possible Varies with site Outcome improved by extensive resection
Mixed neuronal-glial GTR if possible Varies with site Outcome improved by extensive resection
Ependymoma GTR if possible Varies with site Outcome improved by GTR
Choroid plexus tumors GTR if possible Hemiparesis, subdural hygromas, life-threatening blood loss in infants Outcome improved by GTR
PNET GTR if possible Varies with site Outcome possibly improved by extensive resection, but confirmatory data lacking
Supratentorial Midline      
Chiasmatic-hypothalamic glioma Resection reserved for exophytic lesions with substantial mass effect; biopsy used if the diagnosis is uncertain; in NF-1, the diagnosis is often made by imaging Vision loss common with biopsy; hypothalamic dysfunction and hemiparesis with extensive resection In infants, possible to delay need for adjuvant therapy by extensive resection; remains controversial whether this improves results over those obtained with biopsy and chemotherapy
Craniopharyngioma Depending on lesion type, GTR, subtotal resection, or stereotactic approaches potentially optimal Increased neuroendocrine deficits common after extensive resection; vision loss, personality changes, and neurologic impairments also common Progression-free survival possibly improved by extensive resection, but potential for morbidity also increased
Germinoma Biopsy to establish diagnosis Risk of deterioration low with stereotactic or open biopsy Excellent outcome with adjuvant therapy alone
Malignant germ cell tumors Biopsy, GTR if possible; resection of residual disease after adjuvant therapy, if possible Extraocular muscle dysfunction Often only scar tissue or differentiated teratoma disclosed on extensive resection of residual disease after chemotherapy
Pineal parenchymal tumors (pineoblastoma, PNET, pineocytoma) Biopsy, GTR if possible Extraocular muscle dysfunction Outcome possibly improved by extensive resection, but confirmatory data lacking
Medulloblastoma (PNET) GTR or near-total resectionc if possible Dysmetria, ataxia, cranial nerve palsy, paresis, transient mutism Outcome improved by extensive resection in patients with M0 disease
Cerebellar astrocytoma GTR if possible Dysmetria, ataxia GTR usually curative
Ependymoma GTR if possible Dysmetria, ataxia, cranial nerve palsy, paresis, transient mutism Outcome improved by GTR
Diffuse (malignant) brainstem glioma No; generally an imaging diagnosis, except in rare instances Cranial nerve dysfunction, paresis Benefits unproven
Benign (focal) brainstem glioma Craniotomy, near-total resection feasible in selected cases Cranial nerve dysfunction, paresis Excellent outcome for dorsally exophytic and cervicomedullary tumors after near-total resection; for other focal tumors, remains uncertain whether results after resection are superior to those after biopsy and irradiation
DIG, desmoplastic infantile ganglioglioma; DNET, dysembryoplastic neuroepithelial tumor; GTR, gross total resection; NF-1, neurofibromatosis type 1; PNET, primitive neuroectodermal tumor.
aAlthough this schema summarizes general management approaches, a more extensive presentation of treatment caveats for individual tumor types is presented within the text. Because the details of treatment for many tumor types have evolved over time and will likely continue to evolve, treatment decisions for individual patients are best made in the context of a multidisciplinary approach.
bPerioperative measures, such as the use of corticosteroids to reduce mass effect and the use of anticonvulsants for lesions at high risk for seizures, are provided in the text.
ccTumor that infiltrates the brainstem surface generally is not resected aggressively (hence the term near-total rather than gross total), because this small amount of surgically visible residual disease (which generally is not detectable on MRI) does not appear to affect prognosis adversely. This decreases the risk of cranial nerve morbidity.
A major limitation to the widespread incorporation of extensive surgical resections in the management of childhood brain tumors has been the fact that aggressive resections may increase the risk of immediate and long-term morbidity, particularly for tumors in functionally critical locations. Although morbidity generally is less than 10% for polar supratentorial gliomas and less than 20% for cerebellar astrocytomas, more deep-seated lesions, such as ependymomas and craniopharyngiomas, carry morbidity rates in excess of 40% with extensive resection.243 Although some studies have observed that morbidity is lower if operations are performed by neurosurgeons who do such operations frequently,244,245 recent studies also indicate that pediatric neurosurgeons are more likely to attempt extensive removals, on the basis of their recognition that this influences prognosis, and therefore may have overall rates of management morbidity comparable to those of general neurosurgeons, albeit with a higher frequency of complete or nearly complete tumor resections.246
During the last 10 to 20 years, a number of intraoperative modalities have been developed or refined to allow tumor resection to be performed more safely and efficiently. Foremost among these are the progressive improvements in operative microscopy, which facilitates illumination and visualization of the interface between neoplasm and normal brain. Localization techniques, such as frame-based and, more recently, frameless stereotactic guidance systems, allow preoperative targeting of the tumor so that the surgical approach can be tailored precisely to minimize manipulation of normal brain structures and to maximize the extent of resection of deep-seated subcortical lesions.247,248
Ultrasonographic guidance also is useful in this regard. Recently, intraoperative MRI units have become available in some centers and may help to refine further the accuracy of intraoperative decision making,249 provided that issues of cost and the difficulties involved in operating in or adjacent to a high field-strength magnet can be resolved effectively.
In children whose tumors are in and around functionally critical brain regions, intraoperative monitoring of visual, auditory, and somatosensory pathways and direct assessment of motor and speech pathways often are used in an attempt to improve the safety of the tumor resection. In addition, areas of essential cortex overlying a deep-seated tumor may be delineated using cortical stimulation techniques to plan an approach to the tumor that avoids traversing important structures. Functional MRI also provides a useful way of noninvasively localizing important cortical areas250 to identify a safe trajectory to an underlying lesion. Finally, in children with intractable seizures from cerebral neoplasms, intraoperative or extraoperative electrocorticography (ECOG) may be used to define areas of epileptogenic cortex in and around the tumor to increase the likelihood that seizure control will be obtained postoperatively.251
For supratentorial craniotomies, children generally are placed in the supine or lateral position or prone for occipital lesions. For infratentorial craniotomies, the prone or lateral position is used more often than the “sitting” position because of concern over potential venous air embolism. Intraoperatively, the head of an infant often is positioned on a soft headrest, rather than being held by pins that can perforate the skull or cause a depressed fracture. For cortical and many subcortical tumors, the surgical approach follows the most direct trajectory to the lesion. However, for deep-seated lesions that are subjacent to functionally critical regions of the brain, alternate approaches often are required. Details of the operative approaches for specific tumor types are provided in subsequent sections of this chapter.
The actual tumor resection often is aided by the use of ultrasonic aspiration, which provides a relatively atraumatic way to debulk many pediatric brain tumors. The surgical CO2 laser also may be used, depending on the consistency and location of the tumor. In general, tumors are resected “from the inside out.” With many extraaxial tumors and a small percentage of intraparenchymal lesions, a clearly defined peritumoral plane is encountered through which the tumor may be dissected carefully from the surrounding brain, cranial nerves, and vessels after the central portion of the mass has been debulked. However, for most intraparenchymal tumors, a well-defined tumor capsule is not present, and the resection must proceed via gradual internal debulking until the boundary between tumor and normal brain is reached.
Because the extent of resection is so important in defining prognosis and choice of subsequent therapy for many tumor types, objective confirmation of the volume of residual tumor, if any, is essential before embarking on further therapy. Because a surgeon’s impression of the extent of tumor resection is subject to error,252 postoperative confirmation of the extent of resection generally is established by CT or, preferably, by MRI. This imaging typically is performed within the first 24 to 72 hours postoperatively to minimize the impact of postoperative inflammation on the delineation of areas of residual tumor.
A recent trend in the surgical management of selected types of brain tumors has been the concept of second-look surgery. For large, relatively vascular tumors in which an initial complete resection cannot be obtained, the patient is treated with several courses of postoperative chemotherapy in the hope of making the tumor amenable to complete resection at a second procedure. This approach has been applied anecdotally in ependymomas and malignant germ cell tumors, two groups of lesions in which the extent of residual disease before initiation of radiation therapy has a substantial impact on long-term outcome. What remains to be determined is whether patients who undergo a second-stage complete resection have as good a prognosis as those who were amenable to complete resection initially; this issue is being examined systematically in studies of the Children’s Oncology Group (COG).
Surgical resection has been used increasingly as a component of the management of recurrent disease, particularly in children without evidence of tumor dissemination. For children with malignant lesions, this relieves mass effect in preparation for additional phase I or phase II chemotherapy. Some recurrent tumors, such as JPAs and craniopharyngiomas, can be treated with reoperation alone, without the need for additional adjuvant therapy, if a gross total resection (GTR) can be achieved.253 For children with recurrence of other, more

malignant tumors that may also be subject to dissemination, data are lacking for a survival benefit from re-resection.
Postoperative Considerations
The postoperative recovery of patients who undergo a complete or partial resection of a posterior fossa tumor, particularly a tumor in or around the cerebellar vermis, may be complicated by posterior fossa syndrome. Posterior fossa syndrome, also known as “cerebellar mutism,” may range from a mild to severe disorder that is characterized by mutism, ataxia, hemiparesis, cognitive impairment, behavioral changes, cranial nerve palsies, bulbar palsy, and tremor.254,255 The onset is typically within the first week after surgery255,256,257 and appears to occur more commonly with MB than with other posterior fossa tumors.258
The etiology of the posterior fossa syndrome is unknown but appears to be associated with edema in the brachium pontis or with decreased blood flow during surgery. Previously reported as a rare complication of posterior fossa surgery, for unclear reasons this syndrome is now observed in 10% to 25% of patients following tumor resection in the posterior fossa.258,259 Although recovery may be complete, particularly for the mutism, there are patients in whom it is incomplete with long-term neurologic sequelae.255,260,261
Radiation therapy is a central component of curative therapy for a majority of children with CNS tumors. The potential efficacy of irradiation in MB, ependymoma, craniopharyngioma, and many of the astroglial tumors has been apparent since the middle of the 20th century. During the last 25 years, the potential detrimental effects of irradiation in the developing and mature CNS have been quantified. Recognition of the unique CNS vulnerabilities in children and the concurrent demonstration of brain tumor responsiveness to chemotherapy introduced a paradigm of delaying or avoiding irradiation in children. With the introduction of sophisticated, three-dimensional image-guided radiation techniques capable of relative sparing of the normal brain structures, the focus of pediatric brain tumor trials has recently shifted to investigating broader indications for irradiation as a key element in achieving disease control. Current studies are addressing ways to optimize the risk-to-benefit ratio of accurate, limited-volume radiation delivery, sometimes in the setting of reduced radiation dose, as well.
The rational application of radiation therapy in pediatric brain tumors requires an understanding of the development of the brain, the probable biologic effect of ionizing radiation on the brain of a child, the behavior and natural history of the different brain tumors, radiobiology and physics, the techniques and technology of radiation therapy, and the probable interaction of irradiation with other treatment modalities, such as chemotherapy.
Indications for Radiation Therapy
The indications for radiation therapy depend on the histology of the tumor. Pretreatment histologic diagnosis is required except in cases of diffusely infiltrating pontine gliomas and visual pathway gliomas, both diagnosed on the basis of neuroimaging and neurologic findings. For ependymomas and many of the astroglial tumors, the use and timing of irradiation also depend on the anatomic site of involvement and the degree of resection. Specific indications for radiation therapy and controversies concerning its use are discussed later under the headings of the individual tumor types.
Radiation Volume
The radiation target volume is determined by the tumor histiotype, anatomic extent, and the known patterns of spread and failure. Advances in neuroimaging (particularly MRI, with current treatment planning often based on fusion of MRI and CT imaging and the use of fMRI and PET imaging currently under exploration)262,263 have made tumor localization more accurate.
Local target volumes are used for tumors that are typically confined to, a single anatomic location (e.g., ependymomas, craniopharyngiomas, and most astroglial tumors). Determining the target volume requires a complex integration of preoperative and postoperative/preirradiation imaging (accounting for the reconfiguration of the brain following surgery and, in clearly defined settings, response to chemotherapy) to identify the “tumor” or “tumor bed” as the gross target volume (GTV). Depending on the tumor type, a margin for potential microscopic infiltration (the clinical target volume or CTV) is identified by a three-dimensional expansion of the GTV by approximately 1 cm (e.g., discrete JPAs, craniopharyngioma, ependymomas), 1 to 2 cm (e.g., MB when boosting only the tumor bed region), or 2 cm (e.g., high-grade gliomas, diffuse infiltrating brainstem gliomas, or infiltrating WHO grade II astrocytomas).264 Finally, a volumetric expansion of the CTV (typically by 0.5 cm) defines the planning target volume (PTV), recognizing some variability in daily set-up or patient positioning despite considerable attention to immobilization and reproducibility (see Fig. 13.2).
Craniospinal irradiation is a technically demanding technique that provides homogeneous irradiation to the cranium and spine, targeting the entire subarachnoid space in tumors with known potential for (or established) intracranial and/or spinal leptomeningeal metastasis (e.g., MBs and AT/RTs).
Radiobiologic Considerations
The biology of radiation cell lethality for tumors and normal tissues is outlined in Chapter 13. There are differences in inherent tumor cell radiosensitivity, for instance MB cell lines show significantly greater cell death after 2-Gy exposure than glioblastoma multiforme cell lines.265 Fractionation (the principle of multiple, relatively small doses of irradiation protracted over time) is particularly important in the normal tissue tolerances seen in the brain, spinal cord, and several other critical organs. Most data has been derived from radiation effects following conventional fractionation (i.e., use of a single daily fraction of 1.8 to 2.2 Gy, typically on a 5 days per week schedule). Hyperfractionated delivery (specifically delivering two daily fraction of 100 to 120 cGy each to total doses up to 20% to 30% higher than “tolerance levels” defined following conventional fractionation) was purported to show

a significant benefit in childhood brainstem gliomas.266 The therapeutic ratio (i.e., risk-to-benefit ratio) was felt to be improved after high-dose hyperfractionated irradiation. The primary theoretical benefit of hyperfractionated radiation is the opportunity to escalate the total radiation dose without increasing damage to the normal brain. In theory, benefit from hyperfractionation should result because the antitumor effects from irradiation are primarily related to the total dose rather than the dose per fraction (dose rate), whereas the side effects of irradiation are highly correlated with dose per fraction (dose rate).267
Radiation effects on brain generally are characterized as delayed reactions, reflecting slow turnover times of normal brain parenchymal cells or indirect effects on the cerebrovasculature. Prospective data from a number of cooperative group and single institution experiences clearly show no apparent benefit in tumor response or outcome following hyperfractionated irradiation in brainstem gliomas.268 The only current clinical investigation using hyperfractionated delivery in a major multiinstitutional trial is the SIOP-4 study, further evaluating the role of hyperfractionated CSI in MB.
Techniques in Radiation Therapy
Conventional External-Beam Radiation Therapy
Conventional radiotherapeutic techniques are appropriate for pediatric brain tumors only when large volumes are treated (e.g., craniospinal or full cranial irradiation). Simple geometric field arrangements (typically two or three per radiation volume) with customized blocking provide relatively homogeneous irradiation within the defined target volume.
Three-Dimensional Conformal Radiation Therapy
The advances in neuroimaging and sophisticated three-dimensional computerized treatment planning systems have greatly improved the ability to target the tumor while significantly sparing the surrounding normal tissues. Three-dimensional conformal radiation therapy (3D-CRT) typically involves multiple, individually shaped (or collimated) fields, arranged in coplanar, nonaxial, and noncoplanar orientation delivered in either static or dynamic modes.269 Compared to conventional radiation therapy, 3D-CRT more accurately targets the chosen CTV while significantly reducing the volume of normal brain exposed to high-dose irradiation.270
Intensity-Modulated Radiation Therapy
Intensity-modulated radiation therapy (IMRT) is a more complex form of 3D-CRT, combining two advanced concepts with 3D-CRT: (a) inverse treatment planning (in which both the target volume and the adjacent or subtended normal tissues are assigned specific dose levels), allowing optimization of beam trajectories and weights in an overall plan and (b) computer-controlled intensity modulation of the radiation beam during treatment. IMRT allows a high degree of flexibility in reducing the dose to the surrounding normal tissues by the creation of so-called avoidance areas during the treatment planning process. Most IMRT programs prioritize dose avoidance over target volume homogeneity; the latter is a potential consideration in intrinsic brain tumors.271
Stereotactic Radiosurgery
Radiosurgery can be administered by a Gamma-Knife unit (based on a highly collimated dose array using 201 fixed cobalt-60 sources) or by a modified linear accelerator unit. Either unit is designed to deliver a high radiation dose to a small intracranial target in one fraction by focusing multiple small radiation beams from different directions to the target. The application of radiosurgery usually is limited to tumors measuring 3 cm or less in maximum diameter, ensuring a steep dose gradient at the treatment-field edge. Lesions within or immediately adjacent to critical structures (optic nerve/chiasm, brainstem) are approached only with some difficulty and risk. Radiosurgery commonly is used for the management of brain metastases, acoustic neuroma, and meningioma and is being investigated as a boost treatment for small, high-grade gliomas. In children and adolescents, radiosurgery has been used quite selectively in well-circumscribed, intrinsic LGAs (e.g., JPAs in the midbrain) not amenable to complete resection and in focal areas of residual craniopharyngiomas.270,272,273,274 Note that subacute reactions (intralesional necrosis, often associated with transient expansion of a space-occupying lesion) sometimes limit enthusiasm for lesions intrinsic to the brainstem.275,276
Particle Beam Irradiation
There is considerable current interest in the use of particle beams (most often protons) in radiation therapy for children with CNS tumors. Proton beam irradiation provides an advantage over conventional or photon irradiation in that the proton energy can be modulated to provide irradiation to a selected depth, nearly obviating exposure of underlying tissue (compared to photon beams that pass through the target region with gradually diminishing dose intensity). Current explorations focus primarily on MB, providing a potential advantage in dose distribution when targeting the primary tumor bed (especially with regard to cochlear and temporal lobe sparing) and the neuraxis (sparing exit from the spinal field, especially for thyroid and breast tissues).277,278
Brachytherapy, or interstitial irradiation, involves the implantation of radioactive sources directly into brain tumors. Iodine-125 and iridium-192 are radioisotopes that have been used for either permanent or temporary implants. Brachytherapy has been used for patients with glioblastoma as an additional boost after external-beam radiotherapy and for selected patients with recurrent high-grade glioma. Anecdotal favorable results have also been reported in patients with LGAs.279 The homogeneity and ready availability of 3D-CRT or IMRT have supplanted much of the enthusiasm for technically demanding implant procedures in CNS tumors. Intralesional brachytherapy (typically phosporus-32) has been utilized with some success in cystic neoplasms, most often craniopharyngioma cysts.280,281

Radiation and the Developing Brain
Brain development is most rapid during the first 3 years of life. Axonal growth and synaptogenesis are most active during the growth phase. The rate of growth and development slow down after 6 years of age. However, maturation of the brain, judged by degree of myelinization, is not complete until puberty.282 For infants and young children, white matter alterations after irradiation appear to mediate the functional and neurocognitive changes most concerning with regard to functional integrity.283 Younger children are more vulnerable to white matter changes (which may be localized within or adjacent to the high-dose PTV or may extend, presumably by axonal degeneration, along white matter tracts remote from the primary tumor and PTV).284
Radiation Effects on the Brain
CNS responses following irradiation are classically divided temporally as (a) acute reactions, occurring during treatment; (b) subacute or early delayed reactions, occurring a few weeks to 2 months after irradiation; and (c) late reactions, occurring several months to years after treatment.285 The pathogenesis of early and, more often, subacute radiation-induced brain injury includes inflammatory-like changes with associated intra- or perilesional edema; direct damage to oligodendrogliocytes resulting in inhibition of myelin synthesis and consequent white matter degeneration/loss; and damage to the vascular endothelium, resulting in areas of hypoxia or release of necrosis factors with attendant white matter necrosis.286,287,288 Immediate peritumoral edema or intralesional necrosis is uncommon with conventional fraction sizes (i.e., 100 to 300 cGy per fraction).
Subacute reactions include imaging and clinical findings that may mimic the primary tumor or results in constitutional symptoms (e.g., lassitude, low grade fever, less often alterations in recent memory; diffuse white matter changes identified as leukoencephalopathy); these reactions are typically time-limited, resolving within several weeks to a few months.
Late reactions are more clearly dose- and volume-dependent, occurring beyond 6 to 12 months after irradiation; these effects are typically permanent. Late reactions include focal radiation necrosis, a more diffuse pattern of radiation- and/or chemotherapy-induced leukoencephalopathy, neuropsychological effects, cerebrovascular effects, and secondary neoplasms (both benign and malignant). Late effects can be progressive, irreversible, and sometimes fatal. Late neuropsychological effects are particularly concerning in younger children and include intellectual impairment, memory deficits, and limited ability to acquire new knowledge.289 Impairment in cognition is most pronounced in children younger than age 4 to 7.290 Deterioration in IQ is more prevalent in children following whole-brain or “focal” supratentorial irradiation than after treatment confined to the posterior fossa.289
The presence and severity of radiation reactions depend on (a) irradiation treatment factors, including total dose, fraction size, interfractional interval, and treatment volume; (b) patient factors, such as age, presence of preexisting brain injury by tumor or surgery, infection, and vascular diseases; and (c) other treatment modalities, most commonly surgery and chemotherapy. The influence of certain factors, such as fraction size, treatment volume, and dose homogeneity, can be modified or optimized to limit the incidence and severity of brain injury.285
Radiosensitivity of Specific Structures in the Central Nervous System
In the modern radiotherapy era, incidental brainstem necrosis is quite rare; there are no data suggesting the brainstem is more sensitive to irradiation than other normal brain structures. Subacute radiation effects, such as those that occur in the management of diffusely infiltrating brainstem gliomas or large, focal intrinsic JPAs, can be problematic. For example, imaging changes can be difficult to differentiate from tumor progression and clinical signs may be quite pronounced, particularly focal intrinsic brainstem reactions.291 With 3D-CRT to “tolerance” levels of 54 to 60 Gy, one can see subacute white matter changes on MRI several months after irradiation; changes are often asymptomatic and transitory but may progress to frank necrosis.292,293
Spinal Cord
In view of its location, the tolerance of spinal cord to irradiation is a major dose-limiting factor in delivering high-dose irradiation to tumors of the head and neck region, the thorax, and the upper abdomen; cord tolerance is more problematic in general adult radiation oncology than in pediatrics, in which tolerance levels are often approached only in intrinsic spinal cord tumors or with sizable metastatic subarachnoid foci. In pediatric radiation oncology, the spinal cord is more often the targeted tumor volume than an unintended critical structure. Craniospinal irradiation is common, especially in the treatment of standard- and high-risk MB; dose levels to the entire spine rarely approach or exceed the 40-Gy level identified as “safe” in the United States using conventional fractionation.285,294 More limited spinal volumes are often treated to dose levels approximately 45 to 50 Gy or, occasionally, 54 Gy.285,295 With contemporary three-dimensional, image-guided irradiation for low-lying posterior fossa lesions (e.g., fourth ventricular ependymomas), subacute white matter lesions, as described previously for the brainstem, can occur in the cervicomedullary or upper cervical cord regions.292 Although most subacute effects are transitory (white matter on MRI, Lhermitte’s syndrome of a shock-like sensation radiating down the extremities associated with neck flexion), such changes can be associated with significant neurologic signs.
Frank postirradiation myelopathy can occur from 1 year to several years after treatment.295 The traditional dogma concerning the pathogenesis of radiation myelopathy rests on postmitotic cell death in the endothelial cells or oligodendrocytes (or both).285 Current concepts view radiation as producing cell death that, in turn, induces a complex pathophysiologic reaction in which the response of surviving cells may contribute to the impact of radiation on tissue integrity and functions. Cytokines, such as tumor necrosis factor and interleukin-6 (IL-6), appear to play important roles.296 Also,

some researchers suggest that the tolerance of the spinal cord is 5% to 10% lower in children than in adults.296 Large fraction sizes (≥250 cGy) are disproportionately associated with untoward biologic effect on the spinal cord. An apparent volume relationship suggests that the dose to the spinal cord should be reduced when the irradiated length is large.285 Results of recent primate studies indicated that an increase in treatment volume reduces the threshold and steepens the slope of the sigmoid dose-response curve for myelopathy.296
Cranial Nerves
Most cranial nerves are relatively resistant to radiation-induced damage. Two cranial nerves, the second (optic) and eighth (vestibulocochlear), are particularly worth mentioning in the radiation treatment of pediatric cancers. The optic nerve and visual pathway can be damaged during the delivery of therapeutic radiation to periorbital tumors (e.g., orbital rhabdomyosarcoma, optic glioma, paranasal and nasopharyngeal tumors) and suprasellar tumors (e.g., craniopharyngioma, pituitary adenoma, germ cell tumor, hypothalamic-chiasmatic glioma).297 The risk of radiation-induced optic neuropathy is related to total radiation dose, fraction size, and the irradiated volume. It was shown recently that no injuries were observed in 106 optic nerves that received a total dose of less than 59 Gy. The 15-year actuarial risk of optic neuropathy after a dose of greater than 60 Gy was 11% when treatment was administered in fraction sizes of less than 1.9 Gy, as compared with 47% when given in fraction sizes of greater than 1.9 Gy.298
The vestibular cochlear nerve and auditory apparatus must be considered in the delivery of high-dose radiation to posterior fossa tumors, such as MB, ependymoma, and astrocytoma. Although the incidence of early ototoxicity is typically related to cisplatin delivery (and potentially enhanced when administered with or after irradiation), radiation exposure to more than 50 to 55 Gy is associated with a small but finite incidence of late radiation-induced ototoxicity, presumably as a late effect on the vestibular nerves.299,300 The combination of irradiation and cis-platinum may be associated with a greater incidence of ototoxicity, highlighting the importance of prospective audiologic studies in children with CNS tumors receiving both therapies.301
The retina, a specialized neural end-organ supplied by an end-arterial system, is sensitive to vascular injury and has little ability for repair.294 It is sensitive to radiation as well. Deterioration of vision, resulting from radiation-induced progressive obliteration of small retinal vessels, can occur 1.5 to 6.0 years after irradiation. The dose-response curve is steep (with increasing incidence at dose levels between 50 and 60 Gy), and 45 Gy produces a 5% risk of visual injury within 5 years.294 Again, as the fraction size increases up to 2.5 Gy or more, the frequency of injury increases.302
The lens is one of the most radiosensitive organs, even to very low doses of radiation. For example, 1 Gy can lead to cataract formation. From total body irradiation data, the risk of developing a cataract requiring surgery was 20% for fractionated doses of 12 to 16 Gy.303 The dose that could produce a 5% risk of damage to the lens within 5 years is 10 Gy.
Hypothalamic-Pituitary Axis
Irradiation of the region of the hypothalamus and pituitary gland can result in significant neuroendocrine abnormalities and long-term sequelae. This is especially important in children. The hormones affected include growth hormone (GH), thyroid-stimulating hormone, adrenocorticotropic hormone (ACTH), and follicle-stimulating hormone–luteinizing hormone (FSH-LH). The largest volume of data concerns the effect of cranial irradiation on GH production and release. The irradiated anatomic site responsible for GH deficiency has been shown to be the hypothalamus.267 Impaired serum GH response with provocative testing is apparent in 60% to 80% of children who have survived brain tumors.267,304 A dose-response relationship is seen with a threshold of 18 to 25 Gy. The higher the dose of radiation, the earlier the GH deficiency occurs.
Deficiencies of other hypothalamic-pituitary hormones have also been described. The responsible irradiated site can be the hypothalamus, the pituitary, or both. Constine et al.305 have described non-GH abnormalities (thyroidal, gonadal, prolactin, and adrenal) in 20 children with brain tumors not involving the hypothalamic-pituitary region and treated with either cranial or craniospinal irradiation. In patients receiving only cranial irradiation, the hypothalamic-pituitary region is estimated to receive a mean dose of 53.6 Gy (40 to 70 Gy).305
The role of chemotherapy in the treatment of childhood brain tumors has become increasingly important over the past several decades particularly for some of the embryonal and low neoplasms. The specific indications for chemotherapy in childhood CNS tumors are described in the tumor-specific sections that follow.
Factors Influencing Drug Exposure in the Central Nervous System
The BBB and blood–CSF barrier are natural membrane barriers in the CNS that profoundly influence the penetration of most substances in the CNS. The BBB is located at the level of the endothelial lining of brain capillaries, whereas the blood–CSF barrier is located in the epithelium of the organs (e.g., choroid plexus, median eminence, and area postrema) that surround the ventricles. Metabolic enzymes and transporters such as p-glycoprotein (PGP), multidrug resistance-associated proteins (MRP1 and MRP3), and organic acid transporters (OAT) that influence drug transport are not present in normal endothelial cells but are present in the endothelial cells of brain capillaries and may also be present in the epithelial cells at the level of the blood–CSF barrier.306,307,308,309,310,311,312,313,314,315 Factors that influence CNS tumor penetration of an agent across the

BBB include the physiochemical properties of the agent, the degree of protein binding, and the affinity of the agent for carriers that facilitate transport of endogenous compounds into the CNS. Characteristics that negatively impact BBB penetration include poor lipid solubility, significant ionization, and high protein or tissue binding.316,317,318 The blood–tumor barrier is another important variable that may restrict the delivery of systemically administered chemotherapy to tumor tissue.310,313,315
These barriers to CNS drug delivery are not uniformly intact in brain tumors, as reflected by the intra- and intertumor variability in the degree and the amount of tumor enhancement following administration of water-soluble contrast agents such as gadolinium. Thus, the commonly held notion that water-soluble chemotherapeutic agents are unlikely to be useful in the treatment of CNS tumors is not entirely correct. In fact, a number of compounds, such as the classic alkylators and the platinum analogues, are of clinical value in these tumors. Thus, there is controversy regarding the magnitude of the role of the BBB in the resistance of CNS tumors to chemotherapy.
Administration of drugs that increase the systemic clearance of those chemotherapeutic agents that are substrates for cytochrome P-450 isoenzymes is another important variable that may negatively impact delivery of systemically administered agents to the CNS. Particularly relevant to patients with CNS tumors are the use of the enzyme-inducing anticonvulsants such as phenytoin or phenobarbital and the concomitant use of dexamethasone. For example, phenytoin has been shown to dramatically decrease the systemic exposure, and therefore equivalently decrease the CSF exposure, of agents such as topotecan and irinotecan.319,320 Corticosteroids may also decrease CNS drug exposure both through induction of CYP3A4-mediated drug clearance and reduction in the transcapillary transport of various compounds. Another clinical variable that may impact CNS drug exposure is concomitant radiation therapy, which at higher doses may enhance drug exposure through increased transcapillary transport.321,322,323 Increase drug exposure may result in enhanced cytotoxicity but also has the potential to increase neurotoxicity.
Drug Delivery Strategies
A variety of approaches have been employed to either disrupt or circumvent the BBB in an attempt to enhance drug delivery to the target tumor site(s) within the CNS. These approaches include (a) BBB disruption with osmotic agents such as mannitol324 or vasoactive compounds such as the bradykinin analog labradimil (Cereport, RMP-7),325 (b) administration of very high-dose systemic chemotherapy,326,327,328,329 (c) regional chemotherapy approaches (e.g., intrathecal therapy,330 intraarterial therapy,331 or intratumoral therapy using biodegradable polymers332 or convection-enhanced approaches), (d) radiation therapy to disrupt the BBB prior to administration of chemotherapy, and (e) inhibition of drug efflux from the CNS using inhibitors of transporters for PGP.
Blood–Brain Barrier Disruption
Osmotic opening of the BBB using infusions of hypertonic arabinose or mannitol can enhance the penetration of different agents into the CNS. Exposure of capillary endothelial cells to the hyperosmolar solution leads to cell shrinkage and stress on the tight junctions. This pulls the junctions apart, allowing increased capillary permeability.333,334 Although the effect is brief (generally reversible within 10 minutes), increases in CNS and CSF drug levels have been documented and correlated favorably with clinical responses in some but not all instances.335,336,337,338 Although BBB disruption is feasible, the actual effectiveness is uncertain. This approach has many drawbacks including the need for general anesthesia and intra-BBB catheterization. It may also be associated with profound and unpredictable side effects such as pulmonary embolus, stroke, visual loss, hearing loss, and seizures.337 In addition, because the effects are nonspecific, that is, not limited to the tumor, this approach may be associated with an increased potential for neurotoxicity. Pediatric experience with osmotic BBB disruption is very limited and cannot be recommended outside of a clinical trial setting.
A number of chemical agents that are derivatives of normal vasoactive compounds such as labradimal (Cereport, RMP-7), a synthetic analog of bradykinin, interleukin-2; leukotriene C4; and others have been investigated in the preclinical or clinical setting with regard to their ability to disrupt the BBB.339,340,341,342,343,344,345 Labradimil has a selective effect on increasing the permeability of the blood–tumor barrier to a variety of agents at doses lower than those required for disruption of the BBB, thereby avoiding the potential for unpredictable neurotoxicity associated with high drug concentrations in normal brain tissue.339,344,345 Results of preliminary studies indicated that labradimil plus carboplatin is well-tolerated in children with CNS tumors.346 The role of this strategy in children and adults with CNS tumors has not yet been defined.
High-Dose Systemic Therapy
Tumors of the CNS may fail to respond to standard-dose chemotherapy as a result of inherent or acquired drug resistance or because of limited and/or heterogeneous drug exposure in the tumor tissue. In an attempt to overpower these resistance mechanisms and maximize the therapeutic potential for agents with a steep-dose response cure, the use of high-dose chemotherapy with autologous bone marrow or peripheral blood stem cell rescue has been explored by a number of investigators. The nitrosoureas, which are extremely lipophilic, were among the first agents to be studied using this approach. However, substantial dose escalations were not feasible owing to unacceptable neurologic toxicity.347,348 Subsequent trials have used classic (cyclophosphamide, melphalan, and thiotepa) and nonclassic (carboplatin) alkylating agents, often combined with etoposide.
High-dose systemic chemotherapy approaches in children with CNS tumors have been most widely evaluated in infants, for whom postponement of radiation therapy is desirable because of its potential late neurologic toxicities, and in patients with recurrent tumors349,350,351,352,353,354,355 Feasibility has been clearly demonstrated and encouraging results have been seen in patients with MB and germ cell tumors, as well as in infants with embryonal tumors.356 Clinical trials wherein these regimens are compared prospectively to those using standard-dose chemotherapy have not been performed. The high-dose chemotherapy experience in children with gliomas, including

those of the brainstem, has been less promising.357,358,359 Patients who appear most likely to benefit from this approach are those with minimal disease at entry into myeloablative therapy, those whose tumors have shown response to standard-dose chemotherapy, and those with little prior exposure to chemotherapy. The optimal timing of this approach, either as a consolidative or a salvage treatment, remains uncertain, particularly as primary chemotherapy regimens are intensified. There are no prospectively established indications for this approach, which at the present time should continue to be limited to clinical trial settings.
Regional Chemotherapy
Intrathecal Chemotherapy
Delivery of drug directly into the intrathecal space, through either a lumbar puncture or a ventricular reservoir, is a form of regional chemotherapy designed to circumvent the limited penetration of most systemically administered agents across the blood-brain and blood-CSF barriers. This approach has been successfully utilized in the front-line treatment of childhood CNS leukemias and lymphomas. The primary advantage of intrathecal chemotherapy is that very high drug exposures can be attained in the CSF using a relatively small drug dose because the initial volume of distribution in the CSF is very small relative to that of plasma. As a result, total drug exposure is lower, minimizing the potential for systemic toxicity.360,361,362 A primary limitation to this approach is that drug penetration into the brain parenchyma or tumor is only a few millimeters, which limits the utility for patients with either parenchymal or bulky leptomeningeal disease.363 Another limitation to intrathecal therapy, particularly in patients with CNS tumors, include the fact that distribution throughout the neuraxis is uneven.364 This is particularly problematic for patients with hydrocephalus, ventriculoperitoneal shunts, or bulky leptomeningeal disease.365 Intrathecal therapy should not be administered to patients with abnormal CSF flow dynamics because of the potential for increased neurotoxicity and/or decreased efficacy.
Unfortunately, there are a limited number of anticancer drugs, specifically methotrexate, cytarabine, and hydrocortisone, that are specifically available for intrathecal use. Although these agents utilized are routinely used for hematologic malignancies, they have only modest or no activity against childhood CNS tumors. Thus, a role for intrathecal chemotherapy in the front-line treatment of CNS tumors with a predilection for leptomeningeal dissemination has not yet been defined.
The U.S. Pediatric Brain Tumor Consortium (PBTC) is currently evaluating the feasibility of intrathecal drug delivery in infants with newly diagnosed embryonal CNS tumors using intrathecal mafosfamide, a preactivated cyclophosphamide derivative, as one component of the front-line therapy of infants with newly diagnosed embryonal tumors. Topotecan and gemcitabine are other agents that are undergoing evaluation in phase I and phase II clinical trials.
Intratumoral Chemotherapy
Another strategy to enhance intratumoral delivery of therapeutic agents involves direct administration of the agent into the tumor bed. This approach has been extensively evaluated in adults with recurrent high-grade gliomas using microencapsulated, drug-loaded, biodegradable polymers that are implanted into the tumor tissue or tumor cavity at the time of surgery. The agent Gliadel, a polymeric “wafer” impregnated with BCNU (carmustine), passively diffuses from the polymer over a period of several weeks, thereby providing high drug concentrations to the tumor tissue or tumor bed while minimizing systemic drug exposure. This approach has resulted in a modest prolongation of survival in adults with recurrent (median increase from 23 weeks to 31 weeks) or newly diagnosed (median increase from 11.6 to 13.9 months) high-grade gliomas.332,366,367 Current trials in adults are evaluating the use of surgically implanted BCNU wafers administered in conjunction with systemic O6-benzylguanine, to counteract one of the primary mechanisms of BCNU-mediated drug resistance, specifically increased alkylguanyl alkyltransferase activity.368
An alternate approach to intratumoral therapy, which is better suited for the delivery of larger molecules, involves the direct infusion of a soluble agent into the tumor using an implanted catheter connected to an external infusion pump. High concentrations of a therapeutic agent are provided to the tumor and peritumoral brain tissue as a result of bulk flow of the agent through the interstitial spaces of the brain. Interstitial drug delivery (also known as convection-enhanced delivery or intracerebral clysis)369 has been anecdotally applied to the delivery of conventional chemotherapeutic agents, such as BCNU,370 or most recently in the delivery of mutated toxin genes conjugated to ligands that target receptors (e.g., EGFR, transferrin receptor, IL-13 receptor) expressed at levels higher within the brain tumor than in the surrounding normal brain.371,372,373 The toxin component of the conjugate, such as diphtheria toxin and Pseudomonas exotoxin, contains mutations within the domains necessary for cell internalization, which restricts toxin entry to those cells that express the receptor for the ligand conjugate.374 Because each of the receptors that has been targeted to date is expressed also on cells outside the CNS, the infusion approach to delivery minimizes the concentration of toxin that reaches other “vulnerable” cells while maximizing the amounts that are delivered to the tumor. Pediatric trials evaluating this interstitial drug deliver for children with recurrent or refractory high-grade gliomas are ongoing in the Pediatric Brain Tumor Consortium.
Interarterial Therapy
In theory, intraarterial delivery of chemotherapy for CNS tumors offers the potential for achieving higher drug concentrations in the tumor bed without a concomitant increase in systemic exposure and toxicity. The best candidates for this approach are agents that are rapidly cleared after systemic administration or that are metabolized or inactivated after their first pass through the liver.375 Although the use of intraarterial nitrosoureas and cisplatin has resulted in a modest number of clinical responses, there is not yet a demonstrated clinical benefit to routine intra-arterial drug delivery.348,376,377 A potential disadvantage to this approach is that drug penetration into normal brain tissue appears to increase as well, resulting in focal neurologic toxicity, particularly to the retina.378,379 Some of the other observed toxicities may have been due to

non-uniform mixing of drug and blood at the infusion site, resulting in “streaming” during arterial delivery and to exposure of the brain to alcohol-containing diluents.379,380
Novel Agents and Approaches
Differentiating Agents
The potential utility of differentiating agents, such as retinoic acid and inhibitors of histone deacetylase [phenylbutyrate and suberoylanilide hydroxamic acid (SAHA)], in the treatment of MB or high-grade gliomas, has been demonstrated preclinically.381 In vitro studies suggested in addition to inducing differentiation and suppressing tumor growth381,382,383,384,385,386,387,388,389,390 that these agents may also cause a direct apoptotic effect391,392 or, as in the case of retinoic acid, modulate autocrine growth loops through inhibition of the kinase activity of EGFR.393 This molecule often is amplified in adult high-grade gliomas and is possibly related to malignant transformation. To date, phase I trials of both trans- and cis-retinoic acid have been reported preliminarily in adult glioma patients; an objective response rate of approximately 12% and a stable disease rate of 12% to 35% have been reported after oral administration in small numbers of patients.394,395 Further evaluation of cis-retinoic acid in children with high-risk MB and recurrent malignant glioma is being conducted in COG studies. Current investigations with histone deacetylase inhibitors include a phase I studies of valproic acid, which has also been shown to enhance nuclear receptor activity through mitogen-activated protein (MAP) kinase activation396,397 and phase I trials of newer generation histone deacetylase inhibitors such as depsipeptide, SAHA, and PXD101.
Antiangiogenic Agents
Abnormal angiogenesis has been implicated in the development of many tumor types including brain tumors, and presents an attractive strategy for the targeted development of new anticancer agents. Among the anticancer agents that have been studied in children with brain tumors are SU5416, a small molecule inhibitor of vascular endothelial growth factor;398 TNP-470, a fumagillin analog, and thalidomide.399 A phase I trial of lenalidomide (CC5013), a potent analog of thalidomide, will soon be evaluated in children with CNS tumors. A comprehensive review of this strategy is found in Purow et al.400 Although antiangiogenesis represents an exciting area of new drug development, it is too early to know whether any of these antiangiogenic agents will ultimately have a place in the treatment of childhood brain tumors.
Small Molecule Inhibitors
There is increasing knowledge about the underlying molecular biology of pediatric CNS tumors, particularly with respect to aberrant signal transduction pathways. A number of small molecule inhibitors for abnormal signaling pathways are currently in various stages of preclinical and clinical development, particularly agents that inhibit receptor tyrosine kinase pathways. Some of the agents that are currently being evaluated in cooperative groups are briefly discussed.
Gleevec, which competitively inhibits the bcr-abl tyrosine kinase that results from the Philadelphia (9,22) chromosome translocation in chronic myelogenous leukemia,401 also inhibits platelet-derived growth factor receptor (PDGF-R), stem cell factor (SCF) receptor, and c-kit mediated signaling.402,403,404 Because aberrant PDGF-mediated signaling may play a role in brain tumors, particularly gliomas,405 trials with Gleevec were pursued in children with newly diagnosed brainstem gliomas. The PBTC is currently conducting a phase II of Gleevec, commencing approximately 4 weeks after completion of XRT, to determine if this agent can improve the progression-free survival (PFS) in children with newly diagnosed tumors.
ERBB is another receptor tyrosine kinase family that has been shown to play a role in critical cell cycle functions involved in proliferation, apoptosis, migration, survival, and differentiation.406,407,408,409,410,411 ERBB1 amplification and overexpression have been demonstrated in pediatric high-grade gliomas and brainstem gliomas.102,412 Expression of ERBB2 alone or in association with ERBB4 is associated with a worse prognosis in patients with MB.413,414,415 Likewise coexpression of ERB2 and ERB4 plus a high proliferative index as determined by the Ki-67 labeling index are associated with an aggressive tumor phenotype in pediatric ependymomas.414 Pediatric phase I and II clinical trials of a number of ERBB inhibitors are currently ongoing, including trials of gefitinib (ZD-1839), erlotinib (OSI-774), and lapatinib.
The CNS is a relatively immunologically privileged site. The brain lacks defined lymphatic drainage,416 the expression of major histocompatibility complex antigens is low,417 and the BBB limits the interaction of the peripheral host immune system and the brain.418 Nevertheless, the privilege is not absolute. For example, patterns of allogeneic and xenogeneic tissue transplant rejection from immunologically naïve419 and nonnaïve brains420 suggest that peripheral T-cell activity may be carried into the CNS.421
Immunotherapy of CNS tumors is based on the hypothesis that stimulation of the immune system, or blocking of the immunosuppressive effects of tumors, might enhance an antitumor response. Immunotherapy has been studied primarily in the preclinical setting but also more recently in adult phase I studies of patients with malignant gliomas.422,423 Pediatric data are in their very early stages. Strategies of immunotherapy are based on eliciting systemic antitumor immune responses that are carried into the CNS and on inducing a primary immune response in the brain itself. Adults with malignant gliomas are known to have, to some degree, altered immunity, owing to effects on T-cell proliferation, natural killer cell activity, and immunoglobulin production. The current thought is that these effects most likely are due to production of transforming growth factor b. These observations form the basis for another immunotherapy strategy, that of decreasing tumorigenicity of malignant gliomas by blocking the immunosuppressive effects of transforming growth factor b. A variety of approaches have been studied: administration of cytokines, such as interleukins and interferons; delivery of monoclonal antibodies; and the use of adoptive immunotherapy (i.e., the transfer of immune T lymphocytes). Gene transfer therapy for immunomodulatory genes is discussed later.

Interferons show cytostatic and cytotoxic effects on human glioma cell lines and xenografts.424,425,426 However, phase I and phase II clinical trials of interferon-α and interferon-β to boost systemic immune responses against intracranial tumors have yielded clinical responses in only a minority of patients.376,427,428,429 In addition, clinical trials of interferon-γ, which is a much more potent inducer of major histocompatibility complex class I and II antigen expression, have demonstrated unacceptable toxicities with little clinical benefit.430,431 Although the interferons appear to have anti-CNS tumor properties, their clinical potential has not yet been realized.
IL-2, a cytokine that can increase the antitumor activity of T cells and natural killer cells, has been used in the treatment of melanoma and renal cell carcinoma. Systemic and intratumoral administration of IL-2 to boost the antitumor cellular immune response of patients with malignant gliomas has resulted in significant neurotoxicity, primarily from cerebral edema, and in little clinical benefit.418 Adoptive immunotherapy using IL-2 and lymphokine-activated killer cells has resulted in inconsistent clinical and neurotoxic effects.432,433 IL-2 and tumor-infiltrating lymphocytes appear to have activity against CNS tumors in vitro and in extracerebral sites but not in the brain.434 In yet another adoptive approach, stable disease and prolonged survival, albeit with disease, were demonstrated in a recent adult phase I clinical trial of cytotoxic T-lymphocyte therapy for patients with primary or recurrent malignant gliomas.423 Taken together, these data indicate an in vitro antitumor potential of adoptive immunotherapy of CNS tumors that has yet to be fully realized in the clinical setting.
A potential role for monoclonal antibodies in the diagnosis and treatment of brain tumors also has been investigated. Systemically administered radionuclide-conjugated antibodies have prolonged survival in mice with human glioma xenografts, but human applications of these products have shown only limited efficacy. Although disrupted by tumor, the BBB may be sufficiently intact as to block penetration of large-molecular-weight antibodies.435 The application of these products currently is limited by tumor heterogeneity and rapid immune antibody clearance and, for radioconjugated antibody therapy, by dehalogenation and loss of radionuclides131 and excessive radiation to nontarget tissues.436 Intrathecal or intraventricular delivery of monoclonal products may bypass some of the limitations of systemic administration. Such studies are limited but promising.437,438
Gene Transfer Therapy
Gene transfer therapy, the process through which genetic material is transferred into cells for the purpose of eliciting a therapeutic response, is a new and innovative approach to the treatment of brain tumors and of other malignancies and disease processes (see Chapter 17). The gene of interest generally is transferred to the target brain or tumor cell using a virus-mediated delivery system. The postmitotic environment of the CNS tissue may offer an advantage over other tissues in that it may allow more specific targeting of the viral vectors to only mitotically active tumor. Genes that may be transferred may result in cell killing, either directly through cellular toxins or indirectly through the expression of drug-mediating enzymes.439 The therapeutic response of gene therapy can be through immunomodulation or antiangiogenesis as well.421,439,440,441
The herpes simplex virus thymidine kinase type 1 (HSV-Tk1) gene is a type of suicide gene that can be transferred to tumor cells. When exposed to systemically administered ganciclovir, the gene product causes phosphorylation of the drug that results in death not only of transfected tumor cells but of surrounding tumor cells.421,442,443 Through a limited institutional phase I study, Packer et al.444 recently demonstrated the feasibility and safety of HSV-Tk1 gene therapy in children with recurrent supratentorial malignant brain tumors. Significant toxicities associated with the injected vector or ganciclovir exposure occurred in 4 of 12 treated patients and included seizures, headache, lethargy, weakness, cerebral edema, and symptoms of increased ICP. All these symptoms resolved spontaneously or with a short course of glucocorticoid therapy.
Modulation of the immune response to brain tumors through the use of cells genetically modified for secretion of IL-2, interferon-γ, tumor necrosis factor, and IL-4 has been relatively well studied preclinically and with mixed results.441,445,446,447,448 IL-4 has emerged as a particularly potent antiglioma cytokine.418 In an animal model, the efficacy of an IL-4-transduced 9L glioma cell vaccine in eradicating tumor and prolonging survival has been shown.449,450 On the basis of these results, adult phase I trials have been initiated with favorable responses.451 Vaccine therapy using native or genetically modified dendritic cells for antigen presentation also is under investigation.450,451,452,453 Preliminary data indicate that such vaccines may be as potent as those based on cells modified for cytokine production.452
Collectively, these preclinical and early clinical data support a continued investigation of the potential role for gene therapy against CNS tumors. The majority of work has been done with adult gliomas. Pilot studies with high-risk pediatric brain tumors are currently in progress.
Clinical Trials Groups
The use of chemotherapy for CNS tumors now is commonplace and, for specific tumors, is considered the standard of care (as discussed in later sections). The COG, consisting of pediatric cancer programs from North America, Europe, and Australia, conducts numerous clinical trials of chemotherapy for nearly every brain tumor type and for children of all ages. New agents, or new schedules of established agents, are studied in clinical trials for the primary therapy of newly diagnosed disease or as treatment for recurrent disease. Extensive correlative biologic and genomic studies are being conducted by COG investigators in several tumor types, including MB and ependymoma. Similar national and international cooperative groups exist worldwide. In 1999, the National Cancer Institute also established the PBTC, a consortium of 10 member institutions that collectively diagnose disease and treat a large proportion of U.S. children with primary brain tumors. The objectives of the PBTC are to rapidly evaluate new therapeutic agents and treatment strategies for children with high-risk CNS tumors. In addition, the PBTC has a dedicated neuroimaging consortium to pilot new imaging techniques for CNS tumors. Results from PBTC studies are made available to the COG and other international cooperative groups for confirmatory testing in larger phase II and phase III studies.
Physicians and interested families can learn more about investigational chemotherapy protocols by contacting any of

these cooperative groups or by contacting foundations that support pediatric brain tumor research. Details are presented at the end of this chapter in the section on information on clinical trials.
Infants and very young children with CNS tumors are generally treated with protocols separate from older children with the same diagnoses because of the higher risks of treatment-related morbidity in the developing nervous system. In practice, the upper age of eligibility for studies in infants and very young children is generally set at 3 years but varies from study to study, ranging from between 1 and 5 years of age.454,455,456,457
Approximately one third of CNS tumors in children younger than age 15 years are diagnosed before the age of 5.458 In the most recent SEER registry data reflecting incidence of brain tumors from 1975 to 1995, the annual incidence rates for brain tumors are highest in the first 5 years of life and are fairly constant from year to year.2 Infant boys are affected more commonly than girls, but the ratio between them is dependent on both tumor type and age. For example, although MB occurs overall more commonly in boys, girls are affected more commonly in the first year of life, with a ratio to boys to girls of 1.27:1. During the first 2 to 3 years of life, brain tumors are more common in whites than in blacks.2,4 Survival rates are proportional to age. During the period 1986–1994, the 5-year relative survival rates for all CNS cancers were 45% for younger than 1 year old and 59% for 1 to 4 years old.4
Special Clinical and Pathologic Considerations
Astrocytomas (primarily LGAs) are the tumors that occur most commonly in infants, followed in incidence by PNETs (consisting primarily of MBs but including pineoblastomas and cerebral neuroblastomas as well), “other gliomas,” and ependymomas.2 These registry data generally are reflected in the accrual onto the infant brain tumor trials open to all malignant diagnoses; in them, MB and ependymoma are most common.454,456,459,460
Certain CNS tumors have a predilection for infants and young children. Among the LGAs, these tumors include optic chiasm and hypothalamic gliomas (discussed in the section on tumors of the optic pathway). The DIG, a mixed neuronal-glial tumor, frequently is mistaken for a more malignant process. Recognition of this entity is important because, despite its malignant appearance, complete resection alone appears to be curative and chemotherapy is not indicated.125,181 The DNET is another of the mixed tumors that occur more commonly in infants. Other less common low-grade tumors encountered primarily in infants and young children include the mature and immature intracranial teratomas.
Among the malignant CNS tumors, AT/RTs primarily occur in infants and younger children with a predominance of males. Historically, treatment regimens have varied, but survival is poor.118,461,462 Although the tumors are responsive to chemotherapy, the response is generally of short duration. Dissemination of disease along with local recurrence is the most common pattern of treatment failure, and event-free survival generally is less than 1 year.462
Clinical Presentation and Differential Diagnosis
Although the presentation of CNS tumors has been described earlier, there are several unique signs and symptoms of CNS tumors in infants and very young children. The evolving anatomy of the infant skull, together with the predilection for supratentorial sites in the first 1 to 2 years, influences the presentation of these tumors. The principal calvarial sutures do not fuse until approximately 6 months of age. As a result, tumors in infants may not become clinically evident until they have reached a large size. An increasing head size, with or without a bulging fontanel, is particularly characteristic of CNS tumors that occur during infancy. General signs of failure to thrive, such as poor feeding, emesis, and lethargy, are also common. These symptoms may be mistaken for the much more common anatomic problems, infections of the gastrointestinal tract or infections of the ears or sinuses. Loss of developmental milestones and seizures may also occur. A mass lesion can cause a hemiparesis manifested by early hand preference. Evaluation of these general symptoms, particularly when no obvious cause is found or when they persist despite other intervention should include imaging of the brain to rule out a CNS tumor.
Treatment and Prognostic Considerations
The principles of surgical management of brain tumors in infants are essentially the same as for similar tumors in older children. However, infant brain tumors pose a number of unique challenges that add to the difficulty and potential morbidity and mortality of surgical intervention.463,464,465,466,467 First, these lesions often are fairly large at presentation. Second, infant brain tumors are, as a group, more uniformly malignant than are those observed in older children and tend to be extremely invasive into the surrounding brain, increasing the chances of postoperative morbidity. Third, these lesions tend to be extremely vascular. This factor, coupled with the small blood volume of an infant, increases the risk of life-threatening hemorrhage during the course of the tumor resection. Fourth, because postoperative radiation therapy carries significant long-term sequelae in infants and young children, a greater onus is placed on surgeons to attempt as complete a resection as is safely feasible. The management of these tumors requires not only significant surgical expertise but also skilled anesthetic management, with attention to adequate replacement of blood products and clotting factors during the resection.
Radiation Therapy
In general, the survival rates from CNS tumors in infants and young children have historically been lower than those in their older counterparts.458,468 The reason for this finding

may be biologically based for certain tumors but probably is due in equal measure to the administration of lower doses of radiation therapy or the avoidance or deferral of irradiation because of concerns about long-term adverse sequelae.469,470 Craniospinal irradiation to young children is associated with significant neuropsychological, neurodevelopmental, and neuroendocrine deficits that are more profound with earlier age at irradiation.471,472,473,474,475,476 Recognition of these effects and their relation, in part, to radiation therapy led in the 1980s to investigational treatment approaches that relied on postsurgical chemotherapy to delay or obviate the use of primary radiation therapy. This strategy resulted in inadequate cure rates that were similar or inferior to those in historical reports for infants with MB and ependymomas.455,477 The majority of young children treated without irradiation showed progression within 6 to 12 months of surgery. Craniospinal irradiation provides control of disease in 25% to 40% of infants and young children with a progressive tumor following primary radiation therapy. The strategy of salvage irradiation (following postchemotherapy failure in contradistinction to planned consolidative therapy after a predetermined interval of chemotherapy) seems to be inappropriate as the required radiation doses and volumes for chemoresistant disease lead to unacceptable functional compromises.478
Because most instances of tumor progression or recurrence are at the local tumor site, current studies are examining the earlier use of conformal posterior fossa irradiation to improve local disease control. The rationale is supported by a pilot study from France indicating excellent long-term disease control in infants receiving intensive chemotherapy with local irradiation only for M0 MB following progression on a typical infant chemotherapy regimen.352 Current approaches for localized MB in both the COG and the PBTC incorporate early introduction of localized irradiation using response-adjusted doses based on the presence or absence residual disease on imaging studies. The use of reduced neuraxis dose irradiation has been reported in a small pilot series; disease control and functional outcome are not available for a significant cohort of children.479,480 Consideration of consolidation regimens incorporating only 18-Gy craniospinal irradiation have been proposed for children who may show a complete chemotherapy response of metastatic disease. For ependymomas, the international COG protocol uses local irradiation for children ≥12 months old, based on the earlier Pediatric Oncology Group (POG) trial and a more recent study at St. Jude.481,482
Published results of recent infant brain tumor trials using primary chemotherapy are listed in Table 27.6. Results of the first cooperative group study to use postoperative primary chemotherapy, POG 8633,455 showed that the most important prognostic factor for patients as a whole was the degree of tumor resection at diagnosis. The 5-year survival for all patients who had a GTR of their primary tumor, as determined by central review, was 62%; for those with less than complete resections, the rate was 31%. Similar results were seen in the two most common diagnoses; the 5-year survival rates after GTR and less than GTR for MB were 60% and 32%, respectively; and the corresponding survival rates for ependymoma were 66% and 25%, respectively. The presence of metastases did not independently predict outcome for either diagnosis. Except for ependymoma, younger age was not a prognostic factor. This and other smaller studies have demonstrated that, for children with MB and ependymoma, the use of primary chemotherapy after complete resection of disease, with either delayed or lower-dose radiation, results in survival rates that approach those seen in older children.
Alternatively, chemotherapy has been used after reduced-dose craniospinal irradiation (CSI). In a small pilot study of ten patients, craniospinal irradiation was given with vincristine at 18 Gy to the craniospinal axis and 50.4 to 55.8 Gy to the posterior fossa; this was followed by 48 weeks of vincristine, cisplatin, and lomustine (CCNU) therapy.479 Seven of the patients survived without recurrence at more than 6 years from diagnosis. The neurotoxicity with this approach was minimal; IQ scores 3 years after diagnosis were within 8 points of baseline scores.
In several noteworthy studies, a small minority of children, mostly with MB, apparently were cured with surgery and chemotherapy alone.454,460,477,483 The oldest of these series, which was initiated in 1976, was updated by Ater et al. in 1997.483 Children with MBs who were younger than 3 years were treated for up to 2 years with chemotherapy (mechlorethamine, vincristine, procarbazine, and prednisone); radiation therapy was reserved for recurrent disease. Of 12 patients with MB, eight survived at a median of more than 10 years, and six of these never received radiation therapy. Of five patients with ependymoma, two survived, one without having received radiation therapy. Irradiated survivors had a mean IQ of 85, whereas those who were not irradiated maintained a mean IQ of 100. These studies, and a report by Goldwein et al.,479 suggest that neurotoxicity in a subset of young patients with MB can be substantially reduced by lowering, or even omitting, the radiation therapy dose.
In both the POG and the Children’s Cancer Group (CCG) infant brain tumor studies, children with pineoblastoma did not survive, whereas those with nonpineoblastoma, nonmedulloblastoma PNETs had prolonged survivals.454,477 A similarly poor outcome has been seen with other primary chemotherapeutic strategies for sPNETs.457
More recent U.S. and European clinical trials have incorporated dose-intensified chemotherapy after surgery and have either restricted or withheld radiation therapy. The benefit of this approach is still under evaluation. The second POG “Baby-Brain” study (POG 9233) randomly assigned patients prospectively to standard-dose chemotherapy, as was given in the first such study, or to a dose-intensive schedule of the same agents. Initial results indicate that although the dose-intensive approach is feasible, it is more toxic than the standard schedule, and does not appear to provide an overall survival benefit for children with MB or ependymoma,484 although event-free survival was significantly higher on the intensive schedule for infants with ependymoma. Again, complete resection of ependymoma and MB was associated with significantly higher rates of event-free and overall survival. Mature data from POG9233, including those from the companion radiation therapy study, are eagerly awaited. (Dose intensification of chemotherapy with stem cell or marrow support was discussed in the section on drug delivery strategies.)


Currently, there is not a “standard” chemotherapy approach for infants with malignant CNS tumors although most regimens utilize multiagent therapy with cyclophosphamide, vincristine, cisplatin, and etoposide. The value of dose-intensified chemotherapy, as compared to standard doses, is not clear. The PBTC is currently studying the addition of intrathecal mafosfamide to a multiagent systemic chemotherapy regimen similar to regimens in earlier infant trials. It is hoped that intrathecal therapy may delay or eliminate the need for neuraxis radiation. Although the use of chemotherapy in standard doses for a limited time period before radiation therapy is a reasonable approach for infants and young children who are not eligible for clinical trials, it cannot be considered curative for the majority of patients. It appears that radiation therapy cannot yet be avoided. Thus, current studies are testing the hypotheses that earlier use of highly conformal radiation therapy may lead to increased local tumor control rates, and perhaps increased cure rates, while potentially decreasing long-term neurocognitive sequelae from decreased exposure of normal brain tissues to irradiation.
Current infant trials also are seeking to determine whether the benefit of total tumor resection observed in other studies can be expanded to a larger proportion of infants. Limited data suggest that neoadjuvant chemotherapy can render tumors more easily respectable.485 On the basis of these data and the results of earlier trials, ongoing studies (e.g., P9934, PBTC 001) have incorporated second-look operations for patients with residual tumor after initial surgery and several courses of chemotherapy.
New molecular technologies offer the hope that the genetic profiling of infant tumors will allow clinical oncologists to prospectively identify subsets of infants and young children who can be cured with surgery and chemotherapy alone versus those with higher-risk disease who may be candidates for innovative treatment approaches. It is anticipated that state-of-the art molecular genetic analyses will facilitate the development of more effective and less toxic treatment regimens. New treatment protocols within the COG are using molecular approaches, specifically FISH and INI1 mutation analysis,486 to distinguish infants with AT/RTs from those with PNETs to specify distinctive treatment approaches for these two tumor types.
MB, the most common malignant brain tumor of childhood, accounts for approximately 20% of all primary pediatric CNS tumors and for approximately 40% of tumors arising from the cerebellum. Although MBs may be diagnosed in teenagers or young adults, most cases occur in the first decade of life with a peak incidence between 5 and 7 years of age.487,488 Boys are affected between one and a half to two times as commonly as girls.1
Pathology and Patterns of Spread
The ongoing debate surrounding the nomenclature of MB and PNET was reviewed earlier. The most recent WHO classification of brain tumors maintains MB as an independent cerebellar embryonal neuroepithelial tumor and classifies separately other similar-appearing embryonal small-cell tumors in other locations.11
Classical MBs are highly cellular, soft, friable tumors composed of cells with deeply basophilic nuclei of variable size and shape, little discernible cytoplasm and, often, abundant mitoses (Fig. 27.15) although they may exhibit surprising histologic heterogeneity. Homer Wright rosettes and pseudorosettes are variably present. Various degrees of glial or neuronal differentiation are noted, suggesting that the primitive cell of origin possesses the capacity for bipotential differentiation.116,117 A histologic variant with an abundant stromal component, desmoplastic MB (Fig. 27.16) occurs dominantly in the lateral cerebellar areas of adolescents and adults130,489 and in the setting of Gorlin syndrome in infants and young children.19 These may have a nodular appearance. A second somewhat different histologic subtype contain nodules is the so-called cerebellar (or cerebral) neuroblastoma.
An aggressive variant of MB, termed large-cell and large-cell anaplastic, has been described.490,491 As the names imply, the histologic features that distinguish this subset of MB are large round nuclei with prominent nucleoli, frequent mitoses, abundant apoptosis, and, in the anaplastic subset, nuclear pleomorphism. These tumors typically express synaptophysin and chromogranin. Monosomy 22 has not been seen in the cases described, but they tend to be associated with MYC or MCYN amplification.492 The large-cell anaplastic variant represented 4% of the nearly 500 cases of MB reviewed by Brown et al.491
MBs often grow to several centimeters and may fill the posterior fossa, invading surrounding CNS structures as they occupy the regional subarachnoid and ventricular spaces. As with other CNS tumors of presumed primitive neuroepithelial origin, widespread seeding of the subarachnoid space may occur (Fig. 27.17). The reported frequency of CNS spread outside the area of the primary tumor at diagnosis is 11% to 43%, and such spread eventually occurs in as many as 93% of patients who come to necropsy.493,494
Of all pediatric CNS neoplasms, the MB has the greatest propensity for extraneural spread. Although this has been observed in 20% to 35% of patients in smaller institutional studies, more recent larger series suggest that the rate of such events is less than 4%.234,495,496,497,498 Bone and bone marrow are the most common extraneural sites, accounting for more than 80% of metastases; lymph nodes, liver, and lung are other reported sites.499
Prognostic Considerations
An analysis of prognostic factors is confounded by the rapid diagnostic and treatment modality changes that have occurred over the last two decades. Although there have been extensive retrospective assessments of biologic factors,500,501,502,503 they have not been prospectively validated. Some of the clinical and biologic characteristics of the disease that have, or have had, prognostic value in MB are discussed here.
Age at Diagnosis
Determining the influence of age as an independent prognostic factor is difficult because of several observations: younger

patients tend to have a higher rate of disease dissemination at diagnosis,494,498 a lower overall rate of complete tumor resection,504 and less aggressive radiation therapy.469 In addition, more aggressive histologic subtypes of MB491 and AT/RT, which in the past may have been misdiagnosed as MB, occur more commonly in the first few years of life. Nevertheless, children younger than 2 or 3 years have a worse prognosis than do older children. The age at which outcome is no longer negatively affected is not clear. In studies of patients older than 2 to 3 years, for whom radiation therapy generally is used as the primary postoperative treatment modality, younger age has been prognostic of outcome in some series237,505 but not in others.506,507
Figure 27.15 Typical histologic features of medulloblastoma (primitive neuroectodermal tumor). Tumor formed by apparently undifferentiated, basophilic, round to oval nuclei with minimal perceptible cytoplasm (hematoxylin and eosin, ×400).
Figure 27.16 Desmoplastic medulloblastoma showing linear arrangement of cells along delicate background fibers (hematoxylin and eosin, ×400).
Figure 27.17 A: Nodules of a medulloblastoma (primitive neuroectodermal tumor) in the cisterna magna (arrow). A hemorrhage is present in the medulla. B: Transverse section of the medulla and spinal cord showing metastatic deposits of medulloblastoma in the subarachnoid space. Tumor is partially hemorrhagic and has invaded the neural tissue to a variable extent at the different levels.
The Chang staging system, shown in Table 27.7, was developed in 1969 and used preoperative imaging studies, the surgeon’s intraoperative impressions, and CSF cytology to assign stages of primary (T stage) and metastatic (M stage) disease. Modifications of the system have been proposed, but, in any form, the Chang stage remains the system used most widely for designating the extent of disease.
Although M stage retains prognostic value (see further on), T stage does not in most recent large series.498,508,509,510 Brainstem involvement by direct infiltration of tumor, designated as Chang T3b, was historically indicative of a worse prognosis. However, with modern treatment regimens employing chemotherapy with radiation therapy, this feature is no longer prognostic.233,238,509,511,512
Extent of disease, or M stage, is consistently predictive of outcome, with two caveats. First, M1 disease, indicating positive CSF cytology without radiographic evidence of disseminated disease, is rare, and its impact on survival is unclear. In CCG study 921, wherein 18% of patients had M1 disease, 5-year PFS was 57% (± 10%), a rate lower than that for M0 disease and higher than that for M2 or M3 disease but not significantly different in either case.510 In the German HIT ’91 trial, 13% of patients had M1 disease, and their survival rate of 65% (± 12%) did not differ from that of patients with M0 disease.505 A similar rate of survival was found for the low number of M1 patients in the French M7 protocol for metastatic MB.507
Stage Definition
T1 Tumor <3 cm in diameter and limited to the midline position in the vermis, the roof of the fourth ventricle and, less frequently, to the cerebellar hemispheres
T2 Tumor >3 cm in diameter, further invading one adjacent structure or partially filling the fourth ventricle
T3 Divided into T3a and T3b
T3a Tumor invading two adjacent structures or completely filling the fourth ventricle with extension into the aqueduct of Sylvius, foramen of Magendie, or foramen of Luschka, thus producing marked internal hydrocephalus
T3b Tumor arising from the floor of the fourth ventricle or brainstem and filling the fourth ventricle
T4 Tumor further spreading through the aqueduct of Sylvius to involve the third ventricle or midbrain, or tumor extending to the upper cervical cord
M0 No evidence of gross subarachnoid or hematogenous metastasis
M1 Microscopic tumor cells found in cerebrospinal fluid
M2 Gross nodule seedings demonstrated in the cerebellar cerebral subarachnoid space or in the third or lateral ventricles
M3 Gross nodule seedings in the spinal subarachnoid space
M4 Extraneuraxial metastasis

Although M2 to M3 disease is prognostic in all series, ascertaining the independent impact of M4 disease, indicating extraneural metastases, is difficult. It is a much less common occurrence than is M1-3 disease, and, in most recent MB trials, patients with M4 disease were either excluded from, or not accrued to the study. As a result, M2 disease and M3 disease usually are combined in outcome analyses, often with M1 disease. In all series, a higher M stage of disease correlates with a lower survival rate.498,505,506,507,510
Accurate assessment of M stage requires an MRI myelogram and CSF cytology. The timing for both of these procedures relative to surgery is very important. Under ideal circumstances, the MRI myelogram with contrast should be obtained preoperatively to avoid confusion because postoperative changes or blood may cause transient changes that are difficult to differentiate from leptomeningeal disease. If preoperative spinal imaging is not possible, a minimum of 14 post days should elapse to minimize imaging changes related to the surgical procedure. Spinal imaging should also be performed prior to acquisition of CSF for cytology to avoid local imaging changes resulting from the lumbar puncture.
CSF for staging purposes should not be obtained until approximately 2 weeks following tumor resection. Preoperative lumbar puncture is contraindicated owing to the mass effect from the tumor and the potential for cerebellar herniation. Of note is the observation that sampling of CSF from different sites may lead to discordant results. Positive cytology from ventricular CSF, obtained through intraoperatively placed ventriculostomies or ventriculoperitoneal shunts, does not correlate consistently with cytology from the lumbar space, which is more sensitive in detecting malignant cells.511 Cytologic examination of CSF obtained by lumbar tap, obtained at least 2 weeks after definitive surgery, remains the standard method for determining CSF disease status for staging purposes.
Bone marrow aspirates and biopsies are reserved for patients in whom there is a clinical index of suspicion for extraneural disease. The positive yield for either of these studies is extremely low in asymptomatic patients.
Extent of Resection
A near-total resection (generally defined as a more than 90%) can be achieved in approximately 80% of MBs using contemporary microsurgical techniques.244 Evidence suggesting that extent of resection correlates with outcome has been provided by several single-institution studies and by a multiinstitutional study of the International Society of Paediatric Oncology.234,235,236,238,508,513 For example, Jenkin et al.513 reported a 5-year PFS of 93% in children undergoing GTR versus only 45% in those undergoing incomplete resection. These observations were in contrast to older multiinstitutional studies that failed to detect an effect of resection extent on outcome.498,514 In neither of these older studies was M stage assessed uniformly. However, more recently, the CCG study 921 demonstrated a clear relationship between extent of resection and outcome but only in the subset of patients who had no evidence of tumor dissemination.508 The same study found that, regardless of the surgeon’s estimate of disease resection, the presence of less than 1.5 cm2 residual disease on postoperative imaging was associated significantly with a higher PFS rate in patients with M0 disease; this effect was greatest in children older than 3 years of age.508
Although the foregoing studies support the performance of extensive surgical resections in children with MB, little convincing evidence substantiates improved outcome when gross total is compared with near-total resections. This distinction is important, because these tumors generally infiltrate the floor of the fourth ventricle. The observation that this microscopic residual disease does not adversely affect prognosis provides a strong rationale for avoiding aggressive removal of small components of brainstem disease, thereby minimizing the morbidity of the resection.
Although some reports have suggested that shunts placed to relieve hydrocephalus are associated with an increased incidence of systemic metastases,234,495,515 other studies have cast doubt on this association.493,516 A more recent review demonstrated fairly conclusively that, in patients with comparable risk factors, insertion of a shunt did not appear to increase the risk of systemic tumor dissemination.517 Ventriculoperitoneal shunts and alternatives to relieving hydrocephalus have already been discussed.
Biologic Markers
Extensive laboratory research efforts are directed at identifying molecular markers that will reliably predict the outcome

of children with MB. Early flow cytometry studies indicated an association between aneuploidy and a more favorable prognosis.518 Later cytogenetic studies further refined the link between genomic mutations and tumor behavior, demonstrating that deletions of chromosome 17p were associated with poor outcome.100 However, more recent studies demonstrated that loss of heterozygosity on 17p may not be a significant predictor of prognosis in MB.519 Amplification of c-myc was linked strongly to poor prognosis but was found to occur in fewer than 20% of all MBs.520
Several recent investigations have focused on gene expression as a marker of MB prognosis. Expression of the neurotrophin-3 receptor TrkC has been found to be associated with a favorable prognosis.500,501,502 The majority of MBs expresses both TrkC and neurotrophin-3, suggesting an autocrine or paracrine receptor activation of the TrkC receptor tyrosine kinase.500,502 Moreover, TrkC activation induces MB apoptosis that possibly may contribute to the more favorable prognosis of tumors with high TrkC expression.502 In contrast, a poorer prognosis has been linked to increased expression of the neuregulin receptors ErbB-2 and ErbB-4 and of c-myc or loss of caspase-8 protein expression.503,521 Erb2, in particular has been strongly linked to outcome and appears to provide prognostic information that supplements information available from clinical risk factors.503 Multigene expression profiling provides an even more powerful tool for identifying molecular prognostic features,142 which also appears to provide prognostic information that supplements, rather than duplicates information provided by clinical features.522 In view of the strong association between the above molecular factors and outcome, their evaluation has been incorporated into the new phase III prospective standard- and high-risk MB studies of the COG, ACNS0331 and ACNS0332, respectively.
These discoveries suggest that inherent biologic differences, reflected in the divergent expression of genes that regulate tumor growth and response to therapy, determine the clinical outcome of morphologically similar-appearing MBs. Efforts now are under way, through multiinstitutional therapeutic trials, to test prospectively molecular markers that might be used for the stratification of patients in future clinical investigations. Furthermore, the molecules identified by these investigations might serve as targets for future, biologically based therapies specifically designed on the basis of molecular mechanisms regulating tumor growth.
Risk Groups
Classification of patients into low-risk (also called favorable risk, standard-risk, or average-risk) and high-risk disease was initially based on the Chang staging system, which clearly defines metastatic disease (high-risk) and nonmetastatic disease (low-risk). Invasion of the brainstem as documented at the time of surgery, a high-risk feature in the Chang staging system, is no longer considered a poor prognostic factor. Most clinical trials in the United States stratify patients with localized disease and GTR or those with ≤1.5 cm2 of residual disease as average-risk and those with residual disease more than 1.5 cm2 and/or metastatic disease as high-risk (Table 27.8). Children who are younger than 3 years are also considered to be high risk (see the section on tumors in infants and young children). All these factors are clinical, and, thus far, biologic characteristics of disease, including histologic variants of MB, have not figured into risk classification, although their contribution is being assessed prospectively in the current COG and limited institution MB studies.
  Standard-Risk High-Risk
Age at diagnosis >3 years ≤3 years
Extent surgical resection <1.5 cm2 residual tumor >1.5 cm2 residual tumor
Extent of disease Stage M0 Stage ≥M1
Current therapy for MB in older children (generally older than 3 years) consists of maximal surgical resection, craniospinal radiation therapy, and chemotherapy.523 Current treatment challenges include a determination of the optimal dose of radiation therapy; the optimal timing, dose, and combination of chemotherapy agents; and the identification of prognostic factors to identify patients who require less intensive therapy.
A fundamental decision in evaluating a child with a posterior fossa tumor is determining whether the tumor is a mass lesion arising from the cerebellar hemisphere, vermis, or fourth ventricular floor or is an intrinsic tumor of the brainstem. Children with the former lesion types (e.g., MBs, ependymomas, cerebellar astrocytomas) undergo surgical intervention to allow for diagnosis and to achieve cytoreduction.
In children with a resectable lesion, such as a MB, the timing of surgery is determined by the clinical status of the children. If they are alert, high-dose corticosteroids are administered and, if feasible, a spine MRI scan is obtained to determine the presence of evidence of leptomeningeal dissemination; the craniotomy then is performed on the following day. This approach is reasonable also if such children are lethargic but become alert after administration of corticosteroids. However, in the rare situations in which such children are extremely somnolent, urgent surgical intervention is preferred.
A ventriculostomy is placed in most cases for temporary CSF diversion, and this is “weaned” by gradual elevation of the drip chamber during the postoperative period. If affected children need CSF drainage for more than 7 days after tumor removal, permanent CSF diversion usually is performed.
The tumor usually is approached through a suboccipital craniotomy or craniectomy, with the patient in a prone or modified prone (Concorde) position. The dura is opened in a Y-shaped fashion. Cerebellar hemispheric lesions, which are encountered most commonly in older children, are exposed fully by a transverse or vertical incision in the cerebellar hemisphere. The more common vermian and fourth ventricular

lesions may be visible within the foramen of Magendie but, if not, are exposed by dividing the caudal 1 to 2 cm of the inferior vermis. Because approximately 20% of patients develop a postoperative syndrome of pseudobulbar symptoms and mutism (the “posterior fossa syndrome”) after vermian tumor resections,524,525 which may be related in part to the extent of the vermis incision, an attempt is made to incise only as much vermis as is needed to provide adequate exposure of the tumor. The central portion of the lesion is then debulked using the ultrasonic aspirator. The plane between the tumor and the vermis is followed rostrally until the roof of the fourth ventricle is reached. Once the tumor has been partially de-bulked, a cottonoid patty is inserted under the caudal aspect of the tumor along the floor of the fourth ventricle, to minimize the risk of injury to the brainstem as additional tumor is removed from within the fourth ventricle. Some surgeons monitor electromyograms of the lateral rectus and facial muscles during the resection to reduce the risk of abducens and facial nerve palsies. Next, tumor extending laterally within the foramina of Luschka or the cerebellar peduncles is removed. Finally, tumor adherent to the floor of the fourth ventricle is carefully removed. The aggressiveness of the resection in this region is guided by the frozen-section diagnosis. For MBs, the tumor can be shaved down to the floor of the fourth ventricle, but removal of tumor below the floor is unnecessary, as it does not appear to improve outcome and clearly increases the potential for morbidity. This recommendation contrasts to the management approach for ependymoma (discussed further on).
Once the tumor has been removed, the dura is closed, generally with a graft. Because a CSF examination is an important component of the postoperative staging of children with MBs and because blood and debris can settle in the lumbar thecal sac within several hours of surgery, complicating interpretation of the cytology for several weeks, many groups either perform a lumbar puncture immediately after the tumor resection or delay the puncture until the third postoperative week. A postoperative MRI is performed, preferably within 24 to 72 hours after surgery, to confirm the extent of resection.
Radiation Therapy
MB is a radiosensitive, radiocurable CNS tumor; for decades, the standard therapy for MB has been postoperative craniospinal irradiation. Attempts at not irradiating the entire neuraxis or omitting radiation therapy altogether have resulted in compromised survival.526 Conversely, craniospinal irradiation is associated with intellectual, growth and endocrinologic morbidities, most marked in younger patients. Therefore, the emphasis of recent trials has been to identify cohorts of children in whom one can effectively use reduced craniospinal radiation dose levels. A prospective, POG-CCG study in which patients with favorable prognostic factors were randomly assigned to standard (36.0-Gy) versus reduced-dose (23.4-Gy) neuraxis irradiation showed an excess number of subarachnoid failures in the reduced-dose arm; late follow-up shows a marginal benefit to standard CSI, sufficient to indicate that regimens incorporating postoperative irradiation alone yet require 36 Gy CSI.527 Given the documentation that neuropsychological function was superior in children who received the lower CSI dose (at least for those ≤8 years),473 further experience in North America has built upon an encouraging, moderate-size, single-arm pilot study indicating disease control in excess of 80% following the same reduced-dose CSI (23.4 Gy/13 fractions) followed by local posterior fossa boost (54 Gy) and the now “standard” cisplatin, vincristine, and CCNU chemotherapy regimen.509 The more recent COG A9961 protocol used 23.4 Gy and randomly assigned patients to one of two different chemotherapeutic regimens, and preliminary reports indicate 3-year event-free survival results in excess of 80% for both treatment arms. The feasibility of further reduction in the CSI dose is being evaluated for children younger than 8 years in the newly opened ACNS0331 study.
The radiation dose to the posterior fossa tumor bed is less controversial. A total dose of less than 50 Gy has been shown to lead to inferior cure rates. The dose used most commonly is 54.0 to 55.8 Gy. The potential benefit of an additional radiosurgical boost in high-risk patients has not been demonstrated. Current considerations include whether irradiation of the entire posterior fossa or just the tumor bed (plus a 1 to 2 cm) margin should be targeted for the full dose of irradiation. The bulk of MB recurrences are in the posterior fossa adjacent to the original tumor site.528,529
Modern conformal techniques that deliver the full dose to the primary target region, reducing the radiation dose to the cochlea and hypothalamic-pituitary axis, to date appear to provide local disease control comparable to that achieve with full posterior fossa irradiation.528,529,530 The feasibility of systematic reduction in the posterior fossa boost volume is being examined as one component of the ACNS0331 study, in which patients will be randomized between image-guided, conformal irradiation to the whole posterior fossa versus only the tumor bed with a 1.5-cm margin. The technique of craniospinal irradiation is well known. However, ensuring at least a 5-mm margin of coverage of the cribriform plate is important, as inadequate irradiation of that area can lead to an increase in recurrence at this location.531 Spinal MRI is key to determining the inferior border of the thecal sac.
The potential benefit of chemotherapy for subsets of patients with MB was demonstrated first by the International Society of Paediatric Oncology and was subsequently confirmed in studies by the CCG and the POG.238,470,514 Since those reports, imaging and surgical technology have advanced significantly, and tumor assessment has become more standard. MB is one of the most chemotherapy-sensitive of all brain tumors, which has facilitated the investigation of chemotherapy in different contexts: to increase disease control and patient survival, to decrease adverse effects of radiation therapy, and to postpone or obviate the need for radiation therapy in very young children. Table 27.9 outlines some relatively recent studies that have utilized chemotherapy as an important treatment component for MB. The most commonly used agents are the classic alkylators, platinum analogues, etoposide, and vincristine. Toxicities limit the extent to which these agents can be used, particularly following neuraxis radiation.
Historically, chemotherapy has been evaluated in both the neoadjuvant (preirradiation) and adjuvant (postirradiation) settings. In patients with residual or metastatic disease, studies



of neoadjuvant chemotherapy offer the potential advantage of identifying active new agents, of delivering chemotherapy to a BBB that is possibly more advantageously disrupted owing to surgery, of treating micrometastatic disease throughout the CNS, and of decreasing gross residual tumor burden before the delivery of conventional radiation therapy. Chemotherapy before irradiation is better tolerated by the nonirradiated bone marrow as well as by the hearing apparatus. Several active drugs and drug combinations that have been identified through neoadjuvant chemotherapy trials include cyclophosphamide;532 cisplatin and vincristine;533 “8-in-1”;534 cisplatin and etoposide;535 cisplatin, vincristine, and cyclophosphamide;536 and carboplatin with etoposide.537 A recently completed European trial, PNET-3 Study, used a neoadjuvant chemotherapy approach and demonstrated the benefit of neoadjuvant chemotherapy versus radiation therapy only in M0/M1 patients.538 Unfortunately, the potential benefit of decreasing tumor burden before craniospinal radiation therapy, particularly in higher M-stage patients, is accompanied by a risk of disease progression before initiation of radiation therapy. Progressive disease rates of 20% to 30% have occurred during various neoadjuvant regimens.510,536,537,539 In addition, delay of radiation therapy may be detrimental to the outcome of certain patients, as suggested in a study by Kortmann.505
The sensitivity of MBs to chemotherapy has rendered possible the modulation of treatment modalities based on relative risk of recurrence in balance with potential long-term sequelae of treatment. In this context, chemotherapeutic trials generally have employed two risk strata for purposes of treatment: higher-risk disease, defined as that which is resected incompletely or is metastatic, and lower-risk disease, defined as completely resected and without metastases (Table 27.9). The goal of adding chemotherapy to low-risk disease is to facilitate a safe reduction in the radiation dose outside the posterior fossa to decrease the risk of late, adverse neurocognitive effects. Available data from the POG-CCG study of standard (36-Gy) radiation therapy versus reduced-dose (23.4-Gy) radiation therapy, without the addition of chemotherapy, for lower-risk MB patients indicate that neurotoxicity is indeed lower after reduced-dose radiation therapy.473 Whether the addition of chemotherapy to irradiation increases the survival of lower-risk patients is not clear. A limited institutional study by Packer et al.506 that used chemotherapy in addition to standard-dose radiation therapy demonstrated the benefit of this approach for patients considered at high risk for disease recurrence. Many patients in that study, including those with M0 tumors and with totally or near-totally resected tumors, subsequently were considered to have lower-risk disease. Although considered standard therapy, the apparent benefit of chemotherapy for low-risk patients has not been confirmed in larger controlled clinical trials.
In 1993, the CCG and the POG tried to evaluate the role of chemotherapy in lower-risk patients by performing a randomized study of standard radiation therapy versus reduced-dose radiation therapy plus chemotherapy. Unfortunately, the study was stopped early, because of poor patient accrual resulting from parental and physician bias in favor of chemotherapy.540 As noted earlier, a recently completed COG trial compared two cisplatin-based adjuvant regimens, substituting cyclophosphamide for CCNU in one arm, after reduced-dose (23.4 Gy) craniospinal irradiation and achieved outcome results that were at least comparable, if not superior, to those with standard dose radiation therapy alone (Packer et al., personal communication, 2004). Future trials will examine further reduction of the radiation therapy dose. Outside a clinical trial, the use of chemotherapy for lower-risk patients is generally considered standard of care. However, the lowest safe dose of radiation therapy, in combination with chemotherapy, has not yet been determined.
Chemotherapy studies for patients with higher-risk disease have focused on increasing tumor control and patient survival. In some series, the addition of chemotherapy to craniospinal irradiation has increased survival rates over those achieved with standard irradiation alone.238,498,541 Conversely, reduction of craniospinal radiation therapy dose, in conjunction with adjuvant multiagent chemotherapy, may result in lower survival rates.542 Ongoing studies for high-risk patients are exploring chemotherapy dose intensification as well as radiosensitization. The feasibility of administering several courses of high-dose chemotherapy with peripheral blood stem cell support after irradiation recently was demonstrated in a limited institution trial.356 A similar approach has been studied in the COG as well, although survival results are pending. Finally, the feasibility of radiosensitization with carboplatin, followed by adjuvant chemotherapy was demonstrated in the CCG-99701 study, and the efficacy of this approach will be assessed as a randomized question in the ACNS0332 study.
The effects of new agents on measurable disease and the addition of intrathecal chemotherapy also are being explored in patients with recurrent or progressive MB.
The term primitive neuroectodermal tumor was coined by Earle and Hart in 1973. The histologies of all these tumors are similar, but as noted previously, their nosology has been controversial. Various designations for these tumors have been based on the site of tumor origin and by lines of differentiation within the tumor. Thus, the sPNET also has been called cerebral or central neuroblastoma, cerebral MB, and pineoblastoma. Ependymoblastoma commonly has been included in the sPNET category as well. MB is the designation for PNET of the posterior fossa. In this section, only sPNET is reviewed.
Collectively, sPNETs in childhood are rare, accounting for only 2.5% to 6.6% of CNS tumors in most series.472,543,544,545,546 These tumors are more common during the early first decade.463 The median age of onset varies between several months and several years.544,547,548,549 Many series show a male-to-female ratio near unity; in others, boys predominate nearly 2:1.544,549 Whites appear to be affected more commonly than nonwhites.550
Pathology and Patterns of Spread
Despite the presence of considerable microscopic extension, sPNETs generally appear as well-circumscribed masses. Grossly, they are lobulated, soft, hemorrhagic, often cystic

masses. Microscopically, sheets of uniform embryonal-appearing, small, round cells with hyperchromatic oval nuclei and frequent mitoses are noted (Fig. 27.18). Homer Wright rosettes and perivascular pseudorosettes and areas of necrosis are common. Although foci with various degrees of glial, neuronal, or ependymal differentiation are seen in 70% of tumors, light microscopy and ultrastructural evaluation generally reveal most individual cells to be primitive and undifferentiated. They may contain a prominent connective tissue component or display the nodular pattern that has been called “neuroblastoma.”
Figure 27.18 A: Typical field of medulloblastoma (primitive neuroectodermal tumor of the posterior fossa). Tumor is formed by apparently undifferentiated, basophilic, round to oval nuclei with minimal perceptible cytoplasm. (Hematoxylin and eosin, ×400.) B: Typical field of pineoblastoma (primitive neuroectodermal tumor of the pineal). Note the similarity to the photomicrograph in A. (Hematoxylin and eosin, ×400.) C: One of several nests of malignant astrocytes in primitive neuroectodermal tumor of the cerebrum, displaying mild pleomorphism and several mitotic figures. (Hematoxylin and eosin, ×400.) D: Another similar field from the same tumor shown in C stained with glial fibrillary acidic protein (GFAP) showing a few cells staining for GFAP (GFAP, 3400).
Wide local and regional tumor extension is common, and transcallosal extension into the opposite hemisphere also has been reported.551 Although the incidence of diffuse leptomeningeal or spinal subarachnoid disease was as high as 30% at diagnosis in many early studies,547,552,553,554,555,556,557,558,559,560 reliable statistics for the incidence of metastatic disease are not available, owing to the relative rarity of these tumors and to changes in diagnostic imaging modalities and practice over the last few decades.561 Despite the initially low incidence of dissemination at diagnosis, the eventual occurrence of leptomeningeal spread may be as high as 70% at relapse. Systemic metastases have not been reported commonly, but they appear to favor bone and lung.
Prognostic Considerations
Historically the treatment of sPNETs has been similar to that for MB. However, the overall outcome for children with sPNETs is worse than for those with MBs,562 suggesting that there are inherent biologic differences in these disease entities. Factors that may affect the outcome for children with sPNETs that have emerged from various series include the tumor site, stage, and extent of surgical resection. In the CCG 921 study of radiation therapy and two different chemotherapeutic regimens, the absence of metastases and pineal location predicted a higher chance of survival.550 The PFS for nonmetastatic tumors, when adjusted for site, was 50%, whereas the PFS for patients with M+ disease was 0%.550 Children with pineal tumors had a 3-year PFS rate of 61% versus 33% for tumors of nonpineal origin. In contrast, in infant series, the prognosis for patients with sPNETs in the pineal location is dismal.477,504 In some series, patients who had GTRs or no evidence of residual disease as measured by postoperative MRI had a better outcome than patients with incomplete resections; however,233,548 other studies have failed to identify such an association.477,551,558,560 Given the small numbers of patients with primary pineal sPNET the benefit of tumor resection conclusively is difficult.561
Gross total and near-total tumor resection are less frequent for sPNETs than for MBs because sPNETs are usually large, fairly vascular tumors that invade functionally important cortex.233 Accordingly, the surgical management for sPNETs is similar to that for other cerebral malignant tumors, such as high-grade glioma, in which the primary goal is to relieve local mass effect by removing as much tumor as is safely feasible.
Radiation Therapy
Standard therapy for patients with sPNET includes craniospinal irradiation. Current treatment recommendations are craniospinal irradiation to 36 Gy with a boost to 54 Gy to the initial tumor. Radiation therapy alone, as the only postoperative treatment, has been reported in only two retrospective studies; only 1 of 22 patients in the combined studies survived beyond 2 years.563,564 Most treatment series have used both irradiation and adjuvant chemotherapy. Despite the improvements in survival rates noted with combined-modality therapy, PFS is no better than 33%.233,543,550,551 Preliminary data from a four-institution trial may indicate superior outcome using aggressive surgery, reduced neuraxis-dose irradiation (23.4 Gy CSI for

M0 disease) with a local, targeted primary tumor bed boost, followed by four cycles of dose-intensive chemotherapy.356
Figure 27.19 Left: Axial T2-weighted imaging through the posterior fossa demonstrating a T2 hypointense right cerebello-pontine angle mass deforming the brachium pontis with associated surrounding edema (arrow). Center: Apparent diffusion coefficient map with evidence of restricted diffusion returned from the tumor (arrow). Right: T1 magnetization transfer contrast (MTC), which demonstrates moderate enhancement following Gd administration (arrow). This was proven to be an atypical teratoid-rhabdoid tumor, which has similar imaging characteristics to a medulloblastoma.
Various single-agent and combination chemotherapeutic regimens have shown activity against sPNET.477,535,554,565,566 In recent series, the outcome for patients with sPNET appears to be inferior to that for MB.550,562 In the CCG 921 study in which neoadjuvant and adjuvant 8-in-1 chemotherapy was compared to adjuvant CCNU, prednisone, and vincristine, each given with radiation therapy, 3-year PFS and survival were 45% and 57%, respectively, for all patients with sPNET and did not differ significantly between the two study arms.550 Unlike the experience with MB, neither degree of surgical resection nor extent of residual disease was prognostic, although a strong trend was apparent in the CCG-921 study.233 The authors concluded that the data were insufficient to determine whether the use of chemotherapy increased survival from that historically reported with surgery and radiation therapy alone.551 Ongoing clinical trials offer combined-modality therapy to infants and older children with sPNET, the latter of whom are being treated along with high-risk MBs in current COG studies. Identification of either clinical or biologic prognostic factors is needed to redirect therapy in the future.
AT/RTs are extremely rare tumors for which an actual incidence is unknown. Data from an AT/RT registry reveals that the median age at diagnosis is 24 months.462 Approximately 60% of CNS AT/RTs are supratentorial and the remaining 40% are infratentorial; approximately 15% of occurring in the cerebellopontine angle (Fig. 27.19).118,462 Leptomeningeal dissemination is present in 20% to 25% of patients at diagnosis. There appears to be a predominance of affected males with a male-to-female ratio of 2:1.462
Pathology and Patterns of Spread
Definitive diagnosis of AT/RT is based on histologic and cytogenetic features. Although these tumors exhibit considerable morphologic variability, the most common findings consist of a combination of rhabdoid cells and areas resembling primitive neuroectodermal tumors with little or no evidence of differentiation. However, in some instances Homer Wright, Flexner-Wintersteiner, and ependymal rosettes or primitive neural tube-like structures are present. Some tumors are composed entirely of rhabdoid cells, whereas other features that may be found in these neoplasms consist of epithelium of various types and/or mesenchymal tissue.118 Typically, the rhabdoid cells express epithelial membrane antigen (EMA) and vimentin, and less often smooth muscle actin (SMA). Remarkably, they may also express NFP, GFAP, and keratin. An antibody for the INI1 gene is proving useful. Cytogenetic studies have demonstrated deletion of chromosome 22 and alterations of the hSNF5/INI1 gene in these tumors. An antibody raised to INI1 and used with immunohistochemical techniques shows lack of expression as a characteristic feature of AT/RT.486,567
Prognostic Considerations
This is an extremely aggressive tumor with a median event-free survival of 10 months and a median survival of 16.8 months from diagnosis. As a result of the nonuniform approach to treatment for this disease it is impossible to definitely establish prognostic factors that impact survival. Early experience suggests that there is a uniformly poor outcome with conventional chemotherapy.462 Registry data suggests that the event-free survival may be better for patients with a gross total surgical resection versus those with a partial resection (14 months vs. 9.3 months).462
The optimal treatment for AT/RTs is not known. Although chemotherapy appears to provide short-term benefit, it is generally not curative. The tumor is responsive to irradiation; the superior outcome in older children appears to reflect early use of this modality in children older than 3 years old.568 The COG has developed the first prospective cooperative group trial to prospectively evaluate a multimodality approach for AT/RTs.
The surgical techniques for tumor resection should be based on the tumor location and are similar to those employed for sPNET or MB.

Radiation Therapy
The role of radiation therapy is uncertain because the majority of patients with AT/RTs are infants or young children in whom radiotherapy has historically been omitted or delayed. Participants in a National Cancer Institute AT/RT workshop recommended that early use of involved-field radiotherapy be included as a component of therapy in the COG protocol that is under development. This recommendation was based on data from patients with AT/RT who are currently alive without evidence of disease. Registry data reveals that approximately 60% of these patients, primarily those who were older (median age, 47 months) than the entire cohort (median age, 24 months) received some form of radiotherapy.462 A single-institution study indicates survival in eight of nine children/adolescents following craniospinal irradiation and chemotherapy.568
AT/RTs appear to be chemosensitive tumors with retrospective series reporting complete or partial responses in more than 50% of children with incomplete resections.462 A wide variety of chemotherapy regimens have been used, most commonly alkylators and platinum based combinations. Unfortunately, for most patients multiagent chemotherapy with or without radiation therapy is not curative.462,569 The optimal chemotherapy and treatment regimen for AT/RTs remains unknown. In addition to the proposed COG study, there are several limited institution clinical trials that are addressing the role of high-dose chemotherapy with stem cell rescue for children with AT/RT.462
According to recent SEER registry data, ependymomas constitute approximately 9% of all primary CNS tumors in children.1 The tumors usually arise within or adjacent to the ependymal lining of the ventricular system or the central canal of the spinal cord. Ninety percent of the tumors are intracranial, and up to two thirds of these occur in the posterior fossa.226,570,571 In children younger than 3 years more than 85% of tumors may occur in the posterior fossa.482 The highest incidence of ependymoma in children occurs in the first 7 years of life,1 with a second peak in the third to fifth decades of life.572 Whereas in the past, the ratio of male to female patients was reported to be near unity, recent series report a male-to-female ratio of between 1.3 and 2.0.1,224,228,573,574 Of note in the potential etiology of these tumors is that DNA sequences identical to the SV40 polyomavirus and the corresponding viral large T-antigen have been found in several tumors; SV40 and related polyomaviruses can induce ependymoma and choroid plexus tumors in monkeys and other mammalian species.575
Pathology and Patterns of Spread
Ependymomas are generally well-demarcated tumors that often display areas of calcification, hemorrhage, and cysts. Although uncommon, the ependymal rosette is a characteristic and diagnostic microscopic feature composed of radially aligned ependymal elements about a central lumen. More common is the conspicuous pseudorosette (Fig. 27.20), an eosinophilic halo composed of cells with tapering processes surrounding a blood vessel.
Figure 27.20 Section from a fourth ventricular ependymoma displaying typical perivascular pseudorosettes (hematoxylin and eosin, 3400).
Ependymomas vary from well-differentiated tumors with no anaplasia and little polymorphism to highly cellular lesions with significant mitotic activity, anaplasia, and necrosis that may resemble glioblastoma multiforme. Ependymomas have been classified descriptively as cellular, epithelial, or papillary. None of these terms or a proposed classification that grades these tumors on the basis of their degree of anaplasia has gained wide acceptance, although they often appear in older literature. An exception to the use of such terminology exists for the ependymal tumors arising in the region of the conus medullaris and filum terminale, termed myxopapillary tumors because of their unique histologic features. Current terminology for ependymomas located elsewhere in the CNS distinguishes between benign (low-grade) tumors and malignant (high-grade or anaplastic) tumors. The WHO classification system uses the terms ependymoma and anaplastic ependymoma to distinguish these two types. Although the impact of histology on disease behavior and outcome has been debated for 2 decades, more recent series do suggest a significant correlation between anaplastic histology and a higher rate of disease recurrence.228,489574,576,577,578
Ependymomas are locally invasive tumors that spread contiguously into adjacent brain. Tumors arising in the posterior fossa frequently infiltrate the brainstem. In as many as one third of these cases, tumor may project through the foramina of Luschka and/or Magendie to involve the cerebellopontine angle and upper spinal canal.574 The incidence of spinal subarachnoid dissemination has been estimated to be 7% to 12% in combined patient series. One meta-analysis suggests that such events are potentially most common in high-grade and posterior fossa tumors.572,579 Systemic metastases are rare and, when present, show a predilection for liver, lung, and bone.

Prognostic Considerations
The single most important prognostic factor that emerges from review of single- and multiinstitutional experience with ependymoma is the extent of tumor resection.222,223,224,225,226,227,228,455,570,579,580,781 Whether gauged by the surgeon’s estimate or measured by postoperative MRI, the survival rate is higher following a gross total (66% to 75%) versus a less complete resection (0% to 11%).222,223,224,225,226,227,228,582 Perhaps related to degree of resection has been the finding in some series that location of primary intracranial tumor and patient age are prognostic.583,584 However, these data are inconsistent with the exception of ependymomas of the spinal cord, which are associated with the best outcome. Younger children are more likely to have tumors arising from the posterior fossa, and in this location, tumors tend to be more invasive, making a GTR more difficult. However, when lower age (<2 to 4 years) has been found to have a negative impact on survival, it appears to be confounded by the fact that lower radiations doses were typically administered to these younger patients. Finally, histologic subtype, particularly anaplastic ependymoma, has inconsistently been associated with a worse prognosis versus that for patients with a well-differentiated ependymoma. Because this factor has seemed to be an important prognostic variable in larger studies, especially those with consistent central neuropathology review, it is being examined prospectively as a therapeutic stratification variable in the ongoing COG ACNS0121 study.
Interestingly, preliminary studies using CGH suggest that cytogenetic analyses of ependymomas may help in the classification of ependymomas and provide leads concerning their initiation and progression.585 The relationship of these aberrations to patient outcome will also be assessed in the COG ACNS0121 study.
Techniques for the resection of posterior fossa ependymomas are similar to those used for resecting MB, although the rationale for intraoperative monitoring of evoked potentials and cranial electromyography may be even greater because of the higher frequency of brainstem infiltration. Supratentorial ependymomas, often located subcortically, are resected in a fashion similar to that used in other deep-seated gliomas (described further on).
The prognostic benefit of complete tumor resection has been stated. However, such a result has been feasible in only approximately 50% to 66% of ependymomas in most series. GTRs generally are more difficult for posterior fossa ependymomas than for supratentorial ependymomas because of the propensity for infratentorial lesions to infiltrate the brainstem and to surround cranial nerves and vessels lateral and ventral to the brainstem, precluding complete removal without unacceptable neurologic morbidity. Infants are particularly likely to have large infratentorial ependymomas with significant ventrolateral extension, which in part accounts for their less favorable prognosis in most series.481,571,507586,587,588 In such patients, multiple lower cranial nerve palsies as a result of both tumor and surgery often necessitate a tracheostomy and gastric feeding device. Resolution (if any) of neurologic impairment may be delayed for several weeks to months.587,589 Because the prognosis in children with incompletely resected ependymomas is so poor, several recent treatment protocols have selectively incorporated second-look surgery in children with objective evidence of residual disease after an initial procedure. This generally has been attempted after a short course of neoadjuvant chemotherapy, administered in the hope of reducing the vascularity and invasiveness of the residual disease. A multiinstitutional study, designed to examine the ability of this approach to improve outcome in such children without an unacceptable trade-off of morbidity, is currently being conducted in the COG, ACNS0121.
Radiation Therapy
Local postoperative radiation therapy has increased the overall survival rates of patients with ependymoma from 15% to 25% to 35% to 63%. The indications for radiation therapy are strongly supported in the literature; only amongst differentiated supratentorial ependymomas is there suggestive data indicating GTR need not require postoperative irradiation.581,590
Posterior fossa tumors are frequently intertwined with cranial nerves or adherent to the pontomedullary region and/or the cerebellopontine angle. Thus resection is typically followed by local irradiation to include the tumor bed. In addition, attention to potential extension into the foramina of Luschka or below the foramen magnum along the cervical spinal cord is critical in targeting ependymomas.
Longstanding debate regarding the appropriate volume for radiation therapy has shifted from identifying cases that might require craniospinal irradiation to diminishing the target volume from cranial compartments to the tumor/operative bed.222,574,591 An image-guided approach using 3D-CRT [with narrow (1 cm) margins around the tumor/operative bed defining the CTV] to 59.6 Gy has shown excellent tolerance and disease control.482
In intramedullary spinal cord or cauda equina ependymomas, postoperative radiation therapy may be withheld if a microscopically complete surgical excision is documented.592,593
Single- and multiagent chemotherapeutic regimens have been used in ependymoma therapy. In clinical trials, the platinum analogues appear to have the greatest activity.594,595 (For a comprehensive review of chemotherapeutic experience with ependymoma, the reader is referred to a review by Bouffet and Foreman.596 The use of chemotherapy for infants with ependymoma is discussed earlier.) Despite the demonstrated activity of various multiagent regimens, the use of chemotherapy has not improved the overall survival for patients with either completely or incompletely resected ependymoma.597 Thus, for older children, chemotherapy is recommended only as part of a clinical trial. The COG ACNS0121 trial will evaluate whether two cycles of multiagent chemotherapy in patients with subtotal resections will make their tumors more amenable to a GTR prior to the initiation of conformal radiation therapy.

Low-grade glial neoplasms are a diverse group of tumors that include JPA, fibrillary (also called protoplasmic or diffuse) astrocytoma, oligodendroglioma, ganglioglioma, and such mixed tumors as oligoastrocytoma. Their unifying features are their generally slowly evolving, clinical behavior and relatively benign histologic appearance. Generally high rates of long-term survival are characteristic as well, despite low but steady rates of disease progression even 10 years from diagnosis.598
Cerebellar astrocytomas are the most prevalent, representing 15% to 25% of all CNS tumors, followed in prevalence by cerebral hemispheric astrocytomas and tumors of deep midline structures (each representing 10% to 15% of all CNS tumors) and tumors of the optic pathway (accounting for approximately 5% of all CNS tumors).599 Seventy percent to 75% of cerebellar astrocytomas occur in childhood,600,601 most commonly in the first 2 decades of life. The average age at diagnosis ranges from 6.5 to 9.0 years.174,601,602,603,604,605 Boys are affected more commonly than are girls.606 Neuraxis dissemination of LGAs from any location in the brain is distinctly uncommon, occurring in only approximately 5% of cases.607,608,609,610 Tumors arising from the hypothalamus and periventricular areas may be more likely to disseminate.
Pathology and Patterns of Spread
Classifications, such as that of Kernohan, St. Anne/Mayo, and WHO, identify low-grade tumors primarily on the basis of their grade or degree of anaplasia rather than on histologic type.108 Neoplasms that are only modestly cellular and contain few or none of the histologic criteria of malignancy are designated as low-grade or grade I and grade II lesions in these classifications. The WHO classification uses the grade I designation for the typical pilocytic astrocytoma and grade I or grade II for the mixed neuronal-glial tumors. The low-grade fibrillary astrocytomas that make up most adult low-grade lesions are designated as grade II. Grade III and grade IV tumors are high-grade lesions characterized by aggressive clinical behavior and malignant histology (considered separately in the section on supratentorial high-grade gliomas). Although the utility of such grading systems has been questioned because of their subjective nature and their reliance on often small biopsies from tumors that may be heterogeneous, these systems remain popular because applying and understanding them is simple and because they have some prognostic value.
Supratentorial Low-Grade Gliomas
Most supratentorial LGAs are astrocytomas, a diverse group of neoplasms generally composed of GFAP-positive bipolar or stellate cells. Such designations as fibrillary, protoplasmic, gemistocytic, xanthomatous, and pilocytic often are used to describe the appearance of the astrocytes and their various histologic patterns.108,611 Only the pilocytic and fibrillary varieties are seen commonly in children. Pilomyxoid astrocytomas (PMAs), are another type of LGA that appear to previously been diagnosed as pilocytic astrocytomas. In contrast to the typical biphasic pattern of pilocytic astrocytomas, PMAs demonstrate monomorphous piloid cells in a loose fibrillary and myxoid background and do not display Rosenthal fibers, which are characteristic of pilocytic astrocytomas.612,613 PMAs appear to have a more aggressive course than typical pilocytic astrocytomas and typically occur as hypothalamic/chiasmatic tumors in a younger children.612 Anaplastic transformation of LGAs to more malignant-appearing and clinically aggressive entities, such as anaplastic astrocytoma and glioblastoma multiforme, is a common event in adults but occurs less frequently in children and young adults.489,590,614,615,616
Oligodendrogliomas, a separate category of glial neoplasms, are characterized by a generally monotonous collection of uniform spheroidal cells with more homogenous nuclei than are seen in the fibrillary astrocytoma, the tumor that represents the principal diagnostic alternative. An abundant clear cytoplasm surrounding a dark nucleus produces the appearance of a perinuclear halo that gives a distinctive fried-egg appearance. As in astrocytomas, grading of oligodendrogliomas appears to identify groups with differing prognoses. Although several schemes for grading have been proposed, no consensus over which tumor features are the most prognostically significant has emerged. Most investigators reserve the use of the terms high-grade and anaplastic oligodendroglioma for tumors with increased cellularity, mitotic activity, and other features of more malignant glial tumors (outlined earlier and in the section dealing with malignant gliomas).617 In children, at least one half of oligodendroglial tumors occur as a part of mixed astrocytic-oligodendroglial tumors. These tumors have a predilection for the frontal and temporal lobes.618
Mixed neuronal-glial cell neoplasms, such as ganglioglioma, gangliocytoma, DNET, and DIG, are grouped for convenience with the LGAs. Gangliogliomas are the predominant type of mixed tumors, and they are composed of glial tissue and a disorderly array of ganglion cells, some of which may be binucleate, that must be distinguished from normal neurons in the area involved. The glial component most commonly is astrocytic, but it may be oligodendroglial. Although anaplastic gangliogliomas are uncommon, when they do occur, they usually involve anaplasia within the glial component; anaplastic involvement of the neuronal component is unusual.619
Low-Grade Cerebellar Tumors
LGAs occurring in the cerebellum typically are astrocytomas. Two principal histologic variants have been described. The classic, or pilocytic, astrocytoma accounts for 80% to 85% of these tumors (Fig. 27.21), and it is composed of fusiform astrocytes loosely interwoven with a fine fibrillary background and no (or rare) mitoses. A frequent microcystic component and the presence of Rosenthal fibers, thought to represent degenerative changes in astrocytes, also are common. Large macrocystic structures filled with proteinaceous fluid and containing a mural nodule are seen in as many as 50% of patients. The walls of these cysts may be highly vascular, leading to occasional instances of spontaneous hemorrhage. Although this tumor often displays features otherwise associated with malignant behavior, such as nuclear atypia and focal leptomeningeal invasion, it rarely behaves in other than

a benign fashion.174,489 Although most tumors remain confined to the cerebellum, direct extension through the cerebellar peduncles to involve the brainstem may occur. On rare occasions, both pilocytic and fibrillary cerebral astrocytomas have demonstrated neuraxial dissemination or late malignant transformation, a behavior that belies their typically low-grade histologic features.620 In contrast to supratentorial astrocytic tumors, high-grade astrocytomas, such as glioblastoma, are uncommon in the cerebellum.
Figure 27.21 Typical biphasic pattern of a pilocytic astrocytoma. Note the dense, relatively anuclear fibrillar areas alternating with looser honeycombed fields (hematoxylin and eosin, 3250).
The second variety of cerebellar astrocytoma is the diffuse or fibrillary astrocytoma, which accounts for 15% of cerebellar astrocytomas and is similar to the diffuse LGA of the cerebral hemispheres. This tumor is more densely cellular, lacks the microcysts and Rosenthal fibers common to the pilocytic tumors, is more widely infiltrative, and is more likely to undergo anaplastic change than is its counterpart.489
Prognostic Considerations
Published reports of the management of LGA in children are complex, and the identification of consistent prognostic factors is difficult. Most reports include adult and pediatric cases, tumors from all sites, and patients treated over several decades, during which time diagnostic and therapeutic techniques have changed. The very good outcome reported by most authors and the indolent natural history of these tumors confounds analysis as well. Even so, certain factors consistently emerge in analyses but with inconsistent results. Complete resection of tumor seems most important for achieving prolonged disease-free survival in most, but not all, series.590,600,601,604,606,621,622 After a radical resection (i.e., greater than 90% of tumor resected), 5-year PFS rates for cerebral astrocytomas exceed 75%176,621,623 (Table 27.10) versus less than 50% after incomplete resections.176,177 However, the amenability of these tumors to second surgical explorations results in survival rates that may not differ from tumors completely resected at diagnosis.
The independent influence of the histology of tumors—generally pilocytic versus nonpilocytic—is controversial. Some reports support superior survival rates with pilocytic histology,604,624,625,626,627,628 and others report equivalent outcomes for pilocytic and nonpilocytic tumors.177,252,606,609 Within the histologic category of pilocytic astrocytomas some retrospective reports suggest that tumors with a higher proliferative fraction, measured by either a MIB-1 labeling index or a bromodeoxyuridine (BrdU) labeling index, are associated with a shortened PFS or outcome, respectively.629,630 Pilomyxoid astrocytomas, which in retrospective series were previously classified as pilocytic astrocytomas, appear to have more aggressive biologic behavior with a predilection for the hypothalamic/chiasmatic region, a tendency to occur among young children, and a greater potential for leptomeningeal dissemination, compared to pilocytic astrocytomas.612 Other factors that are somewhat related to histology and to the degree of resection include the invasiveness of the tumor into surrounding structures and the amount of residual tumor. For example, pilocytic and oligodendroglial tumors, which often are well circumscribed, appear more often to be amenable to extensive resection and more likely than fibrillary astrocytomas to have a favorable prognosis.177,590,618,621,628 The tendency toward invasiveness among nonpilocytic astrocytomas is in keeping with the less favorable prognosis that some have reported for these tumors.624,626 However, separating the effects of histology and the extent of resection is difficult. Invasion into the brainstem is a primary factor limiting complete resection of cerebellar LGA600,604,631 and is an independent prognostic factor in one of these series.631 Volume of tumor residual emerged in another series as the most important predictor of cerebellar LGA progression, emphasizing the importance of maximal tumor resection.604
Treatment Survival in yr (%) References
5 7 10
Complete resection 76–100 86
60 (d)
69–100 590, 621, 685
Incomplete resection 62 67–87 590, 638
Incomplete resection plus irradiation 80 67–94  
58–93 (d) 77 (d) 67 (d) 590, 638
d, diencephalic.
Although the long-term benefit of radiation therapy is not entirely clear, higher-dose radiation (i.e., greater than 53 Gy) significantly improved length of survival in Shaw’s series.590,628 Young age is noted consistently to increase the risk of progressive disease.600,606,632. The CCG9891/POG9031 natural history study, which accumulated more than 700 cases of LGAs with centrally reviewed pathologic material and radiologic studies, confirmed the strong association between resection extent and outcome and may help to address more conclusively the independent contributions to prognosis of histology and tumor location.239

The goals of surgery for LGAs are to obtain tissue for diagnosis and to remove as much tumor as is safely feasible. Current operative mortality rates are less than 1%; morbidity depends largely on tumor location and is highest in diencephalic tumors,464 in which the incidence of hemiparesis or visual field deficits may be 10% to 20%. Although gross total excisions are possible in 40% to 80% of hemispheric tumors, fewer than 40% of diencephalic tumors are similarly resectable.633,634 Although, the use of microsurgical techniques has led to high rates of resection in selected diencephalic tumors, the efficacy of chemotherapy for deep-seated LGAs in young children has tempered enthusiasm for aggressive resections of such tumors in view of the potential for substantial morbidity.464
Cerebral Hemisphere Gliomas
Although complete resection usually is the operative goal for cerebral hemisphere LGAs, its achievement may be difficult for nonpilocytic tumors, which rarely have a distinct tumor–brain interface. Because malignant degeneration of nonirradiated cerebral LGAs is uncommon, lesions that progress after an initial operation often are amenable to repeat resection. This possibility contrasts with the situation in adults, in which the majority of LGAs exhibits malignant features at the time of progression.635
For cortical and many subcortical tumors, the surgical approach follows the most direct trajectory to the lesion. In recent years, a variety of physiologic monitoring tools have been implemented to facilitate extensive resection of gliomas in “eloquent” regions of the brain. These include fMRI and direct cortical mapping using strip, grid, and bipolar contact electrodes. Although it is difficult to prove that any or all of these modalities are essential to achieving tumor resection, they clearly increase the comfort level of a surgeon attempting resection of a tumor in, or adjacent to, a functionally critical region of the cortex. Similarly, investigators have debated the need for intraoperative ECOG (direct cortical electroencephalographic monitoring) in children with tumor-associated epilepsy; 75% of patients whose tumor resections are performed without ECOG are free of seizures postoperatively, versus 85% of those with ECOG.251,636,637 If used, ECOG is likely to be of most value in treating those patients with long-standing or severe seizure disorders.
For deep or poorly circumscribed superficial lesions, imaging-based neuronavigation and intraoperative imaging using ultrasonography and MRI guidance are helpful for planning an approach to the tumor that avoids traversing critical regions and for monitoring the progress of the resection. These strategies are particularly valuable for lesions that arise from or extend into the thalamus and basal ganglia.
The approach to deep subcortical lesions also is influenced substantially by the predominant direction of tumor growth. Lesions deep within the temporal lobe or temporal horn of the lateral ventricle are approached through a corticotomy in the middle temporal gyrus or sulcus. Lesions that grow medially and encroach on or expand within the lateral ventricle can be approached transcallosally or transfrontally, through the middle frontal gyrus, whereas tumors that extend laterally in the nondominant hemisphere may be approached through the insula after the sylvian fissure has been opened. Laterally extending lesions within the dominant hemisphere and tumors that arise more posteriorly within the thalamus may be reached using a posterior parietal approach situated behind the sensorimotor cortex and above the angular gyrus or through a posterior incision in the middle temporal gyrus. Such lesions can be reached also via an occipital transtentorial trajectory via an opening in the pulvinar. Finally, tumors that project anteriorly and laterally can be reached from a paramedian frontal trajectory, provided that care is taken to avoid injury to the motor pathways.
Cerebellar Gliomas
As with supratentorial hemispheric gliomas, a close correlation exists between the extent of resection and outcome in cerebellar gliomas; complete tumor excision is associated with improved long-term and disease-free survival. Because the survival of patients with GTRs is as high as 90% for up to 30 years, aggressive attempts at resection are warranted, except when the tumor has invaded the cerebellar peduncles or brainstem.174,603,624,632,638 The operative approach is similar to that described earlier for resection of a MB. Data suggest that as few as 36% of patients with subtotal resections remain free of relapse at 6 years, and actuarial survivals similarly may decline with longer follow-up times of 10 to 20 years.174,252,603,639,640 Thus, the appearance of resectable residual tumor on a postoperative scan frequently is an indication for reoperation.
Most pilocytic astrocytomas have a distinct margin and can be separated from adjacent cerebellum with reasonable safety; currently, as many as 90% of patients have GTRs with an operative mortality rate of less than 1%.174,624,626,640,641 Although complete resection is also feasible for the majority of nonpilocytic astrocytomas, its achievement is more difficult because these lesions rarely are as well circumscribed as the pilocytic tumors.
Gangliogliomas and Other Benign Neuroepithelial Tumors
The approach to gangliogliomas is similar to that for other LGAs. Complete resection of cerebral gangliogliomas may be associated with survival rates in excess of 90% at 10 years, and recurrences are infrequent. The recurrence rate is substantially higher for deep-seated lesions, such as those within the diencephalon, because of the difficulties in achieving a GTR. However, even after partial resection, long-term progression-free intervals may ensue. The response of the desmoplastic mixed neuronal-glial tumors appears similar; GTRs generally are curative, and incomplete removals have been associated with local recurrence or tumor progression.181,642,643,644,645,646
Radiation Therapy
The role of radiation therapy in patients with LGAs depends on the anatomic location and tumor extent, age of the child, and the degree of resection. With current neuroimaging, children with completely resected tumors or small amounts of apparent residual are followed without further intervention.239,606,647,648,649 For unresectable tumors (e.g., midbrain, many thalamic or large temporal lobe lesions), radiation therapy has been shown to be effective as gauged by disease response and durable disease control.648,650,651,652 A major

retrospective review of childhood astrocytomas confirmed the improvement in long-term event-free survival following irradiation for incompletely resected astrocytomas, while showing no apparent benefit in overall survival.590 There are specific histologic subtypes of the LGAs that are associated with more aggressive biology (e.g., PMAs and pilocytic tumors with high MIB-1); the potential role of early postoperative irradiation in these settings has yet to be fully explored.613
Recent experience with LGAs has been based on image-guided radiation techniques, including 3D-CRT, stereotactic radiation therapy (essentially, fractionated radiosurgery), and more recent experience with IMRT or proton beam irradiation.648,653,654,655 In most recent series, local irradiation has been quite narrowly applied for discrete JPAs or focal astrocytomas that are not histologically characterized. Margins in several instances are ≤1 cm beyond the enhancing tumor, allowing truly conformal techniques for many of the focal presentations. The use of conformal irradiation for recurrent and high-risk newly diagnosed children older than 10 years with LGAs will be evaluated prospectively in the COG ACNS0221 study.
The role of radiation therapy in incompletely resected cerebellar tumors is unclear. All reported comparisons involve retrospective reviews of patients accumulated over periods as long as 40 years, in most instances following patients after incomplete resection or with inconsistent indications for and techniques of irradiation. As suggested for other LGAs, patients with incomplete resections of cerebellar tumors should be considered for further surgery, and radiation therapy should be reserved for progressive residual or unresectable recurrent tumors. Radiosurgery has been utilized for highly focal recurrences in difficult to resect locations, such as the cerebellar peduncles or brainstem.273,274
Irradiation has been proposed for patients who have gangliogliomas and have undergone incomplete resections or disease recurrence. However, the reported numbers of patients treated with radiation therapy are insufficient to estimate reliably the long-term utility of such treatment. Long periods free of tumor progression may follow incomplete resection alone.
A few investigators have suggested that irradiation may be associated with late anaplastic transformation of low-grade tumors.656 The phenomenon has not been suggested in other major institutional series with long-term follow-up of astrocytic tumors after irradiation; it is difficult to be categorical regarding an etiologic relationship between tumor dedifferentiation and radiation exposure in childhood astroctyomas.649,650
LGA within the cerebral hemispheres and cerebellum is primarily a surgical disease. If tumors recur after complete or incomplete resection, subsequent resection may lead again to prolonged disease-free status. However, aggressive primary or secondary surgery may be unsafe for tumors in deep locations or eloquent structures. In such situations, particularly in young children in whom a delay in radiation therapy may be desirable, or for children whose tumor has progressed after irradiation, chemotherapy has been explored. Numerous single-agent and combination chemotherapy regimens, which generally include classic and/or nonclassic alkylators, nitrosoureas, and/or platinum analogues, have been explored in children from infancy through adolescence with LGAs of all sites.657,658,659,660,661,662,663,664,665,666,667,668 The usual benefit of chemotherapy is disease stabilization, although partial responses may occur. Complete tumor regression is rare, and progressive disease occurs in a subset of patients in many series.
The most promising chemotherapy data for children with LGAs come from two reports. In the first, carboplatin and vincristine were administered to 73 children (mean age, 3 years) with newly diagnosed, progressive LGAs that were primarily diencephalic. Radiographic responses were seen in 56% of patients and the 3-year PFS rates were 68%. There were no correlations among tumor histology, location, or maximum response to chemotherapy and duration of disease control. However, children 5 years old and younger had a significantly higher rate of 3-year PFS (74%) as compared with children older than 5 (39%; p < .01).661 Similarly, another chemotherapeutic regimen consisting of 6-thioguanine, procarbazine, dibromodulcitol, CCNU, and vincristine resulted in tumor reduction in 36% of patients, stable disease in 59%, and a 5-year survival of 78%.666 In contrast to the carboplatin-vincristine data, older age was the only factor that improved survival significantly. These two regimens form the basis of an ongoing randomized study within the COG, A9952, for patients with progressive LGA.
Although chemotherapy appears to be a viable treatment option for children in whom either aggressive surgery or radiation therapy is inadvisable, the natural history of LGAs, characterized by recurrence or progression rates many years after diagnosis or after irradiation, will require up to 20 years of follow-up to determine its long-term benefit. In the shorter run, comparative activity of various agents will be determined. Biologic factors may guide chemotherapy further in the future.
Optic pathway tumors (OPTs) generally are those neoplasms that arise in the optic nerves, chiasm, and hypothalamus and may extend from these sites to adjacent brain structures. Together, they represent approximately 5% of pediatric intracranial tumors.669,670 Nearly two thirds of OPTs are diagnosed in the first 5 years of life.670 Seventy-five percent of OPTs will become symptomatic in the first decade of life, and 90% will become so before the age of 20.671 Boys and girls are affected equally.
OPTs are prevalent in patients with NF-1. The incidence of NF-1 in reported series of OPT ranges as high as 28%,672 whereas as many as 70% of OPTs may be associated with NF-1.206 The sites of OPT involvement appear to differ in patients with and without NF-1. Unilateral or bilateral optic nerve involvement alone is seen almost exclusively in patients with NF-1, whereas chiasmal involvement is significantly more common in patients without NF-1.672,673
Pathology and Patterns of Spread
Histologically, OPTs usually are low-grade pilocytic and occasionally fibrillary astrocytomas with microscopic features virtually identical to those of the classic cerebellar astrocytoma

and other midline pilocytic tumors. Malignant degeneration is rare.674 Although these tumors usually are confined to the structures of the visual pathways and extend in contiguity along them (Fig. 27.22), they may extend also into the frontal lobes, hypothalamus, thalamus, and other midline structures. Such events are more frequent in chiasmal tumors. Overall, tumor growth is slow, although alternating periods of clinical progression and stability suggest an erratic growth pattern.674,675,676,677,678,679
Figure 27.22 A: Diffuse infiltrating glioma of the right optic nerve (arrow). Note the diffuse enlargement of the nerve and absence of a separate mass lesion. B: Optic nerve glioma formed by elongated, swirling piloid processes of astrocytes, the nuclei of which are inconspicuous. Note the plump Rosenthal fiber (arrow) (hematoxylin and eosin, ×250).
Prognostic Considerations
Three factors appear to have prognostic importance for the outcome of OPTs, although reports are not always consistent. The presence of NF-1 generally is associated with more indolent disease, reflected in longer times to disease progression and higher rates of PFS and survival.178,673,680,681,682,683,684,685,686 For example, 15-year relapse-free survival in Jenkin’s report684 was 84% for patients with NF-1 and 47% for those without (p = .0007). In Imes’s series,683 patients with NF-1 survived their OPT better than did those without NF-1 but died of other causes, including other intracranial tumors and complications of NF-1 resulting in a survival not different from nonneurofibromatosis patients. In other series, NF-1 does not offer a protective effect.650,687,688 Tumors involving the chiasm and hypothalamus have a worse prognosis in most (but not all) series.674,680,684,689,690 Finally, as with most other CNS tumors, the youngest children, generally younger than 3 to 5 years, do worse than do their older counterparts.680,687,691,692
Because of the variability in the growth properties of OPTs, diverse approaches to management have been advocated in different clinical situations and in different institutions. In children with NF-1, the etiology of the lesion rarely is in question. Because the tumor usually exhibits diffuse involvement of the chiasm and nerves, it is not amenable to extensive resection. If the tumor is behaving in a biologically indolent fashion, resection generally is not pursued,27,651,693 and adjuvant therapy, if needed, is initiated empirically.
Lesions that seem particularly well suited to radical excision are those that involve only a single optic nerve and produce progressive, disfiguring proptosis or blindness (or both) and those that grow exophytically from the optic chiasm and produce significant mass effect or hydrocephalus.651,676,677,691 For isolated optic nerve gliomas, which are fairly uncommon, the tumor can be removed with preservation of the globe. In such cases, the resected segment of the optic nerve should be as long as possible, preferably extending close to the chiasm, to diminish the risk of local tumor recurrence.677 Ruling out the diagnosis of NF-1 before embarking on surgery is essential, because such children commonly exhibit widespread involvement of the optic pathways and may have long-term stabilization of vision without aggressive intervention. For exophytic chiasmatic-hypothalamic tumors, resection often is pursued to relieve obstructive hydrocephalus, to reduce local mass effect, and to establish a tissue diagnosis. For lesions amenable to resection, the tumor is approached via a subfrontal, trans-sylvian, or transcallosal exposure, depending on the pattern of tumor growth. Although a complete resection is not feasible because these lesions infiltrate the optic chiasm or hypothalamus (or both), substantial symptomatic improvement sometimes can be achieved.651,691 In occasional cases, removal of a significant portion of the tumor may stabilize the disease and delay the need for additional therapy.691
Open biopsy also is pursued sometimes for lesions involving the chiasm in which the histologic diagnosis is uncertain, before instituting further therapy. This is especially the case in children without neurofibromatosis or those with isolated chiasmatic-hypothalamic lesions without contiguous optic nerve or optic tract involvement. However, this procedure may further compromise vision in a significant percentage of patients.159,691,694 Alternatively, some neurosurgeons prefer to perform a stereotactic biopsy and treat with chemotherapy, reserving open resection for lesions that fail to respond or subsequently progress. Although aggressive surgery may be of potential benefit in some patients with large, progressive lesions, it may be associated with significant morbidity, particularly in the youngest patients, and does not convincingly improve survival in comparison to more limited open or stereotactic biopsy and adjuvant therapy.691,695
Radiation Therapy
Radiation therapy is effective in stabilizing or improving chiasmatic/hypothalamic gliomas. Response is apparent in objective reduction in tumor size and long-term stabilization.647,649,650,684,696 Functionally, irradiation typically results in stable visual acuity; vision has been documented to improve in 20% to 25% of instances and to deteriorate in an equal percentage.647,649,650,685,696 Deterioration is often associated with enlargement of tumor-associated cysts or a postirradiation phenomenon of transient increase in tumor size seen in a significant minority of children within 3 to 9 months after irradiation.650 In older children (institutionally defined over a broad range of older than 4 to 5 years old to after puberty), irradiation is the standard intervention when disease progression

or significant visual compromise prompts therapy. For relatively younger children, particularly those younger than 4 to 5 years, radiation therapy is often employed after chemotherapy, reserving the more durable efficacy of irradiation for children with documented progression postchemotherapy.649,685,697,698 Delaying irradiation is appropriate in view of age-related toxicities, particularly vascular effects (vascular compromise believed to result from cicatricial reduction of the major arteries in the circle of Willis, either focally or globally, resulting in moyamoya syndrome, essentially complete loss of the major intracranial vessels in the suprasellar region) and neurocognitive deterioration.647,697,699
Radiation therapy for OPTs should be highly conformed to the chiasm/hypothalamus for lesions localized to that region, using 3D-CRT, IMRT, stereotactic radiation therapy (i.e., fractionated radiosurgery), or proton beam.648,653,700 In more extensive tumors extending into the optic tracts, margins are somewhat greater for the tumor components outside the suprasellar region. Dose levels have been established at 50 to 54 Gy using fraction sizes approximating 180 cGy.650,684,696
The optic glioma chemotherapy experience is greatest for children who are younger than 3 years and for whom delay of radiation therapy is desirable for avoidance of long-term neuropsychological and neuroendocrine effects. Packer’s regimen of vincristine and actinomycin-D was the first to successfully defer radiation therapy for children who were younger than 5 years with progressive chiasmatic and hypothalamic gliomas. At a median of 4 years of follow-up, 62.5% of the patients remained free of progressive disease and had not received radiation therapy.701 By 7 years, however, only one third of patients were free from progression. Packer also tested a carboplatin and vincristine regimen described for LGAs.661 Recently, the POG reported the results of its phase II study of carboplatin for children who were younger than 5 years with progressive OPT. Carboplatin was administered alone every 4 weeks at a dose of 560 mg per m2. After two courses, patients were evaluated, and those with stable disease or better were continued on therapy for 18 months or until disease progression. Of 50 eligible children, including 21 with NF-1, 39 (78%) had stable disease or better, and 34 completed therapy. Six children, of whom only one had NF-1, died of disease.702 These reports have established carboplatin as an effective chemotherapeutic agent for OPT.
Petronio et al.703 reported improvement or stable disease in 15 of 18 patients with progressive OPT a five-drug regimen consisting of 6-thioguanine, procarbazine, dibromodulcitol, CCNU, and vincristine. Of 15 patients, 11 had a greater than 50% decrease in their tumor mass. Those children whose tumors progressed on or after completion of chemotherapy were successfully treated with radiation therapy. Using chemotherapy, vision was stabilized in 14 of 18 patients and improved in 2.
The ongoing COG study for progressive LGAs includes children with OPT and is a randomization between carboplatin-vincristine and a modified 6-thioguanine, procarbazine, cisplatin, and vincristine regimen. Future studies will examine the efficacy of other agents for OPT. Already, reports of limited numbers of patients have identified cisplatin and vincristine, tamoxifen and carboplatin, oral etoposide, velban, and temozolomide as having activity against progressive OPT.667,668,704,705,706
Most reports of chemotherapy for LGAs include patients whose tumors arose from the optic pathway. (See the section on low-grade glioma for additional chemotherapy data.)
Anaplastic astrocytoma, glioblastoma multiforme, and mixed glial tumors with a preponderance of malignant astrocytic elements collectively compose malignant or high-grade astrocytic tumors in children. These tumors represent 7% to 11% of childhood CNS tumors. When primary brainstem tumors are excluded, combined series suggest that approximately 25% of malignant or high-grade astrocytic tumors occur in deep midline structures of the cerebrum, not more than 15% occur in the posterior fossa, and the majority occur in the cerebral hemispheres.230,231,707,708,709 The median age at diagnosis is 9 to 10 years, and the male-to-female ratio is near unity.230,708,710
Pathology and Patterns of Spread
As noted in the section on low-grade astrocytic tumors various classification schemes have been applied to astrocytic tumors. High-grade lesions, unlike their low-grade counterparts, generally are characterized by the presence of several histologic features of malignancy that include hypercellularity, cytologic and nuclear atypia, mitoses, necrosis, endothelial proliferation, and other anaplastic features; lesions with these features may be termed malignant or high-grade gliomas. The most common malignant glial neoplasms are the high-grade astrocytomas, such as the anaplastic astrocytoma (Fig. 27.23) and glioblastoma multiforme (Fig. 27.24), which may alternatively be termed grade III and grade IV astrocytomas, respectively.108 Similarly, the term high-grade or anaplastic

may be used to describe other, less common glial neoplasms, such as oligodendroglioma, ganglioglioma, or mixed astrocytic-oligodendroglial neoplasms.
Figure 27.23 Highly cellular tumor composed of anaplastic astrocytes. A few vessels with compressed lumina are formed by swollen epithelial cells (hematoxylin and eosin, ×250).
Figure 27.24 A: Typical field in glioblastoma multiforme showing pseudopalisading (upper left), neovascularity, nuclear anaplasia, and multinucleate giant cells (lower right). (Hematoxylin and eosin, ×250.) B: Higher magnification of a different field from the same tumor showing cellular anaplasia, multinucleate cells, and bizarre mitosis (upper left corner) (hematoxylin and eosin, ×400).
The high-grade astrocytomas are clinically aggressive, regionally invasive, and capable of extraneural dissemination to lung, lymph nodes, liver, and bone, particularly in adults. In children, these and the other malignant gliomas occur most commonly in the cerebral hemispheres, in contrast to the more frequent cerebellar and deep midline locations for low-grade tumors. Although the rapid growth and effacement of normal tissue produced by high-grade tumors may produce what appears to be a well-demarcated tumor, microscopic study frequently demonstrates extension for up to several centimeters beyond this margin. Distant neuraxial dissemination, once considered unusual, has been demonstrated in as many as 25% to 50% of high-grade astrocytomas, both at diagnosis and postmortem in carefully evaluated series of patients.709,710,711
High-grade gliomas may have a histologically heterogeneous nature in that areas of low-grade histology commonly are noted in many high-grade tumors, particularly in small biopsies taken from the more superficial areas of tumor. Diagnostic confusion may be reduced by more generous sampling and by directing stereotactic biopsies toward the contrast-enhancing or more central portions of the tumor.
Prognostic Considerations
The extent of surgical resection, regardless of the primary tumor site is the most important clinical prognostic factor for children with high-grade astrocytomas as demonstrated by institutional and cooperative group series.229,230,231,232,707,708,712,713 Outcome is better for patients with a complete or near complete resection of primary disease. For example, in the CCG-945 study, children with greater than a 90% tumor resection had a median PFS that was significantly longer than patients with more limited resections. The difference was more pronounced for patients with anaplastic astrocytomas (31 vs. 12 months) than for those with glioblastoma multiforme (12 vs. 8 months) but was statistically significant for both groups.230,232 The relationship may be surprising, given the fact that tumor often extends well beyond the identified central component of tumor714,715,716 and that, even after a radiologically confirmed GTR of all contrast-enhancing tumor, extensive residual tumor is known to remain. A caveat worth emphasizing is that factors specific to the tumor, such as the pattern of growth and degree of infiltration, may determine which tumors are amenable to extensive resection. Thus, tumors that are amenable to resection may constitute a group biologically more favorable than those that infiltrate extensively into the surrounding brain.717 Insights from genomic analyses of these tumors may provide better understanding of these issues.
In several series, site of disease is independently prognostic of outcome as well, with deep midline tumors showing a survival poorer than that of cerebral hemispheric tumors. However, hemispheric tumors are more amenable to radical resection than midline tumors.230 Data are conflicting in regard to the impact of gender and age at diagnosis; however, female gender and younger age have been shown by some investigators to affect outcome favorably.230,232,709,718 The impact of histology—anaplastic astrocytoma versus glioblastoma multiforme—is debated as well. Anaplastic astrocytoma may be favorably prognostic for subsets of patients,229,230,232 although this association has not been uniformly confirmed, possibly reflecting challenges in the reliable classification of these tumors in historical series.719. The prognosis of biologic markers in children with supratentorial high-grade gliomas is not well defined. Factors that have been associated with outcome, independent of histology, include the presence of TP53 overexpression104 and MIB-1 proliferation index.720 PTEN mutations have been associated with an unfavorable outcome and overexpression of basic fibroblast growth factor have also been associated with an unfavorable outcome.721,722
The surgical techniques employed for tumor resection are similar to those used for supratentorial LGAs. As with the latter tumors, malignant gliomas often are amenable to reoperation at the time of disease progression, because most lesions recur at the primary site. Although the majority of children succumb to further disease progression despite additional intervention, attempted re-resection may be warranted in tumors that are amenable to gross total or radical subtotal removal, in view of recent reports of long-term survival after extensive resection followed by high-dose chemotherapy.327 The efficacy of a single course of highly intensive chemotherapy versus three courses of less intensive therapy for children with extensively resected recurrent disease is being examined in a randomized COG study, ACNS0231.
Radiation Therapy
Radiation therapy is a standard component of postoperative management for children with malignant gliomas. Although rarely curative, the addition of irradiation alone had shown improved survival intervals in adults, and pediatric series demonstrating higher rates of 1- to 3-year disease control and survival have been based on a combination of irradiation and chemotherapy.230,231,707

Except in infants and young children younger than 3 to 4 years, postoperative therapy incorporates wide local irradiation. Volumes are MRI-derived, typically from both the enhancing tumor on T1 (preoperatively and as modified by surgical resection) and the extent of perilesional infiltration best estimated by findings on T2 or FLAIR sequences. There is evolving experience using other imaging modalities, such as PET or multinuclear MRS, to improve the targeting of subclinical disease.723,724
Dose levels are typically 56 to 60 Gy in pediatrics, often incorporating volume reductions after 50 to 54 Gy. The addition of stereotactic radiosurgical or interstitial implants to “boost” the residual tumor or immediate tumor bed has been explored.653,725,726,727
The role of chemotherapy for treatment of high-grade gliomas is evolving; the best results were observed in CCG-945, a prospective randomized trial performed between 1976 and 1981. Children with high-grade astrocytomas (excluding lesions of the brainstem or spinal cord) between 2 and 21 years of age were randomized to receive radiation therapy with or without adjuvant chemotherapy consisting of prednisone, CCNU, and vincristine (PCV). The PFS for children receiving postradiation chemotherapy was significantly higher (46%) than for those receiving radiation therapy alone (16%; p = .026).708 However, other studies in children and adults have not shown a similar benefit. For example, a follow-up CCG study randomized patients to receive adjuvant chemotherapy with 8-in-1 or PCV. No difference was observed between the two chemotherapy arms, and 5-year PFS rates were 33% and 36% for 8-in-1 chemotherapy and PCV, respectively.230 Taken together, these studies suggest that the benefit from addition of chemotherapy, when compared to surgery and radiation therapy alone, is modest at best. The COG is currently conducting a trial to evaluate the role of temozolomide, initiated with radiation therapy and continuing monthly thereafter, in children with newly diagnosed high-grade gliomas.
In contrast to this experience with anaplastic astrocytoma and glioblastoma multiforme, children the in CCG-945 study, with eligible diagnoses other than anaplastic astrocytoma and glioblastoma multiforme, fared well. Their 5-year PFS and survival rates were 64% and 71%, respectively. Although retrospective neuropathology review of this cohort using contemporary WHO classification guidelines demonstrated a high frequency of discordance among these “other eligible” patients with inclusion of a substantial percentage of low-grade tumors,719 the survival of the remaining subgroup was still more favorable than that of children with anaplastic astrocytoma or glioblastoma. Chemotherapy with procarbazine, CCNU, and vincristine has resulted in greater than 50% partial and complete responses in adult patients with anaplastic oligodendrogliomas and anaplastic mixed gliomas, particularly in those tumors that exhibit loss of heterozygosity in chromosomes 1p and 19q.728,729,730 These data suggest that CCNU and vincristine, combined with prednisone or procarbazine and in addition to radiation therapy, may have a positive impact on outcome for children with newly diagnosed high-grade gliomas. However, it is important to point out that anaplastic oligodendroglial tumors of childhood may not be molecularly identical to their adult counterparts731 and, thus, may warrant distinctive management approaches.
As a result of the poor outcome for children with most high-grade gliomas, numerous single-agent phase II studies have been conducted using a variety of agents, including cisplatin, carboplatin, CCNU, procarbazine, cyclophosphamide, ifosfamide, etoposide, topotecan, temozolomide, and irinotecan.663,718,732,733,734,735,736,737,738,739 The data for CCNU and cyclophosphamide have shown the most promise.663,718,736 Although temozolomide was approved for use in adults with recurrent or progressive anaplastic gliomas, the response rate following temozolomide treatment in children with recurrent or refractory high-grade gliomas is low.737 However, the activity of temozolomide may be higher in newly diagnosed patients or when administered in combination with O6-benzylguanine (O6 BG), an agent that may overcome temozolomide resistance mediated by the DNA repair protein O6-alkylguanine alkyltransferase (AGT). The COG ACNS0126 study is evaluating the former strategy and the PBTC is studying temozolomide in combination with O6BG.
Combination regimens of BCNU with cisplatin and cyclophosphamide with etoposide have also been evaluated by the POG. Although activity was demonstrated, superiority over CCNU and vincristine has not been determined.740 Myeloablative chemotherapeutic regimens have shown activity against bulk residual disease but have not demonstrated convincing superiority over standard therapy.327,351,353,741,742 Both procarbazine and topotecan were inactive when evaluated as single agents in patients with newly diagnosed high-grade glioma and measurable disease after diagnostic surgery.743,744
Clearly, new agents and new approaches to therapy are needed and this is an active area of clinical investigation. The PBTC and COG are conducting a phase 1 and 2 studies of novel small molecule inhibitors, including signaling mediators that are putatively involved in glioma growth, such as PDGF-R, EGFR, and Ras. In addition, the PBTC is performing phase 1 studies of convection-enhanced delivery of receptor-toxin conjugates.
Tumors arising in the midbrain, pons, and medulla oblongata now account for approximately 20% of all CNS tumors among children younger than 15 years.7 This apparent rise from 10% in the late 1970s does not reflect a truly increased incidence. Instead, this jump stems from the advent of MRI and increased detection of focal brainstem tumors, subsequently confirmed microscopically as LGAs. The median age of occurrence for all brainstem gliomas is 6 to 7 years.173 The male-to-female ratio is near unity. Brainstem tumors are noted to be increasingly frequent among patients with NF-1.745,746
Pathology and Classification
The term brainstem glioma is an imprecise descriptor suggesting that all these tumors behave in the same way. As a result of advances in neuroimaging over the last 2 decades, a plethora

of terms emerged to subclassify brainstem gliomas (sometimes confusingly) by site of origin or imaging features. Terms used have included midbrain tumor, tectal glioma, pontine glioma, focal medullary tumor, cervicomedullary tumor, diffuse glioma, intrinsic glioma, pencil glioma, dorsal exophytic brainstem tumor, focal glioma, and cystic glioma.609,652,747,748,749,750,751,752,753,754,755,756,757,758,759,760,761,762,763,764,765 Brainstem tumors are sometimes also subcategorized by pathology, as either low-grade (benign) or high-grade (malignant) gliomas.763,766,767,768 However, brainstem gliomas can be parsimoniously and better biologically classified as diffusely infiltrative brainstem gliomas and focal brainstem tumors, categories that combine tumor location and histology.173
Figure 27.25 Typical magnetic resonance imaging findings in a diffusely infiltrating brainstem glioma. A: T1-weighted sagittal view shows diffuse, fusiform enlargement of the pons with tumor spread superiorly and inferiorly. B: T2-weighted transverse view shows engulfment of the basilar artery.
Diffusely infiltrative brainstem gliomas are the classic brainstem tumor having an extremely poor prognosis. Most arise in the pons, and cause diffuse enlargement of that structure (Fig. 27.25). Engulfment of the basilar artery by tumor is specific for the diagnosis of diffusely infiltrative brainstem glioma but is not seen in all cases (Fig. 27.25B).173 A component of exophytic growth is seen in at least two-thirds of cases.749 Neoplastic infiltration extending into the midbrain, cerebral peduncle, cerebellum, or medulla is exceedingly common. These tumors are generally fibrillary astrocytomas by histology, sometimes well-differentiated, WHO grade II or, more often, high-grade anaplastic astrocytoma (WHO grade III) or glioblastoma multiforme (WHO grade IV).173 These diffuse gliomas may occasionally show disseminated neuraxis spread.769,770
Focal brainstem tumors are discrete, well-circumscribed tumors without evidence of infiltration or edema. These tumors may occur in any level in the brainstem but are most frequently seen in the midbrain or medulla rather than the pons (Fig. 27.26). More often, focal tumors are dorsally exophytic to the brainstem, sometimes effacing the fourth ventricle. Histopathology reveals that focal brainstem tumors are most commonly pilocytic astrocytomas or, rarely, gangliogliomas, both WHO grade I.173,762 Sometimes these tumors are cystic.749
Rarely, tumors of histologic types other than gliomas are found in the brainstem. Among infants, the highly malignant

AT/RT can arise in the brainstem.461 Embryonal tumors, including PNETs, can also occur in very young children as localized pontine tumors with an exophytic component and extension beyond the pons. Such tumors may have a predilection for leptomeningeal dissemination at diagnosis.771 Hemangioblastoma may arise at the brainstem also, although more commonly during adolescence or adulthood.
Figure 27.26 T1-weighted transverse magnetic resonance imaging with gadolinium shows a focal brainstem tumor with avid enhancement at the medulla. Biopsy demonstrated the tumor to be a pilocytic astrocytoma.
Prognostic Considerations
Prognosis depends principally on the tumor type but is also influenced by whether or not the patient has NF-1. Patients with diffusely infiltrative gliomas fare dismally, regardless of therapy. In most series, median survival is less than 1 year, and survival rates at 2 years are lower than 10% to 20%.772,773 In contrast, the prognosis for patients with focal brainstem tumors is relatively good if the tumor is accessible. Survival for these patients is reported to be between 50% and 100%.203,745,749,768 Many patients with small focal tumors in the midbrain, particularly the tectum, may do extremely well after shunting or third ventriculostomy alone and can demonstrate PFSs of up to 10 years or more.750,774,775
Among patients with NF-1, brainstem gliomas, whether diffuse or focal, display a generally indolent biologic behavior.745,746 The tumor may stabilize in size or regress without intervention and survival approximates 90% at 5 years. Thus, intervention should be limited to those lesions that exhibit rapid or unrelenting growth on serial neuroimaging or lesions that produce significant clinical deterioration.
The choice of treatment depends largely on whether the tumor is a diffusely infiltrative brainstem glioma or a focal brainstem tumor. Even among focal lesions, the site of the tumor affects selection of therapy.
A major advance in the surgical management of brainstem gliomas followed the recognition that this broad category of tumors encompasses biologically distinct groups that demand individualized therapeutic strategies. At one extreme are the diffuse intrinsic gliomas, which are biologically malignant, highly infiltrative, and not amenable to resection. Although biopsy usually is associated with low morbidity, it no longer is considered necessary because, in the context of a typical clinical presentation and characteristic MRI findings, histologic results do not currently influence treatment.215 At the other extreme are the benign intrinsic tectal gliomas, which also are not appropriate for resection because these lesions are extremely indolent and are best treated symptomatically with CSF diversion and observation (Fig. 27.27).748,750,774
In contrast to the preceding groups, and because of improvements in neurosurgical techniques and postoperative care, surgical resection is a reasonable option for focal gliomas that arise at the cervicomedullary junction and for dorsally exophytic gliomas. These lesions are histologically and biologically benign but, unlike tectal gliomas, commonly show gradual enlargement over time. In particular, extensive resection of exophytic tumors, leaving a thin rim of tumor on the surface of the brainstem, frequently achieves long-term PFS without further treatment.652,761,762,768 Some focal intrinsic, cystic, and solid brainstem lesions are likewise surgically resectable, and several authors have suggested that even after only partial resection, many patients require no further postoperative treatment. Although good survival rates have been achieved with surgical resection of symptomatic focal midbrain and medullary lesions, it remains to be determined whether these results represent an improvement over those obtained with stereotactic biopsy and local irradiation,748,756,776,777 particularly in view of the potential for significant surgical morbidity from aggressive attempts at resection.755
Figure 27.27 T1-weighted sagittal magnetic resonance imaging with gadolinium of a tectal glioma with minimal contrast enhancement (arrow).
Radiation Therapy
Radiation therapy is the mainstay of treatment for children with diffusely infiltrative brainstem gliomas. Improvement in symptoms, signs, and neuroimaging occurs in a majority of children, although the duration of benefit is measured in months, with few long-term survivors. 759,778,779 Attempts to improve outcome in the diffusely infiltrating pontine gliomas focused on altered radiation fractionation in the 1980s and 1990s; hyperfractionated delivery (i.e., twice-daily irradiation with fractions of 100 to 125 cGy) to doses (ranging from 64.8 to 78.0 Gy) showed no significant improvement.778,779,780,781,782 A randomized trial comparing hyperfractionated irradiation at the “apparently best”

dose level of 70.2 Gy to conventional fractionation (180 cGy once daily to 55.8 Gy) showed no difference, leading North American centers to consider conventional fractionation as the standard for current management and trials.268
Recent and ongoing trials are evaluating the addition of a diverse array of radiosensitizers (e.g., cytotoxic agents like cis-platinum or topotecan, chemical sensitizers like gadolinium texaphyrin, and biologic agents such as Gleevec and R115777).772,783,784,785
The use of cis-platinum concurrently with irradiation seemed to show similar or marginally inferior outcome when compared to irradiation alone.772
For dorsally exophytic brainstem tumors, radiation therapy is typically utilized when progression is apparent after initial surgery.652,762,763,768 Durable control after irradiation has been recorded in the majority of cases so managed.652,762 Similarly, focal intrinsic brainstem gliomas, which are often juvenile pilocytic lesions, respond well to fractionated focal irradiation, usually with long-term disease control.752,765,786 The low-grade brainstem lesions are treated with minimal margins circumscribing the immediate tumor region; dose levels are 50 to 54 Gy by conventional fractionation. Postirradiation intralesional necrosis or “swelling” is an infrequent but clinically significant phenomenon that requires vigilance in the postirradiation interval.787
A role has yet to be established for chemotherapy as standard treatment in the management of patients with brainstem tumors. With focal brainstem gliomas, weekly carboplatin with vincristine has shown activity in a very limited number of patients younger than 5 years.788 For diffusely infiltrative pontine gliomas, response rates to various single-agent and multiagent regimens have been disappointingly low, even when chemotherapy is given as first treatment, before radiation therapy.12,504,534,789,790,791,792 Adjuvant chemotherapy using PCV did not improve survival when compared with conventional radiation therapy alone in the only prospective, randomized study that has tested this question.793 Oral etoposide has been reported to have antitumor activity in a small numbers of patients with recurrent diffuse tumors,794 although this approach had less encouraging results in a recent COG trial.
Approaches that are being investigated in ongoing studies of the PBTC include the use of inhibitors of angiogenesis, farnesyltransferase, EGFRs, and PDGF-Rs. A series of radiosensitizing strategies, such as the use of concurrent gadolinium texaphyrin and irradiation, are being examined in the COG. As patients with progressive diffusely infiltrative pontine gliomas survive, on average, at least 3 months after their first relapse, they can be candidates for further treatment in clinical trials.795 Such efforts may delay further disease progression and are requisite for improving the outcome in this disease group.
Tumors of the pineal area account for 0.4% to 2.0% of all primary CNS tumors in children. Three principal groups of tumors—germ cell tumors, pineal parenchymal tumors, and astrocytomas—account for most tumors in this location. In combined clinical series, astrocytomas constitute 15%, the pineal parenchymal tumors 17%, and germ cell tumors 40% to 65% of all neoplasms in this area (Fig. 27.24). Of CNS germ cell tumors, two thirds occur in the pineal region and the remaining one third in the suprasellar region.796,797 Pineal parenchymal tumors are more frequent in the first decade of life and have a male-to-female ratio near unity. Germ cell tumors are most common in the second decade of life or later, have a peak incidence at between 10 and 14 years of age, and are associated with a male-to-female ratio of at least 2:1 and as high as 9:1. Astrocytomas tend to occur in two separate age groups, 2- to 6-year-old children and 12- to 18-year-old teens, and each group has a 2:1 male-to-female incidence characteristic of astrocytomas elsewhere in the CNS.171,798,799,800
Pathology and Patterns of Spread
Pineal Parenchymal Tumors
Pineoblastoma is a primitive undifferentiated tumor that accounts for approximately 50% of the pineal parenchymal tumors. Except for its location, this tumor may be indistinguishable from MB and is considered by some investigators to be a variant of the PNET. This highly cellular tumor has frequent mitoses and areas of focal necrosis. The occasional presence of Flexner-Wintersteiner rosettes and fleurettes indicates differentiation toward retinoblastoma.489 Although the histologic appearance of the pineocytoma (Fig. 27.28) may overlap that of the pineoblastoma, the cells generally are larger and have a recognizable relation with blood vessels, and true rosettes rarely are seen. Occasional evidence of astrocytic, neuronal, or ganglion cell differentiation is noted.489
Germ Cell Tumors
The germ cell tumors include a spectrum of embryonal neoplasms and teratomas believed to be derived from totipotent

germ cells that aberrantly migrated to the cranial midline during embryogenesis. The germinoma accounts for approximately 60% of these tumors (Fig. 27.29). Germinomas have a typical two-cell appearance indistinguishable from that of gonadal germinomas and are composed of large, primitive-appearing cells intermixed with smaller lymphoid cells.171,489,801,802
Figure 27.28 Primary pineal tumor displaying prominent perivascular growth of neoplastic cells characteristic of pineocytoma. Note the papillary pattern (hematoxylin and eosin, ×250).
Figure 27.29 Typical field of germinoma displaying large cells with large nucleoli and focus of lymphocytes (right half of field) (hematoxylin and eosin, ×400).
Teratomas and mixed germ cell tumors harboring various mature and immature elements constitute nearly 30% of the pineal nongerminomatous germ cell neoplasms, and the highly malignant embryonal carcinoma, choriocarcinomas, and endodermal sinus tumors constitute the remaining 10%. Together, these tumors are often referred to as nongerminomatous germ cell tumors. The histologic appearance of these tumors is identical to that of similar tumors occurring outside the CNS. Teratomas generally remain local, well encapsulated, and noninvasive. However, areas with more primitive germ cell elements may be present and are associated with a more aggressive clinical course that may include neuraxis dissemination.489,802
Both contiguous regional extension and distant intraaxial dissemination are common with nonastrocytoma pineal region tumors. The overall incidence of leptomeningeal spread in larger series of patients is approximately 10%, with pineoblastoma and germinoma demonstrating the greatest frequency of such spread.171,798,801,802 Even pineocytomas, once considered slow-growing, locally infiltrative tumors, may have a high incidence of leptomeningeal seeding.803,804
Systemic metastases, although uncommon, may occur with the pineal parenchymal and pineal germ cell tumors. Bone, lung, and lymph nodes are the most common sites of such dissemination.802 Occasional instances of peritoneal metastasis associated with the use of ventriculoperitoneal shunts have also been noted.805
Prognostic Considerations
Tumor histology has prognostic significance. The germinomas and LGAs have the best overall survival rate and response to treatment, followed by teratomas and pineal parenchymal tumors. Five-year survival rates for intracranial germinomas are as high as 95%, but the remaining nongerminoma germ cell neoplasms have, historically, much poorer survival rates, ranging from 20% to 76%.806,807,808 Disease that has spread regionally, involves the hypothalamus, or has spread along leptomeninges also is associated with a worse prognosis. Although age at diagnosis is variably reported as having prognostic significance, this parameter may not be independent of the finding that very young patients (those younger than 3 years) have a higher incidence of disseminated disease at diagnosis and are frequently treated with reduced-dose irradiation.171,550,802,809
Because of the diversity in the biologic behavior and response to treatment of different types of pineal area tumors, biopsy is recommended whenever possible to establish a tissue diagnosis, which will help guide subsequent therapy.171,799,802 One exception is patients with benign intrinsic tectal tumors (discussed earlier with brainstem gliomas). A second exception is patients with malignant germ cell tumors in which α-fetoprotein or β-human chorionic gonadotropin (or both) is detected at high levels within the blood or CSF, in which case biopsy is considered optional.
Current neurosurgical techniques allow stereotactic, endoscopic or open biopsies in most patients, with morbidity generally limited to transient worsening of prior visual symptoms, although new or permanent losses may occur. The mortality rate is generally less than 2%. Direct visually guided biopsy is preferred by many neurosurgeons because of concern that stereotactic biopsies may injure adjacent deep veins. However, a variety of recent reports have demonstrated that stereotactic biopsy can be performed with acceptable morbidity, provided that a low frontal entry point is used to allow access to the tumor below the internal cerebral veins.810 This approach also provides CSF for analysis of α-fetoprotein and β-human chorionic gonadotropin, because the biopsy trajectory often traverses the lateral ventricle. Often, it is feasible to achieve CSF diversion using endoscopic third ventriculostomy while the patient is under the same anesthetic. With advances in endoscopic techniques, many surgeons opt to biopsy the tumor endoscopically under direct visualization, rather than stereotactically, at the same setting as the third ventriculostomy. Although these minimally invasive approaches have significant appeal and appear to carry a lower morbidity than conventional open craniotomy approaches, a major concern is the issue of sampling error. As many as 15% of germ cell tumors have mixed histology, which calls attention to the importance of adequate biopsy and of performing CSF and blood marker studies in such patients.799
If a stereotactic or endoscopic biopsy is nondiagnostic or equivocal, or if the histology of the lesion suggests that an open surgical resection is likely to be of benefit, as in the case of benign teratoma, tumor removal can be accomplished using one of a variety of operative approaches. An infratentorial supracerebellar approach is used for lesions that exhibit predominant growth below the level of the vein of Galen and basal veins of Rosenthal. A suboccipital transtentorial

approach is preferred for larger lesions that extend above the basal veins or down into the rostral fourth ventricle. Both approaches can be accomplished using a prone or modified prone position, which is generally preferred to the sitting position, to minimize the risk of air embolism. After division of the precentral cerebellar vein and surrounding arachnoid, the lesion is subjected to biopsy and then debulked using the ultrasonic aspirator. Great care is taken to avoid injury to the deep veins.
Except for well-encapsulated teratomas, few pineal region tumors are amenable to complete resection, generally because of extensive local or regional disease. Even though subtotal resections are possible for many patients with localized tumors, no evidence indicates that such resections improve outcome. Because many tumors in this area are sensitive to both radiation therapy and chemotherapy, and because aggressive surgery may cause significant morbidity, biopsy alone or limited tumor debulking to relieve hydrocephalus often is the most prudent approach initially, particularly for patients with germinomas. In addition, a variety of current treatment protocols for nongerminomatous germ cell tumors are employing intensive neoadjuvant chemotherapy after an initial open or stereotactic biopsy, followed by second-look surgery to perform a biopsy or remove areas of residual enhancement.807,811,812,813 In some cases, such reoperations indicate only scar tissue without evidence of viable tumor or foci of mature teratoma that may be amenable to resection, with favorable long-term results.
Pineoblastomas are considered to be PNETs, and their treatment and outcome are comparable to those of high-risk MBs. However, because these lesions are uncommon, previous studies have contained insufficient numbers of patients to determine conclusively whether extensive resection favorably influences outcome.548,809
Radiation Therapy
Radiation therapy for pineoblastomas, which generally are not completely resectable, is similar to that for high-risk MB—namely, craniospinal irradiation (in most ongoing studies to 36 Gy), followed by additional radiation to the tumor bed to a total dose of 55.8 Gy. Whether a single-fraction stereotactic “boost” to residual disease adds to overall control is under investigation. For incompletely resected, nondisseminated pineocytomas, local field irradiation to 50 to 54 Gy is appropriate.
Germinomas are exquisitely radiosensitive tumors. Typically, 10-year survival rates exceeding 90% after operation and radiation therapy can be expected.796,799,801,802,806,810,814 Radiation therapy is the standard of care for CNS germinomas in most settings; the use of combined chemotherapy and irradiation is increasingly utilized in children presenting prior to puberty. Whether local irradiation or full neuraxis irradiation is ideal for germinomas treated with radiation therapy alone remains a point of considerable debate. For localized pineal germinomas, series utilizing wide local volumes (defined as “local” or third ventricular) achieve durable disease control (or cure) in less than 80% to 90% of instances; series utilizing low-dose neuraxis irradiation (most often 24 Gy to less than 30 Gy) show few if any failures posttherapy. Given the high rates of disease control, it is difficult to prove a significant difference despite clear impressions favoring one approach over the other.800,815,816,817,818,819,820,821,822,823
Although the rate of neuraxis failure had been reported to be higher with suprasellar than pineal germinomas, most recent reports show little difference in approach or outcome.816,824 Metastasis at diagnosis is limited to 15% to 20% in North American series; such patients are similarly curable, requiring doses of craniospinal irradiation of 30 to 35 Gy, sometimes followed to local boost volumes for sizable metastatic sites.819,821,822,823,825 The preliminary experiences in European studies recently indicated an unexpected rate of intracranial failures, leading SIOP and the proposed COG study to adopt full-ventricular irradiation as the initial volume when using irradiation alone (to 24 to 30 Gy using 150-cGy fractions) followed by similar boost irradiation as noted above.823,824,826 Doses to the primary tumor site (or, in the case of multiple midline germinomas, to the involved region of the third ventricle or the entire third ventricular region) are typically 40 to 45 Gy in recent series.821,822,823,824,827,828,829
With conformal techniques, especially IMRT, one can treat the primary tumor to 45 Gy while irradiating the third ventricle to 30 Gy. When combined chemoradiation is utilized in germinomas, tumors demonstrating complete or near complete response to “induction” chemotherapy are typically treated to local volumes (essentially incorporating only the regions used as boost volumes when using irradiation alone) using doses of 30 to 35 Gy.111,826,830,831,832,833 Because excellent disease control seems to be achieved with either primary irradiation or combined chemotherapy-irradiation, future directions will depend on demonstration of differences in functional outcome sought in the planned prospective COG trial.822,826
In North American studies, craniospinal irradiation is recommended for nongerminomatous germ cell tumors in conjunction with chemotherapy; the only exception is mature teratoma.799,804,809,821,834,835,836 There is only limited data to support the use of more localized radiation volumes for these more aggressive germ cell tumors.824,837,838 Dose levels recommended for the more aggressive “non-germinomatous” germ cell tumors are 36 Gy to the neuraxis (typically in 180-cGy fractions) and 54 Gy to the primary tumor volume.821,835,839
Unlike any other CNS malignant tumors, germ cell tumors of the CNS have a model for chemotherapeutic treatment based on histologically similar systemic disease. Outside the CNS, chemotherapeutic regimens based on cisplatin have been very effective in treating disease. This experience, coupled with the mixed response to radiation therapy, prompted the exploration of chemotherapy for CNS germ cell tumors. For patients with germinoma, chemotherapy has been added to decrease deleterious late effects of radiation therapy. For the nongerminomatous tumors, chemotherapy has been added to try to improve cure rates.
There appears to be no question that germinomas are chemotherapy-sensitive tumors. Regimens that use cisplatin, carboplatin, or cyclophosphamide, along with vinblastine or vincristine, bleomycin, and etoposide, are capable of producing complete and partial response rates as high as 90% in newly diagnosed patients.807,824,830,840,841,842,843 In most of these series, patients have had irradiation as part of their treatment;

thus, the influence of chemotherapy on survival remains uncertain. Administration of chemotherapy to patients with suprasellar lesions who also commonly have diabetes insipidus may be difficult. Fluid and electrolyte imbalances may result in clinically significant hyponatremia or hypernatremia and dehydration. The use of vasopressin infusions during such chemotherapy may help to avert some of these problems.844
With the demonstrated responsiveness of germinomas to chemotherapy, the current debate centers on the optimal balance of chemotherapy and radiation therapy. Recent studies have examined the use of chemotherapy with either reduced-dose radiation therapy or without radiation altogether.807,824,843 In the largest series wherein chemotherapy was used alone for patients with germinomas, high rates of response were demonstrated to chemotherapy.807 Despite this, however, 22 of 45 patients ultimately relapsed, a number higher than would have been anticipated after radiation therapy alone. Although a large proportion of the patients who experienced relapse was successfully treated with additional chemotherapy or irradiation, the 2-year overall survival rate for patients with germinoma was only 84%.807
Other, smaller studies have examined the use of chemotherapeutic regimens followed by radiation therapy at doses reduced to 30.6 and 40.0 Gy.824,843 Response rates to chemotherapy have been high, and survival rates in both series are 100%, with median follow-up periods exceeding 32 months. Similar approaches with even lower doses of radiation are being studied by the various cooperative groups. The use of chemotherapy and radiation therapy for intracranial germinomas is rational and effective, although the optimal schedule must be determined. Treatment with chemotherapy alone is not recommended outside of a clinical trial setting.
Similar chemotherapeutic regimens have been applied to nongerminomatous germ cell tumors, with encouraging results.111,821,831,835,838,845,846,847 In these studies, complete and partial response rates to chemotherapy have approached 80%; survival rates several years after diagnosis have similarly ranged from 48% to 80%. Although all patients were treated with radiation as well, the fields varied. Relapse rates appear to be higher in the patients treated with involved fields only. Therefore, craniospinal irradiation for all patients with nongerminomatous germ cell tumors seems advisable.
For a discussion of chemotherapy for pineoblastoma, the reader is referred to the section in which sPNETs are discussed.
Craniopharyngiomas account for between 6% and 9% of all primary CNS tumors in children. These lesions exhibit a bimodal age distribution, with one peak during childhood at approximately 8 to 10 years of age and a second peak in middle age.848 This tumor rarely is detected before age 2 years.602,849 No gender predilection has been noted. Although these lesions are predominantly suprasellar tumors, which involve the pituitary stalk and hypothalamus, they may occur within the sella turcica or third ventricle as well.
Figure 27.30 Photomicrograph of typical epithelium found in a craniopharyngioma showing the basisquamous character with incarcerated keratin. Note the honeycombed character of the epithelium in areas (hematoxylin and eosin, ×250).
Pathology and Patterns of Spread
Craniopharyngiomas in children are thought to arise predominantly from pharyngeal cell rests left from the embryonic hypophysiopharyngeal duct that connects the infundibular bud with the stomodeum. Although, in adults, these tumors may result from neoplastic transformation of cell rests within the pituitary gland that have undergone squamous metaplasia,850 this mechanism is less likely in children.489 Grossly, these tumors are smooth, lobulated masses with both solid and cystic components. The cyst contents may range from gelatinous to viscous oily fluid rich in cholesterol crystals. Rupture of a cyst into the CSF may cause an intense chemical meningitis. Calcification is frequently apparent. Both the cystic lining and the solid portions of the tumor are characterized by squamous epithelium, usually with some evidence of keratinization (Fig. 27.30).
Although craniopharyngioma is a histologically benign tumor composed of well-differentiated tissue, it may have a malignant clinical course because of its location and its propensity to infiltrate surrounding normal structures. A thick glial layer may encase the tumor, and small islands of epithelial tumor arising within this gliotic scar can extend into adjacent tissues. The tight adherence of this layer to surrounding tissue can make complete resection difficult and hazardous.
Clinical Presentation
Childhood craniopharyngiomas often manifest with short stature, symptoms of increased ICP, delayed puberty, vision loss, and neurobehavioral abnormalities. Hydrocephalus is observed in approximately 50% of children as a result of obstruction of the third ventricle and the foramen of Monro by superior tumor extension. Because of slow tumor growth, papilledema is less common than optic pallor. Visual field defects of various degrees of severity occur in 50% to 90% of patients, homonymous hemianopsia and bitemporal hemianopsia being the most frequent defects encountered.849,851 Despite the fact that many children show evidence of vision

loss on examination, only approximately 25% present with complaints of visual deterioration.852
Figure 27.31 Axial (A) and sagittal (B) T2-weighted magnetic resonance imaging of mixed-density craniopharyngioma with foci of calcification (black).
Various neuroendocrine deficits are present in as many as 90% of patients at diagnosis, decreases in GH and resultant growth delay being the most common findings.170,851 Diabetes insipidus is seen in fewer than 10% of children in whom craniopharyngiomas are diagnosed in the modern era and, if present in a child with a suprasellar lesion, should raise concern about the possibility of a germ cell tumor or histiocytosis.
Differential Diagnosis and Evaluation
The differential diagnosis of craniopharyngioma includes intrinsic hypothalamic gliomas, large chiasmal gliomas, Rathke’s cleft cyst, and suprasellar germ cell tumors or teratomas. Plain skull radiographs, commonly used before the advent of CT and MRI, often show an enlarged or distorted sella with suprasellar tumor calcification. The CT scan characteristically demonstrates a partially cystic, low-density, contrast-enhancing lesion with calcification.
MRI defines the solid and cystic nature of the tumor, its extent, and its relation to adjacent structures better than does any other modality (Fig. 27.31).
Owing to the high incidence of clinical and subclinical neuroendocrine deficits at diagnosis, a thorough evaluation of the hypothalamic-pituitary axis should be undertaken preoperatively. Secondary abnormalities of adrenal function and of the regulation of fluid and electrolyte balance, in particular, can lead to serious perioperative problems if not anticipated. Neuroendocrine evaluations should be repeated postoperatively and periodically thereafter for at least 1 year, because hormonal deficits often increase postoperatively and may take several months to stabilize fully.853
Prognostic Considerations
The extent of tumor resection has been an important factor in series in which the initial treatment has consisted of surgery alone. Patients with totally excised tumors have had considerably better survival rates than those managed by biopsy alone or by subtotal resection.854 However, several studies have shown that the combination of subtotal resection and radiation therapy achieves survival results that rival those obtained with attempted GTR.245 The purported prognostic significance of tumor size probably is not an independent variable but rather is related to the ease and extent of resection. Although the data show only a trend, patients with purely cystic lesions appear to survive longer than those with solid or mixed solid and cystic tumors; in addition, children older than 5 years seem to have a better prognosis than do younger patients.854,855
Preoperative and perioperative considerations for operation on tumors in the region of the pituitary gland have been reviewed earlier in this chapter. Those considerations specific to craniopharyngioma are discussed here.
Because tumor resection may cause or exacerbate endocrine deficiencies, the management of these problems must begin preoperatively and continue through the postoperative period. Stress doses of hydroxycorticosteroids (e.g., hydrocortisone, 100 mg/per m2 intravenously followed by 25 mg per m2 every 6 hours) are administered before, during, and immediately after the surgical procedure, often in addition to dexamethasone, which is used to reduce peritumoral edema. Doses then are tapered but, if postoperative endocrine testing demonstrates a need for long-term steroid hormone replacement, then hydrocortisone is continued at maintenance levels.
Hydrocephalus in patients with craniopharyngiomas generally resolves after the tumor has been resected. If the hydrocephalus is severe, a ventriculostomy can be inserted before the tumor resection is begun and can be removed within several days of surgery if the operative procedure opens the CSF pathways. If the hydrocephalus persists, a ventriculoperitoneal shunt is inserted.

The extent of surgical removal to be attempted is a matter of intense debate. Some authors strongly recommend radical surgery in all cases,856,857 whereas others suggest partial resection followed by local irradiation.245 The debate centers on the prognostic advantage gained by complete resection balanced by the morbidity associated with such a procedure. With the use of microsurgical techniques, a radiologically complete resection can be achieved in 60% to 90% of children.852,856,857,858,859,860,861 Under these circumstances, the likelihood of long-term PFS is 80% to 90%. However, significant neurologic morbidity, memory and cognitive dysfunction, and appetite and neurobehavioral disturbances are encountered in 10% to 30% of patients, mortality ranges from 0% to 5%, and panhypopituitarism develops in 80% to 90% of patients.849,851,856,857,862 The operative morbidity is lower in children operated on by neurosurgeons who perform such procedures frequently.245 Although morbidity may also be lessened somewhat by more limited resections, the likelihood of long-term disease control is substantially lower than with complete resection. Without radiation therapy, the vast majority of subtotally resected tumors progress within 2 to 5 years.863,864 The recurrence rate is diminished significantly with the use of postoperative external irradiation,245,865 although some studies indicate that the results, in terms of PFS, remain inferior to those achieved with total resection.863 Although repeat microsurgical resection is feasible in patients with recurrent tumor, morbidity and mortality may be substantially higher than at the primary operation.856 The primary cause of death in these patients is either recurrent tumor or chronic neuroendocrine problems.866
Craniopharyngiomas are extraaxial tumors, tenaciously attached to surrounding structures, such as the optic chiasm, hypothalamus, and vessels of the circle of Willis. These characteristics impose practical limits on a surgeon’s attempt at tumor resection. Regardless of the degree of surgical resection intended at the outset, the usual cases are approached via a subfrontal or trans-sylvian exposure, working between the optic nerves, carotid arteries, and third nerves, or through the lamina terminalis. Large tumors extending to the roof of the third ventricle can also be approached through the corpus callosum. Other approaches may be used, depending on tumor extent and location and a growing trend is to tailor the treatment strategy based on tumor size and composition. Tumors with a sizable solid component (greater than 3 cm) are treated microsurgically, generally via a transcranial approach; selected lesions that arise within and expand the sella turcica may be amenable to transsphenoidal resection.867 Thin-walled cystic lesions can be treated using intracavitary techniques; if small residual solid components of the tumor subsequently enlarge, stereotactic radiosurgery or microsurgical techniques can be used. The management of small, solid tumors (less than 3 cm in diameter) remains particularly controversial, with most groups favoring microsurgical resection and others employing stereotactic techniques.
Radiation Therapy
There is considerable controversy regarding primary irradiation versus primary surgery for craniopharyngiomas. Long-term results with limited surgery (biopsy, limited decompression, or planned subtotal resection) followed by external beam irradiation approximate the same 80% to more than 90% rate of disease control reported after imaging-confirmed complete surgical resection.245,470,868,869,870,871,872 Similar results are reported following irradiation for disease residual after attempted total resection.862,868,869,873,874 Although most institutions stress initial irradiation for documented disease residual, there is limited experience suggesting adequate outcome when irradiation is deferred until documented postoperative progression.875,876
Contemporary irradiation requires image-guided approaches, using either 3D-CRT or IMRT to target volumes closely approximating visible tumor. The role for limited surgical resection to decompress cystic components or remove peripheral tumor resulting in more limited radiation target volumes is yet under study.869 The target volume dose is 54 to 56 Gy.
Although a controlled trial has not been done, multiple comparisons strongly suggest that patients treated with incomplete tumor resection followed by irradiation have less neuroendocrine dysfunction and fewer serious sensory, motor, and visual deficits than do those who have undergone aggressive attempts at complete tumor resection. These patients may also have an improved level of function and better quality of life than patients treated with radical surgery alone.849,862,869,870 Additionally, neuropsychological function is reportedly better preserved in the combined-therapy group despite the known detrimental effect of irradiation.862,865,871 Predominantly cystic lesions, which account for 20% to 30% of tumors, may be treated with intracystic instillation of yttrium-90 or phosphorus-32 to deliver high-dose intralesional therapy.275,276,525,877 Involution of the cyst is achieved in more than 80% of patients;12,525,877,878 morbidity and mortality are lower than with microsurgical resection, and the recurrence rate is comparable to that achieved with attempted total microsurgical removal.877,878 Because of the limited tissue penetration of the radiation from these β-emitting isotopes, radiation exposure to the adjacent hypothalamus and optic nerves is often minimized; focal high-dose regions adjacent to the chiasm can result in visual complications.869 After intracavitary therapy, small, residual solid-tumor components can be treated with stereotactic radiosurgery.281,877
Chemotherapy has no established role in the treatment of craniopharyngiomas. An anecdotal response to a vincristine, BCNU, and procarbazine combination has been described in one patient.879 Intracystic administration of bleomycin has also led to response and significant second remissions in some patients with recurrent disease,880 but concerns have been raised regarding local toxicity. The use of systemic interferon-α has also been evaluated in a phase II study, with some favorable responses.881
Choroid plexus neoplasms constitute between 1% and 4% of brain tumors in children. Seventy percent of these occur during the first 2 years of life; the median age at diagnosis ranges from 10 to 32 months in recent series.240,241,242,882,883,884,885,886 CPPs, benign tumors treated only surgically, outnumber the

malignant counterpart, choroid plexus carcinoma, by nearly four to one.240,882 Tumors arise from the lateral ventricles approximately 75% of the time, from the fourth ventricle and cerebellopontine angle in 15% of cases (Fig. 27.32), and from the third ventricle in 10% of affected children.883
Figure 27.32 Left: Sagittal T1-weighted pre- and post-Gd imaging of a well-defined posterior fossa mass within the fourth ventricle. Right: It demonstrates avid enhancement following the administration of contrast, typical for a choroid plexus papilloma, which in this case was not causing hydrocephalus.
Pathology and Patterns of Spread
Choroid plexus neoplasms generally arise as functioning intraventricular papillomas capable of secreting CSF. Grossly, CPPs resemble a soft coral, with fronds of tumor attached to a pedicle that floats in the CSF. Microscopically, these tumors are similar to normal choroid plexus and have cuboidal or columnar epithelium and a well-preserved epithelial-stromal border overlying fibrovascular septa (Fig. 27.33). Their neoplastic nature is reflected in the heaping and redundancy of the epithelial component. These tumors tend to be slow growing and, because of their intraventricular location, often reach a size of 60 to 70 g before they are diagnosed. Fewer than 5% are bilateral.
Figure 27.33 Choroid plexus papilloma. The photograph illustrates the luxurious papillary structures composed of a loose fibrovascular core covered by a single layer of cuboidal-columnar epithelium (hematoxylin and eosin, ×250).
The choroid plexus carcinoma (CPC), a more aggressive and anaplastic tumor, accounts for up to 40% of choroid plexus neoplasms.504,883,885 This tumor has lost the well-differentiated papillary structure and the epithelial-stromal border of the CPP (Fig. 27.34). It is a hypercellular tumor with pleomorphic cells, frequent mitoses, and foci of necrosis.887 Both papillomas and carcinomas are capable of leptomeningeal dissemination. In CPPs, the clinical behavior and histology of the isolated and frequently noted deposits are benign, and symptoms are uncommon. Conversely, diffuse and aggressive leptomeningeal spread occurs in CPCs.
Prognostic Considerations
Tumor histology and degree of resection are the primary prognostic factors for choroid plexus neoplasms. The long-term recurrence-free survival after complete resection of CPP approaches 100%.883,888,889,890 Even less-than-complete resection is associated with long periods of PFS. The outcome is less favorable in patients with CPC,241,242,891,892 because these

lesions invade the brain parenchyma and are extremely vascular, making complete resection difficult. In addition, these tumors often disseminate within the CSF. Nevertheless, in reviews of experience with CPC, complete resection of disease appears to be the single variable that most affects long-term survival.240,241,242,884,885 In fact, GTR of CPC may be curative by itself in a proportion of children.504
Figure 27.34 Choroid plexus carcinoma. The papillary character, which is partially retained in tissues on the left side of the field, has been lost in the portion of the tumor on the right side. Note pseudostratified epithelium forming irregular glandular structures on the left and diffuse epithelial growth on the right (hematoxylin and eosin, ×250).
Surgical excision is the primary mode of therapy for both CPP and CPC. These lesions most commonly arise in the trigone; tumors in this location usually are approached through a posterior temporoparietal craniotomy, through a cortical sulcus. Tumors of the anterior third ventricle and the body and frontal horn of the lateral ventricle may be approached transcallosally or transcortically through the middle frontal gyrus. Fourth ventricular tumors are approached by the suboccipital route. Intraventricular tumors outside the posterior fossa may be more easily removed if the ventricles are large; for this reason, preoperative shunts usually are not inserted in patients who are otherwise clinically stable. Contemporary surgical morbidity and mortality rates are less than 20% and 5%, respectively. Complete resections are possible in approximately 80% of patients and is much more easily achieved with CPPs than CPCs, in which brain invasion and vascularity may preclude tumor removal.883 Even after complete tumor removal, persistent hydrocephalus requiring a CSF shunt is present in up to 60% of patients.893 If the postoperative MRI scan demonstrates resectable residual tumor, reoperation is indicated. Appreciation for the importance of complete surgical resection has provided an impetus for efforts to perform second-look surgery with initially unresectable CPCs in those children whose incompletely resected tumors persist but decrease in size after a trial of neoadjuvant chemotherapy.
Radiation Therapy
Postoperative radiation therapy frequently is used to treat CPC, especially if the resection is incomplete or if there is evidence of leptomeningeal dissemination of disease. Although the survival times of some irradiated patients may be marginally better than those of nonirradiated patients, such results are not entirely separable from results with surgery alone. No randomized trial has yet tested the benefit of radiation therapy.
Surgery alone is usually sufficient for cure of CPP; chemotherapy has no role in the treatment of this tumor. As with other uncommon malignant pediatric CNS tumors, however, the role of chemotherapy is difficult to define in the treatment of CPC. Numerous reports of small numbers of patients collectively demonstrate that CPC can respond to different chemotherapeutic regimens at initial diagnosis or following relapse. The regimens given to the highest number of patients were based on platinum240,882,883 or cyclophosphamide504 or included multiple agents.241 However, whereas most patients with CPC treated with GTR and chemotherapy appear to be long-term survivors, cure has been achieved with GTR alone. Furthermore, the majority of children whose tumors are less than completely resected and who also are treated with chemotherapy ultimately die of their disease.240,241,885 Anecdotal reports, however, suggest that chemotherapy may render CPC less vascular and infiltrative and potentially amenable to complete removal, even if the initial attempt was unsuccessful.241,882,891 Thus, chemotherapy may potentially contribute to higher chances of survival. Because the number of children with choroid plexus neoplasms is so low, international collaborative clinical trials will be needed to test hypotheses related to outcome of therapy.
Intrinsic tumors of the spinal cord make up 1% to 10% of pediatric CNS tumors.894,895,896,897,898 These tumors occur throughout childhood, and the median age at diagnosis is 10 years. Male patients are slightly more commonly affected than female patients, in a 1.3:1.0 ratio. Patients with neurofibromatosis appear to have a higher incidence of spinal cord tumors, as they do with other astrocytic neoplasms.
Pathology and Patterns of Spread
In children, up to 70% of intramedullary tumors are astrocytomas. The remaining 30% are ependymomas (10%), other glial neoplasms such as oligodendroglioma and gangliogliomas (10%), and malignant gliomas (10%). Histologically, these tumors are indistinguishable from their intracranial counterparts. Large cysts, both within the tumor and at the superior and inferior margins, are common. In as many as 60% of cases, extensive involvement by tumor and cysts may be present. Slow and contiguous extension across several vertebral segments, with compression and effacement of normal tissues, is the usual mode of growth.899,900,901 Leptomeningeal dissemination has been reported in as many as 58% of patients with high-grade tumors, but it is uncommon in patients with low-grade tumors.901,902 The presence of multiple discrete tumors is associated with neurofibromatosis. Tumor location in the spinal cord appears to be random, with the incidence in each anatomic region (cervical, thoracic, lumbar) roughly proportional to the length of that region. The only exception is the myxopapillary ependymoma, which has a predilection for the conus medullaris and filum terminale.
Prognostic Considerations
Spinal cord tumors in children are rare occurrences. The low number of patients and relative lack of inclusion in clinical trials with prescribed treatment approaches make the identification of published prognostic factors difficult. Bouffet et al.898 reviewed the experience of 13 French treatment centers with spinal cord astrocytoma and found high-grade histology and short duration of presenting symptoms to be associated with poorer survival.898 From review of other series, low-grade

lesions are compatible with long-term survival, whereas this appears not to be the case with high-grade lesions. Although investigators have suggested that complete tumor removal may be associated with longer survival and less frequent local recurrences, the degree of resection has not been associated with outcome in several studies. However, such a determination is problematic in view of the small sizes of the study cohorts.901,903,904,905 Ependymomas may be an exception to this generalization, in that patients who undergo total tumor resection have fewer recurrences than do those who undergo incomplete tumor resections.906
Complete surgical resection is difficult for astrocytoma because a distinct tumor–cord interface often is absent, but extensive subtotal resections may be performed in most instances. Ependymomas are associated with a clearer cleavage plane and can usually be resected completely. Intramedullary tumors usually are approached by an osteoplastic laminectomy, removing as a single unit all laminae covering the solid portion of the tumor. Replacement of the lamina after surgery not only helps to protect the spinal cord but also may diminish the risk of subsequent spinal deformity. Operative mortality and the amount of neurologic recovery are proportional to the severity of preoperative dysfunction and to the definition of the tumor–cord interface. In one report of 69 patients undergoing operations for intramedullary tumors, at a mean follow-up of 54 months, 17% were better than they had been preoperatively, 56% were unchanged, and 31% were worse.907 Postoperative orthopedic follow-up and monitoring for spinal deformity are important. In 25% to 40% of children, the development or progression of such deformity occurs within a mean of 3 years.
Radiation Therapy
No controlled trial of radiation therapy has been conducted in patients with intramedullary tumors, and evidence for its utility is inferred from the treatment of similar tumors in other CNS locations. As with low-grade lesions in the cerebrum, irradiation may be deferred for incompletely resected low-grade tumors in very young children or when there is expectation that subsequent surgery might be truly extirpative. With significant disease residual, common practice often includes local postoperative irradiation to dose levels of 50 Gy.592,908,909,910 Data regarding irradiation versus observation for spinal ependymomas has been summarized in the section on ependymomas. For malignant gliomas, postoperative irradiation is standard therapy.899,901,903,905,906,909 Target volumes typically extend 2 to 5 cm beyond the lesion identified by T1 imaging; dose levels to the primary tumor site are often in the 50- to 54-Gy range.906,911
The overall survival rates for LGAs with various degrees of resection and postoperative radiation therapy are 66% to 70% at 5 years, 55% to 73% at 10 years, and 67% at 20 years.592,900,903,905,906,909,911 Durable control of anaplastic astrocytoma or glioblastoma has been elusive.906,911 For patients with ependymomas, survival rates of 50% to 100% at 5 years and 50% to 70% at 10 years are reported, and local recurrences are relatively high in patients with subtotally resected tumors.592,899,903,908,909,910
Tumors of the spinal cord have been treated as their histologic counterparts in other parts of the brain. Chemotherapy has been employed for high-grade lesions at diagnosis, for recurrent low-grade lesions, and in very young children in whom the avoidance of radical surgery or radiation therapy has been desired.894,896,912 The largest of these series involved 13 children with high-grade astrocytoma of the spine who were treated with 8-in-1 chemotherapy along with radiation therapy, postoperatively. The response to preradiation chemotherapy was not measurable in three patients, complete in one patient, partial in two, stable in four, mixed in two, and progressive in one. After completion of therapy, with the time of median follow-up not stated, 2 of the 13 patients were alive without disease at the report, and 5 were alive with disease. Five-year PFS and survival rates were 46% and 54%, respectively.896
Until biologic factors or other clinical trial results indicate otherwise, it seems rational to use chemotherapy for tumors of the spinal cord only as would be used for tumors of like histologies in other areas of the brain.
Mortality rates for children with brain tumors have declined less rapidly than those of other cancers.913 Nevertheless, the 5-year overall survival for children with brain tumors, aside from the favorable LGAs, has improved to 60%.5 Despite successful therapy, children with CNS tumors may have physical, cognitive, neurologic, endocrinologic, or other deficits as direct sequelae of their tumor or as a result of therapy. As a whole, these patients function at lower intellectual, social, and physical levels than their peers, which in turn leads to a diminished quality of life. The eventual magnitude of these problems may be greatest in patients who are the youngest at diagnosis (see Chapters 13, 15, and 49).
The acute and late effects amongst pediatric brain tumor patients arise from several sources. Before diagnosis, the tumor mass distorts and even destroys normal brain tissue and increases ICP, which may be associated with hydrocephalus. Surgical trauma, postoperative meningitis, shunt infection, or repeat surgery can cause some degree of irreversible neurologic damage.914,915 Likewise, chemotherapy may be capable of producing encephalopathy.916 Radiation therapy has been implicated as the chief cause of many adverse sequelae, which are listed in Table 27.11.
The subacute effects of irradiation, apparent 2 to 6 months or more after therapy, include intralesional necrosis or edema, perilesional white matter changes, or alterations in adjacent white matter or cerebral nuclei (e.g., basal ganglia) that may be silent or accompanied by site-specific symptomatology.285
Cognitive impairment is among the most devastating problems of the child treated with radiation therapy.289,903,917,918,919,920,921

Neurocognitive deficits are inversely related to age at the time of irradiation and correlate directly with both site (supratentorial irradiation ≫ posterior fossa irradiation) and radiation dose.289,290,473,917,919,921 Detailed prospective studies indicate that radiation-related cognitive problems include relative difficulty with selective attending and acquisition of new knowledge; over time, the relatively mild difference in intellectual development equates to significant declines in IQ for children younger than 5 to 8 years when treated, with more subtle changes in older children.289,922,923
Syndrome Onset Cause Clinical Manifestation Treatment and Outcome
Somnolence syndrome 4–8 wk after RT Whole-brain or large-field supratentorial RT 7–14 days of lethargy, malaise, nausea, vomiting, and anorexia Can respond to dexamethasone; usually self-limited
Radiation necrosis Usually 3 mo–3 yr after RT Idiosyncratic, 0.1%–1.0% with conventional RT; increased with higher dose per fraction or total dosage, usually after ≥5,500 cGy Focal neurologic dysfunction, seizures, symptoms of increased intracranial pressure, coma, or death Focal areas resectable; dexamethasone if unresectable or small
Mineralizing microangiopathy with dystrophic calcification Unknown; 9 mo–several years after RT ≥2,000 cGy; perhaps potentiated by intrathecal methotrexate or cytarabine None or possible headaches, seizures, strokes Unclear (often an autopsy finding)
Moyamoya disease 6 mo–15 yr after RT, particularly in children ≤3yr Arterial occlusion after ≥4,000 cGy RT; associated with neurofibromatosis type 1 Headaches, seizures, transient ischemic attacks, strokes, progressive mental deterioration Some stabilize from spontaneous collateral vessel formation; arterial bypass surgery may improve outcome
Endocrinologic dysfunction Biphasic; early dysfunction and then fixed damage 1–5 yr after RT RT to hypothalamus, pituitary gland, thyroid gland, or gonads Growth failure, thyroid deficiency, gonadotropin deficiency Requires hormonal replacement; static damage
Neuropsychological damage Increases with time from RT; unclear whether plateaus Any brain RT; increases with larger total dosage or volume or younger age; also complicated by the effects of the tumor itself, surgery, and chemotherapy Cognitive deficits, learning disabilities, and behavioral abnormalities Requires remediation and special education; may progress over time
Secondary brain tumor 5–25 yr after RT ≥1,800 cGy RT, perhaps potentiated by chemotherapy and genetic predisposition Meningioma, high-grade astrocytoma, or sarcoma Often poor prognosis
RT, radiation therapy.
Full-brain irradiation to dose levels of more than 30 to 36 Gy in infants and very young children has been associated with severe intellectual deficits; such children have median IQ scores of 60 to 65 at 5- to 7-years posttherapy, indicating little likelihood of ultimate independent function as adults.919 Sophisticated local radiation techniques, particularly when tumors are confined to the posterior fossa, have shown rather stable neurocognitive function even in children averaging only 2.8 years of age.289 Detailed assessment of neuropsychological function usually identifies multiple areas of damage in information processing. Some studies have found specific neurocognitive deficits in attention, memory, coordination, fine motor speed, visual motor processing, mathematics, and spatial relations.290,917,919,924,925
Identification of neurocognitive deficits has recently been followed by both pharmacologic and rehabilitative interventions that seem to ameliorate the learning and memory difficulties associated with declines in IQ.289 Methylphenidate has been shown to reverse attentional problems and improve memory/learning at least in time-limited measures.926 Ongoing studies address prospective interventions during and immediately after irradiation.
As another consequence of radiation therapy to whole brain, hypothalamic-pituitary region, or spine, growth failure occurs commonly among brain tumor patients. Irradiation along the spinal axis retards the growth of the vertebral column and spinal cord, leading to a child with a short trunk and disproportionately longer extremities.916 Spinal irradiation alone has been associated with a decrease in eventual height of 9, 7, and 5.5 cm when administered at ages 1, 5, and 10 years, respectively.927 A larger contribution to decreased stature stems from impaired GH secretion. Noted in up to 25% of brain tumor children even before irradiation, deficits appear in almost all children within 1

to 2 years after irradiation that includes the hypothalamic-pituitary axis.304,928 The use of concomitant chemotherapy may increase further the severity of growth retardation.929 The effect of precocious puberty prematurely fusing bony epiphyses also can contribute to short stature.
Hormone replacement therapy should be considered in patients whose growth velocity has declined and who do not respond to provocative GH testing. The decision regarding whether and when to initiate replacement remains emotional and controversial for parents and physicians, most North American neurooncology groups recommend initiation of GH after 1 or more years of disease control or stabilization.930
Although the hormone is mitogenic, it does not appear to produce any increased risk for tumor recurrence.931,932 Other neuroendocrine deficits may also be seen. Primary, or less commonly secondary or tertiary, hypothyroidism may occur in more than half of patients as a result of irradiation to the thyroid gland or the hypothalamic-pituitary axis, respectively.916,933 Puberty may occur prematurely or at normal onset but only seldom is delayed. Abnormalities of gonadotropin or corticotropin secretion may be less common.933 Male patients appear to be less at risk than female patients for gonadal dysfunction secondary to spinal irradiation, but the synergistic effect from commonly used drugs such as cyclophosphamide and nitrosoureas, known to affect oogenesis and spermatogenesis, has not been estimated.916
High-frequency, sensorineural hearing loss is another complication that is frequent among brain tumor survivors, most commonly caused by repeated administration of cisplatin. Cisplatin results in rather immediate, dose-dependent reduction in high frequency hearing; if continued, more pronounced hearing deficits occur into the speech range. Irradiation is associated with a relatively late, pronounced, typically unilateral sensorineural hearing loss that occurs in 10% to 15% of children after doses of more than 50 to 54 Gy. Highly conformal radiation therapy may reduce the risk of ototoxicity.299
The optimal management of the potential acute and long-term needs of children with tumors of the CNS requires a multidisciplinary team. This team should include physicians (oncologists, neurologists, neurosurgeons, radiation oncologists, ophthalmologists, endocrinologists, cancer geneticists, physiatrists) and nurses as well as an occupational therapist, physical therapist, child life worker, educational psychologist, audiologist and social worker. The team and family must remember that the effects resulting from the tumor or its treatment may not occur or become fully manifest until several years after completion of therapy. Consequently, yearly physical examinations and neuropsychological evaluations should be performed for at least the first 5 years after therapy. Children who have intellectual impairments require further evaluation so that proper educational intervention can begin, if not already implemented. Educational interventions have been associated with improvements in spelling and reading, particularly when written feedback was provided to parents and schools, even years after therapy.918
Other routine evaluations should be tailored to the patient’s tumor and prior therapy. An evaluation by an occupational therapist may be indicated to assess fine motor and visual skills. Endocrinologic evaluation should be based on the location tumor location and treatment. For the majority of patients this will include an annual thyroid function studies (for a minimum of 5 years) to monitor for subclinical manifestations of hypothyroidism. Appropriate thyroid replacement is critical to growth, learning, and prevention of thyroid tumors from persistently elevated thyroid-stimulating hormone. Longitudinal audiometry should also be considered for at least several years, particularly in patients who received platinum analogues or radiation to the posterior fossa.
Long-term management of the child with a brain tumor also includes surveillance for disease relapse. Surveillance of brain tumor patients, unlike patients with other childhood cancers, is indeed a chronic issue. Progression or recurrence of tumors such as ependymoma or the LGAs frequently does not occur until 3 to 5 or more years from diagnosis. For MB, with a median time to relapse of 18 to 24 months, recurrence may occur 8 or more years after diagnosis.934 The ideal surveillance modalities and schedule for all the diverse tumor pathologies is unclear. Despite fair agreement that serial clinical examination is important not only to detect signs suggestive of relapse but also to monitor for the already mentioned sequelae of treatment, controversy surrounds the timing and need for surveillance MRI. For malignant brain tumors, the CCG has recommended surveillance MRI every 3 months during the first year after diagnosis, every 4 months during the second year, biannually during years 3 and 4 from diagnosis and, finally, annual to biennial scanning until 12 years after diagnosis.935 Other investigators have questioned whether neuroimaging surveillance affects outcome, as the salvage rate for relapsed tumors such as MB is low.936,937 Further data are forthcoming from the POG experience to help answer this question. Even less clear is the decision regarding when to stop routine MRI scanning, as secondary brain tumors, particularly meningiomas, gliomas, and sarcomas, have been reported 5 to 25 years after treatment of the original tumor.916 Most recommendations today are based on retrospective analyses. Prospective evaluation of neuroimaging surveillance schedules appropriate to the heterogeneous CNS tumor pathologies is needed.
Although prevention of these deleterious sequelae is beginning to receive attention, judging the efficacy of current interventions is difficult. Reduction of radiation effects on the CNS can be achieved by diminishing radiation volume (through use of image-guided 3D-CRT or IMRT), dose, or the use of irradiation (particularly in infants and very young children who require large treatment volumes). Several factors other than radiation therapy can influence neurocognitive and neuroendocrine function, and chemotherapy itself may have deleterious effects on the developing nervous system.914,915,920,938 Furthermore, prolonged chemotherapy with alkylators and etoposide among infants with brain tumors has been associated with an excess risk of second malignancies.481 Other new strategies to mitigate damage have yet to be tested. Conformal techniques can better focus the radiation delivered, such as in the posterior fossa where, in the past, local field treatment still exposed the hypothalamus and hippocampus just beyond the posterior clinoid

processes. The chemoprotectant amifostine has yet to be widely studied with cisplatin administration.
Organization Phone Number Web Site
American Brain Tumor Association 800.886.2282
Brain Tumor Foundation of Canada
Children’s Cause for Cancer Advocacy 301.562.2765
Children’s Oncology Group 800.458.6223
Childhood Brain Tumor Foundation 301.515.2900
Children’s Brain Tumor Foundation 212.448.9494
National Brain Tumor Foundation 800.934.CURE
National Cancer Institute 301.496.6641
Musella Foundation for Brain Tumor Research 516.295.4740
Pediatric Brain Tumor Consortium 301.496.6641
Pediatric Brain Tumor Foundation of the United States 800.253.6530
The Brain Tumor Society 800.770.8287
United Kingdom Children’s Cancer Study Group
Because most patients with neurocognitive deficits do not have observable histopathologic changes, the location and pathogenesis of their problems are not well established. Thus, understanding the ways in which radiation and chemotherapy alter axonal growth, dendritic arborization and pruning, synaptogenesis, and myelination, all of which occur in childhood and adolescence, is presumably critical to understanding the development of these problems and remains largely unknown. Evidence suggests that MRI is capable of demonstrating areas of treatment-related white matter abnormalities not visible on CT that may correlate with the degree of clinical neurologic compromise, particularly in IQ, factual knowledge, and verbal and nonverbal thinking.473 Prospective evaluation with new MRI techniques and other novel neuroimaging modalities may improve understanding of the location and dynamics of damage associated with various therapies.
Imprecise methods, small patient numbers, and limited funding for studying patients with CNS tumors have been barriers to our knowledge of late effects. Closer collaboration with schools for assessment and intervention might help. In the future, comprehensive batteries that take the entire child into account are needed. Given that executive function, memory, pragmatic language, and other cognitive functions change over time, such evaluations should steer assessment away from IQ and toward developmental neuropsychological models. More precise definitions of disability and quality of life are needed. Newly validated instruments that measure quality of life should be implemented. Assessments will also have to consider practical outcomes, such as the ability to hold a job, drive a car, manage finances, or live independently.939 Interventional trials of medications that improve arousal, attention, and memory, such as modafinil, methylphenidate, or donepezil, are needed also.
For more information about clinical trials or other support services for children with brain tumors, a list of resources is provided in Table 27.12.
1. Gurney JG, Smith MA, Bunin GR. Cancer incidence and survival among children and adolescents: United States SEER program 1975–1995. NIH Pub No 99-4649. Bethesda, MD: National Cancer Institute, SEER Program, 1999:51–63.
2. Gurney JG, Wall DA, Jukich PJ, et al. The contribution of nonmalignant tumors to CNS tumor incidence rates among children in the United States. Cancer Causes Control 1999;10:101–105.
3. Grovas A, Fremgen A, Rauck A, et al. The National Cancer Data Base report on patterns of childhood cancers in the United States. 1997;80:2321–2323.
4. Bleyer WA. Epidemiologic impact of children with brain tumors. Childs Nerv Syst 1999;15:758–763.
5. Jemal A, Clegg LX, Ward E, et al. Annual report to the nation on the status of cancer, 1975–2001, with a special feature regarding survival. Cancer 2004;101:3–27.
6. Ries LA, Kosary CL, Hankey BF, et al. SEER cancer statistics review, 1973–1994. NIH Pub No 97-2789. Bethesda, MD: National Cancer Institute, SEER Program, 1997.
7. Smith MA, Freidlin B, Ries LA, et al. Trends in reported incidence of primary malignant brain tumors in children in the United States. J Natl Cancer Inst 1998;90:1269–1277.
8. Legler JM, Ries LA, Smith MA, et al. Cancer surveillance series [corrected]: brain and other central nervous system cancers: recent trends in incidence and mortality. J Natl Cancer Inst 1999;91:1382–1390.
9. Bunin G. What causes childhood brain tumors? Limited knowledge, many clues. Pediatr Neurosurg 2000;32:321–326.
10. Preston-Martin S. Epidemiology of primary CNS neoplasms. Neurol Clin 1996;14:273–290.
11. Kleihues P. World Health Organization classification of tumours. In: Kleihaus P, Cauance WK, eds. Pathology and genetics of tumours of the central nervous system. Lyon, France: International Agency for Research on Cancer Press, 2000:208–241.
12. Evans G, Burnell L, Campbell R, et al. Congenital anomalies and genetic syndromes in 173 cases of medulloblastoma. Med Pediatr Oncol 1993;21:433–434.
13. Cowan R, Hoban P, Kelsey A, et al. The gene for the naevoid basal cell carcinoma syndrome acts as a tumour-suppressor gene in medulloblastoma. Br J Cancer 1997;76:141–145.
14. Kimonis VE, Goldstein AM, Pastakia B, et al. Clinical manifestations in 105 persons with nevoid basal cell carcinoma syndrome. Am J Med Genet 1997;69:299–308.
15. Eberhart CG. Medulloblastoma in mice lacking p53 and PARP: all roads lead to Gli. Am J Pathol 2003;162:7–10.
16. Gailani MR, Bale SJ, Leffell DJ, et al. Developmental defects in Gorlin syndrome related to a putative tumor suppressor gene on chromosome 9. Cell 1992;69:111–117.

17. Hahn H, Wicking C, Zaphiropoulous PG, et al. Mutations of the human homolog of Drosophila patched in the nevoid basal cell carcinoma syndrome. Cell 1996;85:841–851.
18. Johnson RL, Rothman AL, Xie J, et al. Human homolog of patched, a candidate gene for the basal cell nevus syndrome. Science 1996;272:1668–1671.
19. Amlashi SF, Riffaud L, Brassier G, et al. Nevoid basal cell carcinoma syndrome: relation with desmoplastic medulloblastoma in infancy. A population-based study and review of the literature. Cancer 2003;98:618–624.
20. Taylor MD, Mainprize TG, Rutka JT. Molecular insight into medulloblastoma and central nervous system primitive neuroectodermal tumor biology from hereditary syndromes: a review. Neurosurgery 2000;47:888–901.
21. Foulkes WD. A tale of four syndromes: familial adenomatous polyposis, Gardner syndrome, attenuated APC and Turcot syndrome. QJM 1995;88:853–863
22. Hamilton SR, Liu B, Parsons RE, et al. The molecular basis of Turcot’s syndrome. N Engl J Med 1995;332:839–847.
23. Malkin D, Li FP, Strong LC, et al. Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science 1990;250:1233–1238.
24. Srivastava S, Zou ZQ, Pirollo K, et al. Germ-line transmission of a mutated p53 gene in a cancer-prone family with Li-Fraumeni syndrome. Nature 1990;348:747–749.
25. Cawthon RM, Weiss R, Xu GF, et al. A major segment of the neurofibromatosis type 1 gene: cDNA sequence, genomic structure, and point mutations. Cell 1990;62:193–201.
26. Wallace MR, Marchuk DA, Andersen LB, et al. Type 1 neurofibromatosis gene: identification of a large transcript disrupted in three NF1 patients. Science 1990;249:181–186.
27. Listernick R, Charrow J, Gutmann DH. Intracranial gliomas in neurofibromatosis type 1. Am J Med Genet 1999;89:38–44.
28. Pollack IF, Mulvihill JJ. Neurofibromatosis 1 and 2. Brain Pathol 1997;7:823–836.
29. Trofatter JA, MacCollin MM, Rutter JL, et al. A novel moesin-, ezrin-, radixin-like gene is a candidate for the neurofibromatosis 2 tumor suppressor. Cell 1993;72:791–800.
30. Rouleau GA, Merel P, Lutchman M, et al. Alteration in a new gene encoding a putative membrane-organizing protein causes neuro-fibromatosis type 2. Nature 1993;363:515–521.
31. Kandt RS, Haines JL, Smith M, et al. Linkage of an important gene locus for tuberous sclerosis to a chromosome 16 marker for polycystic kidney disease. Nat Genet 1992;2:37–41.
32. van Slegtenhorst M, de Hoogt R, Hermans C, et al. Identification of the tuberous sclerosis gene TSC1 on chromosome 9q34. Science 1997;277:805–808.
33. Liaw D, Marsh DJ, Li J, et al. Germline mutations of the PTEN gene in Cowden disease, an inherited breast and thyroid cancer syndrome. Nat Genet 1997;16:64–67.
34. Latif F, Tory K, Gnarra J, et al. Identification of the von Hippel-Lindau disease tumor suppressor gene. Science 1993;260:1317–1320.
35. Duan DR, Humphrey JS, Chen DY, et al. Characterization of the VHL tumor suppressor gene product: localization, complex formation, and the effect of natural inactivating mutations. Proc Natl Acad Sci U S A 1995;92:6459–6463.
36. Kibel A, Iliopoulos O, DeCaprio JA, et al. Binding of the von Hippel-Lindau tumor suppressor protein to Elongin B and C. Science 1995;269:1444–1446.
37. Ron E, Modan B, Boice JD Jr, et al. Tumors of the brain and nervous system after radiotherapy in childhood. N Engl J Med 1988;319:1033–1039.
38. Kimball D V, Gelber RD, Li F, et al. Second malignancies in patients treated for childhood acute lymphoblastic leukemia. J Clin Oncol 1998;16:2848–2853.
39. Loning L, Zimmermann M, Reiter A, et al. Secondary neoplasms subsequent to Berlin-Frankfurt-Munster therapy of acute lymphoblastic leukemia in childhood: significantly lower risk without cranial radiotherapy. Blood 2000;95:2770–2775.
40. Rimm IJ, Li FC, Tarbell NJ, et al. Brain tumors after cranial irradiation for childhood acute lymphoblastic leukemia. A 13-year experience from the Dana-Farber Cancer Institute and the Children’s Hospital. Cancer 1987;59:1506–1508.
41. Neglia JP, Meadows AT, Robison LL, et al. Second neoplasms after acute lymphoblastic leukemia in childhood. N Engl J Med 1991;325:1330–1336.
42. Bader JL, Meadows AT, Zimmerman LE, et al. Bilateral retinoblastoma with ectopic intracranial retinoblastoma: trilateral retinoblastoma. Cancer Genet Cytogenet 1982;5:203–213.
43. Savla J, Chen TT, Schneider NR, et al. Mutations of the hSNF5/INI1 gene in renal rhabdoid tumors with second primary brain tumors. J Natl Cancer Inst 2000;92:648–650.
44. Hochberg FH, Miller DC. Primary central nervous system lymphoma. J Neurosurg 1988;68:835–853.
45. Meyn MS. Ataxia-telangiectasia and cellular responses to DNA damage. Cancer Res 1995; 55:5991–6001.
46. Penn I. De novo malignances in pediatric organ transplant recipients. Pediatr Transplant 1998;2:56–63.
47. Gold EB, Leviton A, Lopez R, et al. The role of family history in risk of childhood brain tumors. Cancer 1994;73:1302–1311.
48. Draper GJ, Heaf MM, Kinnier Wilson LM. Occurrence of childhood cancers among sibs and estimation of familial risks. J Med Genet 1977;14:81–90.
49. Kuijten RR, Bunin GR. Risk factors for childhood brain tumors. Cancer Epidemiol Biomarkers Prev 1993;2:277–288.
50. Farwell J, Flannery JT. Cancer in relatives of children with central-nervous-system neoplasms. N Engl J Med 1984;311:749–753.
51. Kuijten RR, Strom SS, Rorke LB, et al. Family history of cancer and seizures in young children with brain tumors: a report from the Children’s Cancer Group (United States and Canada). Cancer Causes Control 1993;4:455–464.
52. Taylor MD, Gokgoz N, Andrulis IL, et al. Familial posterior fossa brain tumors of infancy secondary to germline mutation of the hSNF5 gene. Am J Hum Genet 2000;66:1403–1406.
53. Proust F, Laquerriere A, Constantin B, et al. Simultaneous presentation of atypical teratoid/rhabdoid tumor in siblings. J Neurooncol 1999;43:63–70.
54. Moschovi M, Sotiris Y, Prodromou N, et al. Familial medulloblastoma. Pediatr Hematol Oncol 1998;15:421–424.
55. Zwetsloot CP, Kros JM, Gueze HD. Familial occurrence of tumours of the choroid plexus. J Med Genet 1991;28:492–494.
56. Fitzgerald LF. Familial brainstem glioma. Clin Neurol Neurosurg 2000;102:106–108.
57. Dirven CM, Tuerlings J, Molenaar WM, et al. Glioblastoma multiforme in four siblings: a cytogenetic and molecular genetic study. J Neurooncol 1995;24:251–258.
58. Norman MA, Holly EA, Ahn DK, et al. Prenatal exposure to tobacco smoke and childhood brain tumors: results from the United States West Coast childhood brain tumor study. Cancer Epidemiol Biomarkers Prev 1996;5:127–133.
59. Robison LL, Buckley JD, Bunin G. Assessment of environmental and genetic factors in the etiology of childhood cancers: the Children’s Cancer Group epidemiology program. Environ Health Perpect 1995;103[Suppl 6]:111–116.
60. Ali Kahn A, O’Brien DF, Kelly P, et al. The anatomical distribution of cerebral gliomas in mobile phone users. Ir Med J 2003;96:240–242.
61. Hardell L, Mild KH, Carlberg M. Case-control study on the use of cellular and cordless phones and the risk for malignant brain tumours. Int J Radiat Biol 2002;78:931–936.
62. Hardell L, Mild KH, Carlberg M. Further aspects on cellular and cordless telephones and brain tumours. Int J Oncol 2003;22:399–407.
63. Hansson Mild K, Hardell L, Kundi M, et al. Mobile telephones and cancer: is there really no evidence of an association? (review). Int J Mol Med 2003;12:67–72.
64. Kundi M, Mild K, Hardell L, et al. Mobile telephones and cancer—a review of epidemiological evidence. J Toxicol Environ Health B Crit Rev 2004;7:351–384.
65. Khalili K, Del Valle L, Otte J, et al. Human neurotropic polyomavirus, JCV, and its role in carcinogenesis. Oncogene 2003;22:5181–5191.
66. Fine HA. Polyomavirus and medulloblastoma: a smoking gun or guilt by association? J Natl Cancer Inst 2002;94:240–241.
67. Hayashi H, Endo S, Suzuki S, et al. JC virus large T protein transforms rodent cells but is not involved in human medulloblastoma. Neuropathology 2001;21:129–137.
68. Kim JY, Koralnik IJ, LeFave M, et al. Medulloblastomas and primitive neuroectodermal tumors rarely contain polyomavirus DNA sequences. Neuro-oncology 2002;4:165–170.
69. Bigner SH, Bjerkvig R, Laerum OD. DNA content and chromosomal composition of malignant human gliomas. Neurol Clin 1985;3:769–784.
70. Mark SJ, Laerum OD. Modal DNA content of human intracranial neoplasms studied by flow cytometry. J Neurosurg 1980;53:198–204.
71. Christov K, Zapryanov Z. Flow cytometry in brain tumors. I. Ploidy abnormalities. Neoplasma 1986;33:49–55.
72. Giangaspero F, Burger PC. Correlations between cytologic composition and biologic behavior in the glioblastoma multiforme. A postmortem study of 50 cases. Cancer 1983;52: 2320–2333.
73. Shapiro JR, Mehta BM, Fiola MR. Intrinsically chemo- and radioresistant subpopulations identified in freshly resected human gliomas become the dominant population in recurrent tumor samples. Neurology 1990;40:395.
74. Raffel C, Gilles FE, Weinberg KI. Reduction to homozygosity and gene amplification in central nervous system primitive neuroectodermal tumors of childhood. Cancer Res 1990;50:587–591.
75. Douglass EC, Look AT, Kun LE. Cellular DNA content is predictive of early relapse in medulloblastoma. Pediatr Neurosci 1990;15:140.
76. Yasue M, Tomita T, Engelhard H, et al. Prognostic importance of DNA ploidy in medulloblastoma of childhood. J Neurosurg 1989;70:385–391.
77. Biegel JA, Burk CD, Barr FG, et al. Evidence for a 17p tumor related locus distinct from p53 in pediatric primitive neuroectodermal tumors. Cancer Res 1992;52:3391–3395.
78. Thomas GA, Raffel C. Loss of heterozygosity on 6q, 16q, and 17p in human central nervous system primitive neuroectodermal tumors. Cancer Res 1991;51:639–643.
79. Bigner SH, Mark J, Friedman HS, et al. Structural chromosomal abnormalities in human medulloblastoma. Cancer Genet Cytogenet 1988;30:91–101.
80. Griffin CA, Hawkins AL, Packer RJ, et al. Chromosome abnormalities in pediatric brain tumors. Cancer Res 1988;48:175–180.
81. Pietsch T, Waha A, Koch A, et al. Medulloblastomas of the desmoplastic variant carry mutations of the human homologue of Drosophila patched. Cancer Res 1997;57:2085–2088.
82. Raffel C, Jenkins RB, Frederick L, et al. Sporadic medulloblastomas contain PTCH mutations. Cancer Res 1997;57:842–845.
83. Xie J, Johnson RL, Zhang X, et al. Mutations of the PATCHED gene in several types of sporadic extracutaneous tumors. Cancer Res 1997;57:2369–2372.
84. Dong J, Gailani MR, Pomeroy SL, et al. Identification of PATCHED mutations in medulloblastomas by direct sequencing. Hum Mutat 2000;16:89–90.
85. Reardon DA, Michalkiewicz E, Boyett JM, et al. Extensive genomic abnormalities in childhood medulloblastoma by comparative genomic hybridization. Cancer Res 1997;57: 4042–4047.
86. Biegel JA, Zhou JY, Rorke LB, et al. Germ-line and acquired mutations of INI1 in atypical teratoid and rhabdoid tumors. Cancer Res 1999;59:74–79.
87. Seizinger BR, Martuza RL, Gusella JF. Loss of genes on chromosome 22 in tumorigenesis of human acoustic neuroma. Nature 1986;322:644–647.
88. Versteege I, Sevenet N, Lange J, et al. Truncating mutations of hSNF5/INI1 in aggressive paediatric cancer. Nature 1998;394:203–206.
89. Schnitzler G, Sif S, Kingston RE. Human SWI/SNF interconverts a nucleosome between its base state and a stable remodeled state. Cell 1998;94:17–27.
90. Reardon DA, Entrekin RE, Sublett J, et al. Chromosome arm 6q loss is the most common recurrent autosomal alteration detected in primary pediatric ependymoma. Genes Chromosome Cancer 1999;24:230–237.
91. Kramer DL, Parmiter AH, Rorke LB, et al. Molecular cytogenetic studies of pediatric ependymomas. J Neurooncol 1998;37:25–33.
92. Lang FF, Miller DC, Koslow M, et al. Pathways leading to glioblastoma multiforme: a molecular analysis of genetic alterations in 65 astrocytic tumors. J Neurosurg 1994;81:427–436.
93. Coons SW, Johnson PC, Shapiro JR. Cytogenetic and flow cytometry DNA analysis of regional heterogeneity in a low grade human glioma. Cancer Res 1995;55:1569–1577.
94. Pollack IF, Hamilton RL, Finkelstein SD, et al. The relationship between TP53 mutations and overexpression of p53 and prognosis in malignant gliomas of childhood. Cancer Res 1997;57:304–309.

95. James CD, He J, Carlbom E, et al. Chromosome 9 deletion mapping reveals interferon alpha and interferon beta-1 gene deletions in human glial tumors. Cancer Res 1991;51: 1684–1688.
96. Fults D, Pedone CA, Thomas GA, et al. Allelotype of human malignant astrocytoma. Cancer Res 1990;50:5784–5789.
97. Wasson JC, Saylors RL III, Zeltzer P, et al. Oncogene amplification in pediatric brain tumors. Cancer Res 1990;50:2987–2990.
98. Rasheed BK, McLendon RE, Herndon JE, et al. Alterations of the TP53 gene in human gliomas. Cancer Res 1994;54:1324–1330.
99. Litofsky NS, Hinton D, Raffel C. The lack of a role for p53 in astrocytomas in pediatric patients. Neurosurgery 1994;34:967–972.
100. Batra SK, McLendon RE, Koo JS, et al. Prognostic implications of chromosome 17p deletions in human medulloblastomas. J Neurooncol 1995;24:39–45.
101. Louis DN. A molecular genetic model of astrocytoma histopathology. Brain Pathol 1997;7:755–764.
102. Bredel M, Pollack IF, Hamilton RL, et al. Epidermal growth factor receptor expression and gene amplification in high-grade non-brainstem gliomas of childhood. Clin Cancer Res 1999;5:1786–1792.
103. Pollack IF, Finkelstein SD, Burnham J, et al. Age and TP53 mutation frequency in childhood malignant gliomas: results in a multi-institutional cohort. Cancer Res 2001;61:7404–7407.
104. Pollack IF, Finkelstein SD, Woods J, et al. Expression of p53 and prognosis in children with malignant gliomas. N Engl J Med 2002;346:420–427.
105. Zulch K. Brain tumors: their biology and pathology, 3rd ed. Berlin: Springer-Verlag, 1986.
106. Kernonhan JW, Mabon RF, Svien HJ, et al. Pro Staff Meet May. Clinic 1949, 71–75.
107. Daumas-Duport C, Scheithauer B, O’Fallon J, et al. Grading of astrocytomas. A simple and reproducible method. Cancer 1988;62:2152–2165.
108. Kleihues P, Burger PC, Scheihauer B, et al. Histological typing of tumors of the central nervous system. Berlin: Springer-Verlag, 1993.
109. Rorke LB, Gilles FH, Davis RL, et al. Revision of the World Health Organization classification of brain tumors for childhood brain tumors. Cancer 1985;56:1869–1886.
110. Schnipper L. Clinical implications of tumor-cell heterogeneity. N Engl J Med 1986;314: 1423–1431.
111. Kobayashi H. The biological modification of tumor cells as a means of inducing their regression: an overview. J Biol Response Mod 1986;5:1–11.
112. Gould VE. Histogenesis and differentiation: a re-evaluation of these concepts as criteria for the classification of tumors. Hum Pathol 1986;17:212–215.
113. Clark HB. Immunohistochemistry of nervous system antigens: diagnostic applications in surgical neuropathology. Semin Diagn Pathol 1984;1:309–316.
114. Coakham HB, Brownell B. Monoclonal antibodies in the diagnosis of cerebral tumors and cerebrospinal fluid neoplasia. In: Cavanaugh JB, ed. Recent advances in neuropathology. Edinburgh, Churchill Livingstone, 1986:25–53.
115. Perentes E, Rubinstein LJ. Recent applications of immunoperoxidase histochemistry in human neuro-oncology. An update. Arch Pathol Lab Med 1987;111:796–812.
116. Gould VE, Jansson DS, Molenaar WM, et al. Primitive neuroectodermal tumors of the central nervous system. Patterns of expression of neuroendocrine markers, and all classes of intermediate filament proteins. Lab Invest 1990;62:498–509.
117. Molenaar WM, Jansson DS, Gould VE, et al. Molecular markers of primitive neuroectodermal tumors and other pediatric central nervous system tumors. Monoclonal antibodies to neuronal and glial antigens distinguish subsets of primitive neuroectodermal tumors. Lab Invest 1989;61:635–643.
118. Rorke LB, Packer RJ, Biegel JA. Central nervous system atypical teratoid/rhabdoid tumors of infancy and childhood: definition of an entity. J Neurosurg 1996;85:56–65.
119. Bender BL, Yunis EJ. Central nervous system pathology of tuberous sclerosis in children. Ultrastruct Pathol 1980;1:287–299.
120. Nakamura Y, Becker LE. Subependymal giant-cell tumor: astrocytic or neuronal? Acta Neuropathol (Berl) 1983;60:271–277.
121. Bonnin JM, Rubinstein LJ, Papasozomenos SC, et al. Subependymal giant cell astrocytoma. Significance and possible cytogenetic implications of an immunohistochemical study. Acta Neuropathol (Berl) 1984;62:185–193.
122. Powell SZ, Yachnis AT, Rorke LB, et al. Divergent differentiation in pleomorphic xanthoastrocytoma. Evidence for a neuronal element and possible relationship to ganglion cell tumors. Am J Surg Pathol 1996;20:80–85.
123. Perry A, Giannini C, Scheithauer BW, et al. Composite pleomorphic xanthoastrocytoma and ganglioglioma: report of four cases and review of the literature. Am J Surg Pathol 1997;21:763–771.
124. Taratuto AL, Monges J, Lylyk P, et al. Superficial cerebral astrocytoma attached to dura. Report of six cases in infants. Cancer 1984;54:2505–2512.
125. VandenBerg SR, May EE, Rubinstein LJ, et al. Desmoplastic supratentorial neuroepithelial tumors of infancy with divergent differentiation potential (“desmoplastic infantile gangliogliomas”). Report on 11 cases of a distinctive embryonal tumor with favorable prognosis. J Neurosurg 1987;66:58–71.
126. Taratuto AL, VandenBerg SR, Rorke LB. Desmoplastic infantile astrocytoma and ganglioglioma. Lyon, France: International Agency for Research on Cancer Press, 2000.
127. Hassoun J, Gambarelli D, Grisoli F, et al. Central neurocytoma. An electron-microscopic study of two cases. Acta Neuropathol (Berl) 1982;56:151–156.
128. Bailey P, Cushing H. Medulloblastoma cerebelli: a common type of cerebellar glioma of childhood. Arch Neurol Psychiatry 1925;14:192–225.
129. Cushing H. Experiences with the cerebellar medulloblastoma: a critical review. Acta Pathol Microbiol Scand 1930;7:1–86.
130. Rorke LB. The cerebellar medulloblastoma and its relationship to primitive neuroectodermal tumors. J Neuropathol Exp Neurol 1983;42:1–15.
131. Hart MN, Earle KM. Primitive neuroectodermal tumors of the brain in children. Cancer 1973;32:890–897.
132. Kershman J. The medulloblast and medulloblastoma. A study of human embryos. Arch Neurol Psychiat 1938;40:937–967.
133. Fujita S. Quantitative analysis of cell proliferation and differentiation in the cortex of the postnatal mouse cerebellum. J Cell Biol 1967;32:277–287.
134. Fujita S, Shimada M, Nakamura T. H3-thymidine autoradiographic studies on the cell proliferation and differentiation in the external and the internal granular layers of the mouse cerebellum. J Comp Neurol 1966;128:191–208.
135. Swarz JR, Del Cerro M. Lack of evidence for glial cells originating from the external granular layer in mouse cerebellum. J Neurocytol 1977;6:241–250.
136. Fulop Z, Lakos I, Basco E, et al. Identification of early glial elements as the precursors of Bergmann-glia: a Golgi-analysis of the developing rat cerebellar cortex. Acta Morphol Acad Sci Hung 1979;27:273–280.
137. Ghandour MS, Labourdette G, Vincendon G, et al. A biochemical and immunohistological study of S100 protein in developing rat cerebellum. Dev Neurosci 1981;4:98–109.
138. Wechsler-Reya RJ, Scott MP. Control of neuronal precursor proliferation in the cerebellum by Sonic Hedgehog. Neuron 1999;22:103–114.
139. Kenney AM, Cole MD, Rowitch DH. Nmyc upregulation by sonic hedgehog signaling promotes proliferation in developing cerebellar granule neuron precursors. Development 2003;130:15–28.
140. Wetmore C. Sonic hedgehog in normal and neoplastic proliferation: insight gained from human tumors and animal models. Curr Opin Genet Dev 2003;13:34–42.
141. Marino S, Vooijs M, van Der Gulden H, et al. Induction of medulloblastomas in p53-null mutant mice by somatic inactivation of Rb in the external granular layer cells of the cerebellum. Genes Dev 2000;14:994–1004.
142. Pomeroy SL, Tamayo P, Gaasenbeek M, et al. Prediction of central nervous system embryonal tumour outcome based on gene expression. Nature 2002;415:436–442.
143. Pomeroy SL, Sturla LM. Molecular biology of medulloblastoma therapy. Pediatr Neurosurg 2003;39:299–304.
144. Ellison D. Classifying the medulloblastoma: insights from morphology and molecular genetics. Neuropathol Appl Neurobiol 2002;28:257–282.
145. Taylor MD, Liu L, Raffel C, et al. Mutations in SUFU predispose to medulloblastoma. Nat Genet 2002;31:306–310.
146. Aldosari N, Rasheed BK, McLendon RE, et al. Characterization of chromosome 17 abnormalities in medulloblastomas. Acta Neuropathol (Berl) 2000;99:345–351.
147. Rorke L, Hart M, McLendon RE. Supratentorial primitive neuroectodermal tumor (PNET). Lyon, France: IARC Press, 2000.
148. Shapira E, Marom K, Yelin R, et al. A role for the homeobox gene Xvex-1 as part of the BMP-4 ventral signaling pathway. Mech Dev 1999;86:99–111.
149. Tarabykin V, Britanova O, Fradkov A, et al. Expression of PTTG and prc1 genes during telencephalic neurogenesis. Mech Dev 2000;92:301–304.
150. Boquet I, Hitier R, Dumas M, et al. Central brain postembryonic development in Drosophila: implication of genes expressed at the interhemispheric junction. J Neurobiol 2000;42:33–48.
151. Ho KS, Scott MP. Sonic hedgehog in the nervous system: functions, modifications and mechanisms. Curr Opin Neurobiol 2002;12:57–63.
152. Schuurmans C, Guillemot F. Molecular mechanisms underlying cell fate specification in the developing telencephalon. Curr Opin Neurobiol 2002;12:26–34.
153. Rubenstien J, Puelles L. Development of the nervous system. New York: Oxford University Press, 2004.
154. Rorke LB. Pathology of brain and spinal cord tumors. In: Chous M DC, Hockley AD, et al., eds. Pediatric neurosurgery. London: Churchill Livingstone, 1999.
155. Shi SR, Key ME, Kalra KL. Antigen retrieval in formalin-fixed, paraffin-embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections. J Histochem Cytochem 1991;39:741–748.
156. Yachnis AT, Trojanowski JQ. Studies of childhood brain tumors using immunohistochemistry and microwave technology: methodological considerations. J Neurosci Methods 1994;55:191–200.
157. Honig PJ, Charney EB. Children with brain tumor headaches. Distinguishing features. Am J Dis Child 1982;136:121–124.
158. Barlow CF. Headaches and brain tumors. Am J Dis Child 1982;136:99–100.
159. Glaser JS, Hoyt WF, Corbett J. Visual morbidity with chiasmal glioma. Long-term studies of visual fields in untreated and irradiated cases. Arch Ophthalmol 1971;85:3–12.
160. Albright AL, Sclabassi RJ, Slamovits TL, et al. Spasmus nutans associated with optic gliomas in infants. J Pediatr 1984;105:778–780.
161. Yang PJ, Berger PE, Cohen ME, et al. Computed tomography and childhood seizure disorders. Neurology 1979;29:1084–1088.
162. Rich KM, Goldring S, Gado M. Computed tomography in chronic seizure disorder caused by glioma. Arch Neurol 1985;42:26–27.
163. Rasmussen T. Surgical aspects of temporal lobe epilepsy. Results and problems. Acta Neurochir Suppl (Wien) 1980;30:13–24.
164. Spencer DD, Spencer SS, Mattson RH, et al. Intracerebral masses in patients with intractable partial epilepsy. Neurology 1984;34:432–436.
165. Smith KR Jr., Weinburg WA, McAlister WH. Failure to thrive: the diencephalic syndrome of infancy and childhood. J Neurosurg 1965;23:348–351.
166. Torrey EF, Uyeda CI. The diencephalic syndrome of infancy: report of two cases. Am J Dis Child 1965;110:689–696.
167. Packer RJ, Siegel KR, Sutton LN, et al. Leptomeningeal dissemination of primary central nervous system tumors of childhood. Ann Neurol 1985;18:217–221.
168. Rodriguez LA, Edwards MS, Levin VA. Management of hypothalamic gliomas in children: an analysis of 33 cases. Neurosurgery 1990;26:242–246.
169. Fisher PG, Jenab J, Goldthwaite PT, et al. Outcomes and failure patterns in childhood craniopharyngiomas. Childs Nerv Syst 1998;14:558–563.
170. Sklar CA. Craniopharyngioma: endocrine abnormalities at presentation. Pediatr Neurosurg 1994;21[Suppl 1]:18–20.
171. Packer RJ, Sutton LN, Rosenstock JG, et al. Pineal region tumors of childhood. Pediatrics 1984;74:97–102.
172. Sanford RA, Freeman CR, Burger P, et al. Prognostic criteria for experimental protocols in pediatric brainstem gliomas. Surg Neurol 1988;30:276–280.

173. Fisher PG, Breiter SN, Carson BS, et al. A clinicopathologic reappraisal of brain stem tumor classification. Identification of pilocytic astrocytoma and fibrillary astrocytoma as distinct entities. Cancer 2000;89:1569–1576.
174. Gol A, McKissock W. The cerebellar astrocytomas: a report on 98 verified cases. J Neurosurg 1959;16:287–296.
175. Zec N, Cera P, Towfighi J. Extramedullary hematopoiesis in cerebellar hemangioblastoma. Neurosurgery 1991;29:34–37.
176. Mercuri S, Russo A, Palma L. Hemispheric supratentorial astrocytomas in children. Long-term results in 29 cases. J Neurosurg 1981;55:170–173.
177. Palma L, Guidetti B. Cystic pilocytic astrocytomas of the cerebral hemispheres. Surgical experience with 51 cases and long-term results. J Neurosurg 1985;62:811–815.
178. Ettinger AB. Structural causes of epilepsy. Tumors, cysts, stroke, and vascular malformations. Neurol Clin 1994;12:41–56.
179. Daumas-Duport C, Scheithauer BW, Chodkiewicz JP, et al. Dysembryoplastic neuroepithelial tumor: a surgically curable tumor of young patients with intractable partial seizures. Report of thirty-nine cases. Neurosurgery 1988;23:545–556.
180. Drake J, Hoffman HJ, Kobayashi J, et al. Surgical management of children with temporal lobe epilepsy and mass lesions. Neurosurgery 1987;21:792–797.
181. Duffner PK, Burger PC, Cohen ME, et al. Desmoplastic infantile gangliogliomas: an approach to therapy. Neurosurgery 1994;34:583–589.
182. Buetow PC, Smirniotopoulos JG, Done S. Congenital brain tumors: a review of 45 cases. Am J Neuroradiol 1990;11:793–799.
183. Zimmerman RA, Haselgrove JC, Bilaniuk LT, et al. Magnetization transfer suppression in gadolinium enhancement of the child’s brain. Proc XV Symposium Neuroradiologicum, 1994: 267–269.
184. Gonen O, Viswanathan AK, Catalaa I, et al. Total brain N-acetylaspartate concentration in normal, age-grouped females: quantitation with non-echo proton NMR spectroscopy. Magn Reson Med 1998;40:684–689.
185. Wang Z, Sutton LN, Cnaan A, et al. Proton MR spectroscopy of pediatric cerebellar tumors. Am J Neuroradiol 1995;16:1821–1833.
186. Sutton LN, Wang ZJ, Wehrli SL, et al. Proton spectroscopy of suprasellar tumors in pediatric patients. Neurosurgery 1997;41:388–394; discussion 394–395.
187. Girard N, Wang ZJ, Erbetta A, et al. Prognostic value of proton MR spectroscopy of cerebral hemisphere tumors in children. Neuroradiology 1998;40:121–125.
188. Lazareff JA, Gupta RK, Alger J. Variation of post-treatment H-MRSI choline intensity in pediatric gliomas. J Neurooncol 1999;41:291–298.
189. Rock JP, Hearshen D, Scarpace L, et al. Correlations between magnetic resonance spectroscopy and image-guided histopathology, with special attention to radiation necrosis. Neurosurgery 2002;51:912–919; discussion 919–920.
190. Haacke EM, Xu Y, Cheng YC, et al. Susceptibility weighted imaging (SWI). Magn Reson Med 2004;52:612–618.
191. Haacke M, Ogg R, Reichenbach J, et al. Susceptibility weighted imaging (SWI): a new means to enhance image contrast. In: Proceedings of the 10th Annual Meeting of ISMRM, Honolulu, HI, 2002;297.
192. Haacke M, Herigault G, Yu Y, et al. Observing tumor vascularity noninvasively using magnetic resonance imaging. Image Anal 2002;21:107–113.
193. Knopp EA, Cha S, Johnson G, et al. Glial neoplasms: dynamic contrast-enhanced T2*-weighted MR imaging. Radiology 1999;211:791–798.
194. Aronen HJ, Gazit IE, Louis DN, et al. Cerebral blood volume maps of gliomas: comparison with tumor grade and histologic findings. Radiology 1994;191:41–51.
195. Provenzale JM, Wang GR, Brenner T, et al. Comparison of permeability in high-grade and low-grade brain tumors using dynamic susceptibility contrast MR imaging. AJR Am J Roentgenol 2002;178:711–716.
196. Huisman TA, Sorensen AG. Perfusion-weighted magnetic resonance imaging of the brain: techniques and application in children. Eur Radiol 2004;14:59–72.
197. Rollins NK, Lowry PA, Shapiro KN. Comparison of gadolinium-enhanced MR and thallium-201 single photon emission computed tomography in pediatric brain tumors. Pediatr Neurosurg 1995;22:8–14.
198. Maria BL, Drane WB, Quisling RJ, et al. Correlation between gadolinium-diethylenetriaminepentaacetic acid contrast enhancement and thallium-201 chloride uptake in pediatric brainstem glioma. J Child Neurol 1997;12:341–348.
199. Hustinx R, Alavi A. SPECT and PET imaging of brain tumors. Neuromaging Clin N Am. 1999;9:751–766.
200. Erdem E, Zimmerman RA, Haselgrove JC, et al. Diffusion-weighted imaging and fluid attenuated inversion recovery imaging in the evaluation of primitive neuroectodermal tumors. Neuroimaging Clin North Am 2001;43:927–933.
201. Sugahara T, Korogi Y, Kochi M, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging 1999;9:53–60.
202. Sutton LN, Lasner T, Hunter J, et al. Thirteen-year-old female with hemangioblastoma. Pediatr Neurosurg 1997;27:50–55.
203. Lesniak MS, Klem JM, Weingart J, et al. Surgical outcome following resection of contrast-enhanced pediatric brainstem gliomas. Pediatr Neurosurg 2003;39:314–322.
204. DiPaolo DP, Zimmerman RA, Rorke LB, et al. Neurofibromatosis type 1: pathologic substrate of high-signal-intensity foci in the brain. Radiology 1995;195:721–724.
205. Hunter JV, Molloy PT, Needle MN. Neurofibromatosis type I and lesions of the corpus callosum in a pediatric population. 7th Annual European Neurofibromatosis Meeting, Paris, 1997. Abstract.
206. Gonen O, Wang ZJ, Viswanathan AK, et al. Three-dimensional multivoxel proton MR spectroscopy of the brain in children with neurofibromatosis type 1. Am J Neuroradiol 1999;20:1333–1341.
207. Listernick R, Louis DN, Packer RJ, et al. Optic pathway gliomas in children with neurofibromatosis 1: consensus statement from the NF1 Optic Pathway Glioma Task Force. Ann Neurol 1997;41:143–149.
208. Ricci PE, Karis JP, Heiserman JE, et al. Differentiating recurrent tumor from radiation necrosis: time for re-evaluation of positron emission tomography? Am J Neuroradiol 1998;19:407–413.
209. Rock JP, Scarpace L, Hearshen D, et al. Associations among magnetic resonance spectroscopy, apparent diffusion coefficients, and image-guided histopathology with special attention to radiation necrosis. Neurosurgery 2004;54:1111–1117; discussion 1117–1119.
210. Hurd RE, Freeman DM. Metabolite specific proton magnetic resonance imaging. Proc Natl Acad Sci 1989;86:4402–4406.
211. Ryner LN, Sorenson JA, Thomas MA. 3D localized 2D NMR spectroscopy on an MRI scanner. J Magn Reson B 1995;107:126–137.
212. Mayer D, Dreher W, Liebfritz D. Fast IH spectroscopic imaging combined with 2D correlation spectroscopy uncoupled in both frequency domains, Proc Intl Soc Mag Reson Med, 2000;8:370.
213. Skare S, Hedehus M, Moseley ME, et al. Condition number as a measure of noise performance of diffusion tensor data acquisition schemes with MRI. J Magn Reson 2000;147: 340–352.
214. Detre JA, Alsop DC. Perfusion magnetic resonance imaging with continuous arterial spin labeling: methods and clinical applications in the central nervous system. Eur J Radiol 1999;30:115–124.
215. Albright AL, Packer RJ, Zimmerman R, et al. Magnetic resonance scans should replace biopsies for the diagnosis of diffuse brain stem gliomas: a report from the Children’s Cancer Group. Neurosurgery 1993;33:1026–1029.
216. Lunsford LD. Diagnosis of mass lesions using the Leksell system. In: Lunsford LD, ed. Modern stereotactic surgery. Boston: Martinus Nijhoff, 1987:
217. Broggi G, Franzini A, Migliavacca F, et al. Stereotactic biopsy of deep brain tumors in infancy and childhood. Childs Brain 1983;10:92–98.
218. Jones RF, Stening WA, Brydon M. Endoscopic third ventriculostomy. Neurosurgery 1990;26:86–91.
219. Hopf NJ, Grunert P, Fries G, et al. Endoscopic third ventriculostomy: outcome analysis of 100 consecutive procedures. Neurosurgery 1999;44:795–804.
220. Dias MS, Albright AL. Management of hydrocephalus complicating childhood posterior fossa tumors. Pediatr Neurosci 1989;15:283–289.
221. Foy PM, Chadwick DW, Rajgopalan N, et al. Do prophylactic anticonvulsant drugs alter the pattern of seizures after craniotomy? J Neurol Neurosurg Psychiatry 1992;55:753–757.
222. Nazar GB, Hoffman HJ, Becker LE, et al. Infratentorial ependymomas in childhood: prognostic factors and treatment. J Neurosurg 1990;72:408–417.
223. Healey EA, Barnes PD, Kupsky WJ, et al. The prognostic significance of postoperative residual tumor in ependymoma. Neurosurgery 1991;28:666–671.
224. Rousseau P, Habrand JL, Sarrazin D, et al. Treatment of intracranial ependymomas of children: review of a 15-year experience. Int J Radiat Oncol Biol Phys 1994;28:381–386.
225. Sutton LN, Goldwein J, Perilongo G, et al. Prognostic factors in childhood ependymomas. Pediatr Neurosurg 1990;16:57–65.
226. Pollack IF, Gerszten PC, Martinez AJ, et al. Intracranial ependymomas of childhood: long-term outcome and prognostic factors. Neurosurgery 1995;37:655–666.
227. Robertson PL, Zeltzer PM, Boyett JM, et al. Survival and prognostic factors following radiation therapy and chemotherapy for ependymomas in children: a report of the Children’s Cancer Group. J Neurosurg 1998;88:695–703.
228. Horn B, Heideman R, Geyer R, et al. A multi-institutional retrospective study of intracranial ependymoma in children: identification of risk factors. J Pediatr Hematol Oncol 1999;21:203–211.
229. Wisoff JH, Boyett J, K. B, et al. Neurosurgical management and influence of extent of resection on survival in pediatric high-grade astrocytomas: a report on the CCG 934. J Neurosurg 1993;78:344 (abst).
230. Finlay JL, Boyett JM, Yates AJ, et al. Randomized phase III trial in childhood high-grade astrocytoma comparing vincristine, lomustine, and prednisone with the eight-drugs-in-1-day regimen. Childrens Cancer Group. J Clin Oncol 1995;13:112–123.
231. Campbell JW, Pollack IF, Martinez AJ, et al. High-grade astrocytomas in children: radiologically complete resection is associated with an excellent long-term prognosis. Neurosurgery 1996;38:258–264.
232. Wisoff JH, Boyett JM, Berger MS, et al. Current neurosurgical management and the impact of the extent of resection in the treatment of malignant gliomas of childhood: a report of the Children’s Cancer Group trial no. CCG-945. J Neurosurg 1998;89:52–59.
233. Albright AL, Wisoff JH, Zeltzer P, et al. Prognostic factors in children with supratentorial (nonpineal) primitive neuroectodermal tumors. A neurosurgical perspective from the Children’s Cancer Group. Pediatr Neurosurg 1995;22:1–7.
234. Berry MP, Jenkin RD, Keen CW, et al. Radiation treatment for medulloblastoma. A 21-year review. J Neurosurg 1981;55:43–51.
235. Bourne JP, Geyer R, Berger M, et al. The prognostic significance of postoperative residual contrast enhancement on CT scan in pediatric patients with medulloblastoma. J Neurooncol 1992;14:263–270.
236. Caputy AJ, McCullough DC, Manz HJ, et al. A review of the factors influencing the prognosis of medulloblastoma. The importance of cell differentiation. J Neurosurg 1987;66: 80–87.
237. Hughes EN, Shillito J, Sallan SE, et al. Medulloblastoma at the joint center for radiation therapy between 1968 and 1984. The influence of radiation dose on the patterns of failure and survival. Cancer 1988;61:1992–1998.
238. Tait DM, Thornton-Jones H, Bloom HJ, et al. Adjuvant chemotherapy for medulloblastoma: the first multi-centre control trial of the International Society of Paediatric Oncology (SIOP I). Eur J Cancer 1990;26:464–469.
239. Shaw EG, Wisoff JH. Prospective clinical trials of intracranial low-grade glioma in adults and children. Neuro-oncology 2003;5:153–160.
240. Packer RJ, Perilongo G, Johnson D, et al. Choroid plexus carcinoma of childhood. Cancer 1992;69:580–585.

241. Berger C, Thiesse P, Lellouch-Tubiana A, et al. Choroid plexus carcinomas in childhood: clinical features and prognostic factors. Neurosurgery 1998;42:470–475.
242. Pencalet P, Sainte-Rose C, Lellouch-Tubiana A, et al. Papillomas and carcinomas of the choroid plexus in children. J Neurosurg 1998;88:521–528.
243. Cochrane DD, Gustavsson B, Poskitt KP, et al. The surgical and natural morbidity of aggressive resection for posterior fossa tumors in childhood. Pediatr Neurosurg 1994;20:19–29.
244. Albright AL, Wisoff JH, Zeltzer PM, et al. Current neurosurgical treatment of medulloblastomas in children. A report from the Children’s Cancer Study Group. Pediatr Neurosci 1989;15:276–282.
245. Sanford RA. Craniopharyngioma: results of survey of the American Society of Pediatric Neurosurgery 1994;21[Suppl 1]:39–43.
246. Albright AL, Sposto R, Holmes E, et al. Correlation of neurosurgical subspecialization with outcomes in children with malignant brain tumors. Neurosurgery 2000;47:879–885.
247. Kelly PJ. Stereotactic craniotomy. Neurosurg Clin North Am 1990;1:781–799.
248. Barnett GH, Kormos DW, Steiner CP, et al. Use of a frameless, armless stereotactic wand for brain tumor localization with two-dimensional and three-dimensional neuroimaging. Neurosurgery 1993;33:674–678.
249. Martin AJ, Hall WA, Liu H, et al. Brain tumor resection: intraoperative monitoring with high-field-strength MR imaging-initial results. Radiology 2000;215:221–228.
250. Schneider W, Noll DC, Cohen JD. Functional topographic mapping of the cortical ribbon in human vision with conventional MRI scanners. Nature 1993;365:150–153.
251. Berger MS, Ojemann GA, Lettich E. Neurophysiological monitoring during astrocytoma surgery. Neurosurg Clin North Am 1990;1:65–80.
252. Schneider JH Jr, Raffel C, McComb JG. Benign cerebellar astrocytomas of childhood. Neurosurgery 1992;30:58–62.
253. Barrer SJ, Schut L, Sutton LN, et al. Re-operation for recurrent brain tumors in children. Childs Brain 1984;11:375–386.
254. Doxey D, Bruce D, Sklar F, et al. Posterior fossa syndrome: identifiable risk factors and irreversible complications. Pediatr Neurosurg 1999;31:131–136.
255. Pollack IF. Posterior fossa syndrome. Int Rev Neurobiol 1997;41:411–432.
256. Janssen G, Messing-Junger AM, Engelbrecht V, et al. Cerebellar mutism syndrome. Klin Padiatr 1998;210:243–247.
257. Turgut M. Transient “cerebellar” mutism. Childs Nerv Syst 1998;14:161–166.
258. Gajjar A, Sanford RA, Bhargava R, et al. Medulloblastoma with brain stem involvement: the impact of gross total resection on outcome. Pediatr Neurosurg 1996;25:182–187.
259. Pollack IF. Infratentorial primitive neuroectodermal tumors. New York: Thieme, 2001.
260. Steinbok P, Cochrane DD, Perrin R, et al. Mutism after posterior fossa tumour resection in children: incomplete recovery on long-term follow-up. Pediatr Neurosurg 2003;39:179–183.
261. Gelabert-Gonzalez M, Fernandez-Villa J. Mutism after posterior fossa surgery. Review of the literature. Clin Neurol Neurosurg 2001;103:111–114.
262. Ciernik IF, Dizendorf E, Baumert BG, et al. Radiation treatment planning with an integrated positron emission and computer tomography (PET/CT): a feasibility study. Int J Radiat Oncol Biol Phys 2003;57:853–863.
263. Nelson SJ, Graves E, Pirzkall A, et al. In vivo molecular imaging for planning radiation therapy of gliomas: an application of 1H MRSI. J Magn Reson Imaging 2002;16:464–476.
264. Thames HD Jr., Peters LJ, Withers HR, et al. Accelerated fractionation vs hyperfractionation: rationales for several treatments per day. Int J Radiat Oncol Biol Phys 1983;9:127–138.
265. Taghian A, DuBois W, Budach W, et al. In vivo radiation sensitivity of glioblastoma multiforme. Int J Radiat Oncol Biol Phys 1995;32:99–104.
266. Freeman CR. Hyperfractionated radiotherapy for diffuse intrinsic brain stem tumors in children. Pediatr Neurosurg 1996;24:103–110.
267. Halperin EC, Constine LS, Tarbell NJ, et al. Late effects of cancer treatment. In: Halperin EC, Constine LS, Tarbell NJ, et al., eds. Pediatric radiation oncology. New York: Raven Press, 1994:485–554.
268. Mandell LR, Kadota R, Freeman C, et al. There is no role for hyperfractionated radiotherapy in the management of children with newly diagnosed diffuse intrinsic brainstem tumors: results of a Pediatric Oncology Group phase III trial comparing conventional vs. hyperfractionated radiotherapy. Int J Radiat Oncol Biol Phys 1999;43:959–964.
269. Fraass BA. The development of conformal radiation therapy. Med Phys 1995;22:1911–1921.
270. Thornton AF Jr, Hegarty TJ, Ten Haken RK, et al. Three-dimensional treatment planning of astrocytomas: a dosimetric study of cerebral irradiation. Int J Radiat Oncol Biol Phys 1991;20:1309–1315.
271. Kirsch DG, Tarbell NJ. Conformal radiation therapy for childhood CNS tumors. Oncologist 2004;9:442–450.
272. Leibel SA, Scott CB, Loeffler JS. Contemporary approaches to the treatment of malignant gliomas with radiation therapy. Semin Oncol 1994;21:198–219.
273. Hadjipanayis CG, Kondziolka D, Gardner P, et al. Stereotactic radiosurgery for pilocytic astrocytomas when multimodal therapy is necessary. J Neurosurg 2002;97:56–64.
274. Boethius J, Ulfarsson E, Rahn T, et al. Gamma knife radiosurgery for pilocytic astrocytomas. J Neurosurg 2002;97:677–680.
275. Chiou SM, Lunsford LD, Niranjan A, et al. Stereotactic radiosurgery of residual or recurrent craniopharyngioma, after surgery, with or without radiation therapy. Neuro-oncology 2001;3:159–166
276. Plowman PN, Wraith C, Royle N, et al. Stereotactic radiosurgery. IX. Craniopharyngioma: durable complete imaging responses and indications for treatment. Br J Neurosurg 1999; 13:352–358.
277. Miralbell R, Lomax A, Bortfeld T, et al. Potential role of proton therapy in the treatment of pediatric medulloblastoma/primitive neuroectodermal tumors: reduction of the supratentorial target volume. Int J Radiat Oncol Biol Phys 1997;38:477–484.
278. Hug EB, Muenter MW, Archambeau JO, et al. Conformal proton radiation therapy for pediatric low-grade astrocytomas. Strahlenther Onkol 2002;178:10–17.
279. Kortmann RD, Timmermann B, Taylor RE, et al. Current and future strategies in radiotherapy of childhood low-grade glioma of the brain. Part II: Treatment-related late toxicity. Strahlenther Onkol 2003;179:585–597.
280. Hasegawa T, Kondziolka D, Hadjipanayis CG, et al. Management of cystic craniopharyngiomas with phosphorus-32 intracavitary irradiation. Neurosurgery 2004;54:813–820; discussion 820–822.
281. Backlund EO, Axelsson B, Bergstrand CG, et al. Treatment of craniopharyngiomas—the stereotactic approach in a ten to twenty-three years’ perspective. I. Surgical, radiological and ophthalmological aspects. Acta Neurochir (Wien) 1989;99:11–19.
282. Dobbing J, Sands J. Quantitative growth and development of human brain. Arch Dis Child 1973;48:757–767.
283. Mulhern RK, White HA, Glass JO, et al. Attentional functioning and white matter integrity among survivors of malignant brain tumors of childhood. J Int Neuropsychol Soc 2004; 10:180–189.
284. Reddick WE, Glass JO, Langston JW, et al. Quantitative MRI assessment of leukoencephalopathy. Magn Reson Med 2002;47:912–920.
285. Schultheiss TE, Kun LE, Ang KK, et al. Radiation response of the central nervous system. Int J Radiat Oncol Biol Phys 1995;31:1093–1112.
286. Tofilon PJ, Fike JR. The radioresponse of the central nervous system: a dynamic process. Radiat Res 2000;153:357–370.
287. Yuan H, Gaber MW, McColgan T, et al. Radiation-induced permeability and leukocyte adhesion in the rat blood-brain barrier: modulation with anti-ICAM-1 antibodies. Brain Res 2003;969:59–69.
288. Gaber MW, Yuan H, Killmar JT, et al. An intravital microscopy study of radiation-induced changes in permeability and leukocyte-endothelial cell interactions in the microvessels of the rat pia mater and cremaster muscle. Brain Res Brain Res Protoc 2004;13:1–10.
289. Mulhern RK, Merchant TE, Gajjar A, et al. Late neurocognitive sequelae in survivors of brain tumours in childhood. Lancet Oncol 2004;5:399–408.
290. Mulhern RK, Palmer SL, Reddick WE, et al. Risks of young age for selected neurocognitive deficits in medulloblastoma are associated with white matter loss. J Clin Oncol 2001;19: 472–479.
291. Packer RJ, Zimmerman RA, Kaplan A, et al. Early cystic/necrotic changes after hyperfractionated radiation therapy in children with brain stem gliomas. Data from the Childrens Cancer Group. Cancer 1993;71:2666–2674.
292. Fouladi M, Chintagumpala M, Laningham FH, et al. White matter lesions detected by magnetic resonance imaging following radiotherapy and high dose chemotherapy in children with medulloblastoma/PNET. J Clin Oncol 2005;22:4551–4560.
293. Helton K, Edwards M, Steen RG, et al. Late transient MRI findings associated with treatment in pediatric brain tumor patients. J Neurosurg 2004(in press).
294. Emami B, Lyman J, Brown A, et al. Tolerance of normal tissue to therapeutic irradiation. Int J Radiat Oncol Biol Phys 1991;21:109–122.
295. Leibel S, Sheline GE. Tolerance of the central and peripheral nervous system to therapeutic irradiation. Adv Radiat Biol 1987;12:257–288.
296. Ang KK, Stephens LC. Prevention and management of radiation myelopathy. Oncology (Huntingt) 1994;8:71–76.
297. Jiang GL, Tucker SL, Guttenberger R, et al. Radiation-induced injury to the visual pathway. Radiother Oncol1994;30:17–25.
298. Parsons JT, Bova FJ, Fitzgerald CR, et al. Radiation optic neuropathy after megavoltage external-beam irradiation: analysis of time-dose factors. Int J Radiat Oncol Biol Phys 1994; 30:755–763.
299. Merchant TE, Gould CJ, Xiong X, et al. Early neuro-otologic effects of three-dimensional irradiation in children with primary brain tumors. Int J Radiat Oncol Biol Phys 2004;58:1194–207.
300. Grau C, Overgaard J. Postirradiation sensorineural hearing loss: a common but ignored late radiation complication. Int J Radiat Oncol Biol Phys 1996;36:515–517.
301. Miettinen S, Laurikainen E, Johansson EA. Radiotherapy enhances ototoxicity of cisplatin in children. Acta Otolaryngol Suppl 1997;529:90–94.
302. Wara WM, Irvine AR, Neger RE, et al. Radiation retinopathy. Int J Radiat Oncol Biol Phys 1979;5:81–83.
303. Deeg HJ, Flournoy N, Sullivan KM, et al. Cataracts after total body irradiation and marrow transplantation: a sparing effect of dose fractionation. Int J Radiat Oncol Biol Phys 1984;10:957–964.
304. Merchant TE, Goloubeva O, Pritchard DL, et al. Radiation dose-volume effects on growth hormone secretion. Int J Radiat Oncol Biol Phys 2002;52:1264–1270.
305. Constine LS, Woolf PD, Cann D, et al. Hypothalamic-pituitary dysfunction after radiation for brain tumors. N Engl J Med 1993;328:87–94.
306. Bradbury M. The blood-brain barrier. Exp Physiol 1993;78:53–72.
307. de Boer A, Breimer D. The blood-brain barrier: clinical implications for drug delivery to the brain. J R Coll Phys London 1994;28:502–506.
308. Johansson B. The physiology of the blood-brain barrier. Adv Exp Med Biol 1990;27:25–39.
309. Bart J, Groen HJ, Hendrikse N, et al. The blood-brain barrier and oncology: new insights into function and modulation. Cancer Treat Rev 2000;26:449–462.
310. Rao VV, Dahlheimer JL, Bardgett ME, et al. Choroid plexus epithelial expression of MDR1 P glycoprotein and multidrug resistance-associated protein contribute to the blood-cerebrospinal-fluid drug-permeability barrier. Proc Natl Acad Sci U S A 1999;96:3900–3905.
311. Schinkel A. P-glycoprotein, a gatekeeper in the blood-brain barrier. Adv Drug Deliv Rev 1999;36:179–194.
312. Betz A. An overview of the multiple functions of the blood-brain barrier. NIDA Res Monogr 1992;120:5–72.
313. Angeletti RH, Novikoff PM, Juvvadi SR, et al. The choroid plexus epithelium is the site of the organic anion transport protein in the brain. Proc Natl Acad Sci U S A 1997;94:283–286.

314. Pardridge W, Oldendorf W, Cancilla P, et al. Blood-brain barrier: interface between internal medicine and the brain. Ann Intern Med 1986;105:82–95.
315. Wijnholds J, deLange EC, Scheffer GL, et al. Multidrug resistance protein 1 protects the choroid plexus epithelium and contributes to the blood-cerebrospinal fluid barrier. J Clin Invest 2000;105:279–285.
316. Mellet L. Physiochemical considerations and pharmacokinetic behavior in delivery of drugs to the central nervous system. Cancer Treat Rep 1977;61:527–531.
317. Koch-Weser J, Sellers E. Binding of drugs to serum albumin. N Engl J Med 1976;294:311–316.
318. Poplack D, Bleyer W, Horowitz M. Pharmacology of antineoplastic agents in cerebrospinal fluid. New York: Plenum Press, 1980.
319. Zamboni WC, Gajjar AJ, Heideman RL, et al. Phenytoin alters the disposition of topotecan and N-desmethyl topotecan in a patient with medulloblastoma. Clin Cancer Res 1998;4:783–789.
320. Murry DJ, Cherrick I, Salama V, et al. Influence of phenytoin on the disposition of irinotecan: a case report. J Pediatr Hematol Oncol 2002;24:130–133.
321. Jarden JO, Dhawan V, Moeller JR, et al. The time course of steroid action on blood-to-brain and blood-to-tumor transport of 82Rb: a positron emission tomographic study. Ann Neurol 1989;25:239–245.
322. Nakagawa H, Groothuis DR, Owens ES, et al. Dexamethasone effects on [125I]albumin distribution in experimental RG-2 gliomas and adjacent brain. J Cereb Blood Flow Metab 1987;7:687–701.
323. Groothuis DR, Wright DC, Ostertag CB. The effect of 125I interstitial radiotherapy on blood-brain barrier function in normal canine brain. J Neurosurg 1987;67:895–902.
324. Neuwelt EA, Hill SA, Frenkel EP. Osmotic blood-brain barrier modification and combination chemotherapy: concurrent tumor regression in areas of barrier opening and progression in brain regions distant to barrier opening. Neurosurgery 1984;15:362–366.
325. Emerich DF, Dean RL, Osborn C, et al. The development of the bradykinin agonist labradimil as a means to increase the permeability of the blood-brain barrier: from concept to clinical evaluation. Clin Pharmacokinet 2001;40:105–123.
326. Papadakis V, Dunkel IJ, Cramer LD, et al. High-dose carmustine, thiotepa and etoposide followed by autologous bone marrow rescue for the treatment of high risk central nervous system tumors. Bone Marrow Transplant 2000;26:153–160.
327. Finlay JL, Goldman S, Wong MC, et al. Pilot study of high-dose thiotepa and etoposide with autologous bone marrow rescue in children and young adults with recurrent CNS tumors. The Children’s Cancer Group. J Clin Oncol 1996;14:2495–2503.
328. Guruangan S, Dunkel IJ, Goldman S, et al. Myeloablative chemotherapy with autologous bone marrow rescue in young children with recurrent malignant brain tumors. J Clin Oncol 1998;16:2486–2493.
329. Abrey LE, Rosenblum MK, Papadopoulos E, et al. High dose chemotherapy with autologous stem cell rescue in adults with malignant primary brain tumors. J Neurooncol 1999; 44:147–153.
330. Blaney S, Boyett J, Friedman H, et al. Phase 1 trial of intrathecal mafosfamide in infants with embryonal CNS tumors: a Pediatric Brain Tumor Consortium study. International Symposium on Pediatric Neuro-Oncology, London, England, 2002.
331. Shapiro W. Chemotherapy of malignant gliomas: studies of the BTCG. Rev Neurol (Paris) 1992;148:428–434.
332. Brem H, Piantadosi S, Burger PC, et al. Placebo-controlled trial of safety and efficacy of intraoperative controlled delivery by biodegradable polymers of chemotherapy for recurrent gliomas. The Polymer-brain Tumor Treatment Group. Lancet 1995;345:1008–1012.
333. Rapoport SI. Osmotic opening of the blood-brain barrier: principles, mechanism, and therapeutic applications. Cell Mol Neurobiol 2000;20:217–230.
334. Kroll RA, Neuwelt EA: Outwitting the blood-brain barrier for therapeutic purposes: osmotic opening and other means. Neurosurgery 1998;42:1083–1099.
335. Neuwelt EA, Diehl JT, Vu LH, et al. Monitoring of methotrexate delivery in patients with malignant brain tumors after osmotic blood-brain barrier disruption. Ann Intern Med 1981;94:449–454.
336. Neuwelt EA, Balaban E, Diehl J, et al. Successful treatment of primary central nervous system lymphomas with chemotherapy after osmotic blood-brain barrier opening. Neurosurgery 1983;12:662–671.
337. Doolittle ND, Miner ME, Hall WA, et al. Safety and efficacy of a multicenter study using intraarterial chemotherapy in conjunction with osmotic opening of the blood-brain barrier for the treatment of patients with malignant brain tumors. Cancer 2000;88:637–647.
338. Iwadate Y, Namba H, Saegusa T, et al. Intra-arterial mannitol infusion in the chemotherapy for malignant brain tumors. J Neurooncol 1993;15:185–193.
339. Liu Y, Hashizume K, Chen Z, et al. Correlation between bradykinin-induced blood-tumor barrier permeability and B2 receptor expression in experimental brain tumors. Neurol Res 2001;23:379–387.
340. Elliott PJ, Hayward NJ, Huff MR, et al. Unlocking the blood-brain barrier: a role for RMP-7 in brain tumor therapy. Exp Neurol 1996;141:214–224.
341. Cloughesy TF, Black KL, Gobin YP, et al. Intra-arterial Cereport (RMP-7) and carboplatin: a dose escalation study for recurrent malignant gliomas. Neurosurgery 1999;44:270–278; discussion 278–279.
342. Black KL, Chio CC. Increased opening of blood-tumour barrier by leukotriene C4 is dependent on size of molecules. Neurol Res 1992;14:402–404.
343. Gutman M, Laufer R, Eisenthal A, et al. Increased microvascular permeability induced by prolonged interleukin-2 administration is attenuated by the oxygen-free-radical scavenger dimethylthiourea. Cancer Immunol Immunother 1996;43:240–244.
344. de Vries H, Blom-Roosemalen M, van Osten M, et al. The influence of cytokines on the integrity of the blood-brain barrier in vitro. J Neuroimmunol 1996;64:37–43.
345. Dean RL, Emerich DF, Hasler BP, et al. Cereport (RMP-7) increases carboplatin levels in brain tumors after pretreatment with dexamethasone. Neuro-oncology 1999;1:268–274.
346. Thomas J, Lind M, Ford J, et al. Pharmacokinetics of carboplatin administered in combination with the bradykinin agonist Cereport (RMP-7) for the treatment of brain tumors. Cancer Chem Pharmacol 2000;45:284–290.
347. Hochberg FH, Parker LM, Takvorian T, et al. High-dose BCNU with autologous bone marrow rescue for recurrent glioblastoma multiforme. J Neurosurg 1981;54:455–460.
348. Bashir R, Hochberg FH, Linggood RM, et al. Pre-irradiation internal carotid artery BCNU in treatment of glioblastoma multiforme. J Neurosurg 1988;68:917–919.
349. Kalifa C, Hartmann O, Demeocq F, et al. High-dose busulfan and thiotepa with autologous bone marrow transplantation in childhood malignant brain tumors: a phase II study. Bone Marrow Transplant 1992;9:227–233.
350. Bergman I, Jakacki RI, Heller G, et al. Treatment of standard risk medulloblastoma with craniospinal irradiation, carboplatin, and vincristine. Med Pediatr Oncol 1997;29:563–567.
351. Mason WP, Grovas A, Halpern S, et al. Intensive chemotherapy and bone marrow rescue for young children with newly diagnosed malignant brain tumors. J Clin Oncol 1998;16: 210–221.
352. Dupuis-Girod S, Hartmann O, Benhamou E, et al. Will high dose chemotherapy followed by autologous bone marrow transplantation supplant cranio-spinal irradiation in young children treated for medulloblastoma? J Neurooncol 1996;27:87–98.
353. Heideman RL, Douglass EC, Krance RA, et al. High-dose chemotherapy and autologous bone marrow rescue followed by interstitial and external-beam radiotherapy in newly diagnosed pediatric malignant gliomas. J Clin Oncol 1993;11:1458–1465.
354. Dunkel IJ, Boyett JM, Yates A, et al. High-dose carboplatin, thiotepa, and etoposide with autologous stem-cell rescue for patients with recurrent medulloblastoma. Children’s Cancer Group. J Clin Oncol 1998;16:222–228.
355. Mahoney DH Jr., Strother D, Camitta B, et al. High-dose melphalan and cyclophosphamide with autologous bone marrow rescue for recurrent/progressive malignant brain tumors in children: a pilot pediatric oncology group study. J Clin Oncol 1996;14:382–388.
356. Strother D, Ashley D, Kellie SJ, et al. Feasibility of four consecutive high-dose chemotherapy cycles with stem-cell rescue for patients with newly diagnosed medulloblastoma or supratentorial primitive neuroectodermal tumor after craniospinal radiotherapy: results of a collaborative study. J Clin Oncol 2001;19:2696–2704.
357. Finlay JL, August C, Packer R, et al. High-dose multi-agent chemotherapy followed by bone marrow ‘rescue’ for malignant astrocytomas of childhood and adolescence. J Neurooncol 1990;9:239–248.
358. Jakacki RI, Siffert J, Jamison C, et al. Dose-intensive, time-compressed procarbazine, CCNU, vincristine (PCV) with peripheral blood stem cell support and concurrent radiation in patients with newly diagnosed high-grade gliomas. J Neurooncol 1999;44:77–83.
359. Kedar A, Maria BL, Graham-Pole J, et al. High-dose chemotherapy with marrow reinfusion and hyperfractionated irradiation for children with high-risk brain tumors. Med Pediatr Oncol 1994;23:428–436.
360. Poplack D, Riccardi R. Pharmacologic approaches to the treatment of central nervous system malignancy. Boston: Martinus Nijhoff, 1987.
361. Chabner B. The role of drugs in cancer treatment. Philadelphia: WB Saunders, 1982.
362. Collins J. Regional therapy: an overview. Boston: Martinus Nijhoff, 1987.
363. Blasberg RG, Patlak C, Fenstermacher JD. Intrathecal chemotherapy: brain tissue profiles after ventriculocisternal perfusion. J Pharmacol Exp Ther 1975;195:73–83.
364. Blaney S, Poplack D, Godwin K, et al. The effect of body position on ventricular cerebrospinal fluid methotrexate concentration following intralumbar administration. J Clin Oncol 1995;13:177–179.
365. Blaney S, Berg S, Boddy A. CNS directed drug delivery. London: Arnold, 2004.
366. Westphal M, Hilt DC, Bortey E, et al. A phase 3 trial of local chemotherapy with biodegradable carmustine (BCNU) wafers (Gliadel wafers) in patients with primary malignant glioma. Neuro-oncol 2003;5:79–88.
367. Giese A, Kucinski T, Knopp U, et al. Pattern of recurrence following local chemotherapy with biodegradable carmustine (BCNU) implants in patients with glioblastoma. J Neurooncol 2004;66:351–360.
368. Rhines LD, Sampath P, Dolan ME, et al. O6-benzylguanine potentiates the antitumor effect of locally delivered carmustine against an intracranial rat glioma. Cancer Res 2000;60:6307–6310.
369. Kaiser MG, Parsa AT, Fine RL, et al. Tissue distribution and antitumor activity of topotecan delivered by intracerebral clysis in a rat glioma model. Neurosurgery 2000;47:1391–1398; discussion 1398–1399.
370. Bruce JN, Falavigna A, Johnson JP, et al. Intracerebral clysis in a rat glioma model. Neurosurgery 2000;46:683–691.
371. Weber FW, Floeth F, Asher A, et al. Local convection enhanced delivery of IL4-Pseudomonas exotoxin (NBI-3001) for treatment of patients with recurrent malignant glioma. Acta Neurochir Suppl 2003;88:93–103.
372. Kunwar S, Pai LH, Pastan I. Cytotoxicity and antitumor effects of growth factor-toxin fusion proteins on human glioblastoma multiforme cells. J Neurosurg 1993;79:569–576.
373. Laske DW, Ilercil O, Akbasak A, et al. Efficacy of direct intratumoral therapy with targeted protein toxins for solid human gliomas in nude mice. J Neurosurg 1994;80:520–526.
374. Rand RW, Kreitman RJ, Patronas N, et al. Intratumoral administration of recombinant circularly permuted interleukin-4-Pseudomonas exotoxin in patients with high-grade glioma. Clin Cancer Res 2000;6:2157–2165.
375. Saris SC, Blasberg RG, Carson RE, et al. Intravascular streaming during carotid artery infusions. Demonstration in humans and reduction using diastole-phased pulsatile administration. J Neurosurg 1991;74:763–772.
376. Mahaley MS Jr., Hipp SW, Dropcho EJ, et al. Intracarotid cisplatin chemotherapy for recurrent gliomas. J Neurosurg 1989;70:371–378.
377. Newton HB, Page MA, Junck L, et al. Intra-arterial cisplatin for the treatment of malignant gliomas. J Neurooncol 1989;7:39–45.
378. Kapp J, Vance R, Parker JL, et al. Limitations of high dose intra-arterial 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) chemotherapy for malignant gliomas. Neurosurgery 1982; 10:715–719.
379. Blacklock JB, Wright DC, Dedrick RL, et al. Drug streaming during intra-arterial chemotherapy. J Neurosurg 1986;64:284–291.
380. Ross RL, Kapp JP, Hochberg F, et al. Solvent systems for intracarotid 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) infusion. Neurosurgery 1983;12:512–514.

381. Li XN, Parikh S, Shu Q, et al. Phenylbutyrate and phenylacetate induce differentiation and inhibit proliferation of human medulloblastoma cells. Clin Cancer Res 2004;10:1150–1159.
382. Mukherjee P, Das SK. Antiproliferative action of retinoic acid in cultured human brain tumour cells Gl-As-14(S). Cancer Lett 1990;52:83–89.
383. Agrawal A, Martell LA, Ross DA, et al. Retinoic acid modulation of proliferation and differentiation in brain tumors. Proc Am Assoc Cancer Res 1993;34:20.
384. Rodts GE Jr, Black KL. Trans retinoic acid inhibits in vivo tumour growth of C6 glioma in rats: effect negatively influenced by nerve growth factor. Neurol Res 1994;16:184–186.
385. Stockhammer G, Manley GT, Johnson R, et al. Inhibition of proliferation and induction of differentiation in medulloblastoma- and astrocytoma-derived cell lines with phenylacetate. J Neurosurg 1995;83:672–681.
386. Samid D, Ram Z, Hudgins WR, et al. Selective activity of phenylacetate against malignant gliomas: resemblance to fetal brain damage in phenylketonuria. 1994;54:891–895.
387. Samid D, Shack S, Myers CE. Selective growth arrest and phenotypic reversion of prostate cancer cells in vitro by nontoxic pharmacological concentrations of phenylacetate. J Clin Invest 1993;91:2288–2295.
388. Liu L, Shack S, Stetler-Stevenson WG, et al. Differentiation of cultured human melanoma cells induced by the aromatic fatty acids phenylacetate and phenylbutyrate. J Invest Dermatol 1994;103:335–340.
389. Sidell N, Wada R, Han G, et al. Phenylacetate synergizes with retinoic acid in inducing the differentiation of human neuroblastoma cells. Int J Cancer 1995;60:507–514.
390. Carducci MA, Nelson JB, Chan-Tack KM, et al. Phenylbutyrate induces apoptosis in human prostate cancer and is more potent than phenylacetate. Clin Cancer Res 1996;2:379–387.
391. Hallahan AR, Pritchard JI, Chandraratna RA, et al. BMP-2 mediates retinoid-induced apoptosis in medulloblastoma cells through a paracrine effect. Nat Med 2003;9:1033–1038.
392. Gumireddy K, Sutton LN, Phillips PC, et al. All-trans-retinoic acid-induced apoptosis in human medulloblastoma: activation of caspase-3/poly(ADP-ribose) polymerase 1 pathway. Clin Cancer Res 2003;9:4052–4059.
393. Hoi Sang U, Espiritu OD, Kelley PY, et al. The role of the epidermal growth factor receptor in human gliomas: I. The control of cell growth. J Neurosurg 1995;82:841–846.
394. Atiba J, Jamil S, Meyskens FL. Transretinoic acid (tRA) in the treatment of malignant gliomas (MG): a phase II study. Proc Am Soc Clin Oncol 1994;13:178.
395. Phuphanich S, Scott C, Fischbach AJ, et al. All-trans-retinoic acid: a phase II Radiation Therapy Oncology Group study (RTOG 91–13) in patients with recurrent malignant astrocytoma. J Neurooncol 1997;34:193–200.
396. Hao Y, Creson T, Zhang L, et al. Mood stabilizer valproate promotes ERK pathway-dependent cortical neuronal growth and neurogenesis. J Neurosci 2004;24:6590–6599.
397. Jansen MS, Nagel SC, Miranda PJ, et al. Short-chain fatty acids enhance nuclear receptor activity through mitogen-activated protein kinase activation and histone deacetylase inhibition. Proc Natl Acad Sci U S A 2004;101:7199–7204.
398. Kieran M, Supko J, MacDonald T, et al. Phase I and pharmacokinetic study of intravenous SU5416 in pediatric central nervous system tumors. Proc Am Soc Clin Oncol 2003;22: 846(abst. 986).
399. Chintagumpala M, Bomgaars L, Aleksic A, et al. Phase I trial of carboplatin and thalidomide in children with refractory or recurrent solid tumors. J Clin Oncol 2004;22:4394–4400.
400. Purow B, Fine HA. Antiangiogenic therapy for primary and metastatic brain tumors. Hematol Oncol Clin North Am 2004;8:1161–1181.
401. Druker B, Tamura S, Buchdunger E, et al. Effects of a selective inhibitor of the Abl tyrosine kinase on the growth of Bcr-Abl positive cells. Nat Med 1996;2:561–566.
402. Buchdunger E, Cioffi C, Law N, et al. Abl protein-tyrosine kinase inhibitor STI571 inhibits in vitro signal transduction mediated by c-kit and platelet-derived growth factor receptors. J Pharmacol Exp Ther 2000;295:139–145.
403. Carroll M, Ohno-Jones S, Tamura S, et al. CGP 57148, a tyrosine kinase inhibitor, inhibits the growth of cells expressing BCR-ABL, TEL-ABL, and TEL-PDGFR fusion proteins. Blood 1997;90:4947–4952.
404. Heinrich M, Griffith D, Druker B, et al. Inhibition of c-kit receptor tyrosine kinase activity by STI 571, a selective tyrosine kinase inhibitor. Blood 2000;96:925–932.
405. Nistér M, Claesson-Welch L, Eriksson A, et al. Differential expression of platelet-derived growth factor receptors in human malignant glioma cell lines. J Biol Chem 1991;266: 16755–16763.
406. Yarden Y, Sliwkowski MX. Untangling the ErbB signalling network. Nat Rev Mol Cell Biol 2001;2:127–137.
407. Pinkas-Kramarski R, Shelly M, Guarino BC, et al. ErbB tyrosine kinases and the two neuregulin families constitute a ligand-receptor network. Mol Cell Biol 1998;18:6090–7101.
408. Sartor CI, Zhou H, Kozlowska E, et al. 3rd Her4 mediates ligand-dependent antiproliferative and differentiation responses in human breast cancer cells. Mol Cell Biol 2001;21:4265–4275.
409. Adam L, Vadlamudi R, Kondapaka SB, et al. Heregulin regulates cytoskeletal reorganization and cell migration through the p21-activated kinase-1 via phosphatidylinositol-3 kinase. J Biol Chem 1998;23:28238–28246.
410. Zhou BP, Hu MC, Miller SA, et al. HER-2/neu blocks tumor necrosis factor-induced apoptosis via the Akt/NF-kappaB pathway. J Biol Chem 2000;275:8027–8031.
411. Yang HY, Zhou BP, Hung MC, et al. Oncogenic signals of HER-2/neu in regulating the stability of the cyclin-dependent kinase inhibitor p27. J Biol Chem 2000;275:24735–24739.
412. Gilbertson R, Hill A, Hernan R, et al. ERBB1 is amplified and overexpressed in high grade diffusely infiltrative pediatric brain stem glioma. Clin Cancer Res 2003;9:3620–3624.
413. Gilbertson RJ, Jaros E, Perry RH, et al. Mitotic percentage index: a new prognostic factor for childhood medulloblastoma. Eur J Cancer 1997;33:609–615.
414. Gilbertson R, Wickramasinghe C, Hernan R, et al. Clinical and molecular stratification of disease risk in medulloblastoma. Br J Cancer 2001;85:705–712.
415. Gilbertson RJ, Clifford SC, MacMeekin W, et al. Expression of the ErbB-neuregulin signaling network during human cerebellar development: implications for the biology of medulloblastoma. Cancer Res 1998;58:3932–3941.
416. Yoffey JM, Courtice FC. Lymphatics, lymph and the lymphomyeloid complex. New York: Academic Press, 1970.
417. Daar AS, Fuggle SV, Fabre JW, et al. The detailed distribution of MHC Class II antigens in normal human organs. Transplantation 1984;38:293–298.
418. Pollack IF, Okada H, Chambers WH. Exploitation of immune mechanisms in the treatment of central nervous system cancer. Semin Pediatr Neurol 2000;7:131–143.
419. Barker CF, Billingham RE. Immunologically privileged sites. Adv Immunol 1977;25:1–54.
420. Medawar PB. Immunity to homologous grafted skin, III: the fate of skin homografts transplanted to the brain, to subcutaneous tissue, and to the anterior chamber of the eye. Br J Exp Pathol 1948;29:58–69.
421. Fathallah-Shaykh H. New molecular strategies to cure brain tumors. Arch Neurol 1999; 56:449–453.
422. Plautz GE, Miller DW, Barnett GH, et al. T cell adoptive immunotherapy of newly diagnosed gliomas. Clin Cancer Res 2000;6:2209–2218.
423. Plautz GE, Barnett GH, Miller DW, et al. Systemic T cell adoptive immunotherapy of malignant gliomas. J Neurosurg 1998;89:42–51.
424. Yates AJ, Stephens RE, Elder PJ, et al. Effects of interferon and gangliosides on growth of cultured human glioma and fetal brain cells. Cancer Res 1985;45:1033–1039.
425. Yung WK, Steck PA, Kelleher PJ, et al. Growth inhibitory effect of recombinant alpha and beta interferon on human glioma cells. J Neurooncol 1987;5:323–330.
426. Cook AW, Carter WA, Nidzgorski F, et al. Human brain tumor–derived cell lines: growth rate reduced by human fibroblast interferon. Science 1983;219:881–883.
427. Mahaley MS Jr., Dropcho EJ, Bertsch L, et al. Systemic beta-interferon therapy for recurrent gliomas: a brief report. J Neurosurg 1989;71:639–641.
428. Nagai M, Arai T. Clinical effect of interferon in malignant brain tumours. Neurosurg Rev 1984;7:55–64.
429. Olson JJ, James CD, Lawson D, et al. Correlation of the response of recurrent malignant gliomas treated with interferon alpha with tumor interferon alpha gene content. Int J Oncol 2004;25:419–427.
430. Mahaley MS Jr, Bertsch L, Cush S, et al. Systemic gamma-interferon therapy for recurrent gliomas. J Neurosurg 1988;69:826–829.
431. Farkkila M, Jaaskelainen J, Kallio M, et al. Randomised, controlled study of intratumoral recombinant gamma-interferon treatment in newly diagnosed glioblastoma. Br J Cancer 1994;70:138–141.
432. Barba D, Saris SC, Holder C, et al. Intratumoral LAK cell and interleukin-2 therapy of human gliomas. J Neurosurg 1989;70:175–182.
433. Yoshida S, Tanaka R, Takai N, et al. Local administration of autologous lymphokine-activated killer cells and recombinant interleukin 2 to patients with malignant brain tumors. Cancer Res 1988;48:5011–5016.
434. Saris SC, Spiess P, Lieberman DM, et al. Treatment of murine primary brain tumors with systemic interleukin-2 and tumor-infiltrating lymphocytes. J Neurosurg 1992;76:513–519.
435. Neuwelt EA, Specht HD, Barnett PA, et al. Increased delivery of tumor-specific monoclonal antibodies to brain after osmotic blood-brain barrier modification in patients with melanoma metastatic to the central nervous system. Neurosurgery 1987;20:885–895.
436. Colapinto EV, Zalutsky MR, Archer GE, et al. Radioimmunotherapy of intracerebral human glioma xenografts with 131I-labeled F(ab’)2 fragments of monoclonal antibody Mel-14. Cancer Res 1990;50:1822–1827.
437. Lashford LS, Davies AG, Richardson RB, et al. A pilot study of 131I monoclonal antibodies in the therapy of leptomeningeal tumors. Cancer 1988;61:857–868.
438. Johnson VG, Wrobel C, Wilson D, et al. Improved tumor-specific immunotoxins in the treatment of CNS and leptomeningeal neoplasia. J Neurosurg 1989;70:240–248.
439. Lowenstein PR, Cowen R, Thomas C, et al. The basic science of brain-tumour gene therapy. Biochem Soc Trans 1999;27:873–881.
440. Trojan J, Johnson TR, Rudin SD, et al. Treatment and prevention of rat glioblastoma by immunogenic C6 cells expressing antisense insulin-like growth factor I RNA. Science 1993;259:94–97.
441. Yu JS, Wei MX, Chiocca EA, et al. Treatment of glioma by engineered interleukin 4-secreting cells. Cancer Res 1993;53:3125–3128.
442. Culver KW, Ram Z, Wallbridge S, et al. In vivo gene transfer with retroviral vector-producer cells for treatment of experimental brain tumors. Science 1992;256:1550–1552.
443. Ram Z, Culver KW, Oshiro EM, et al. Therapy of malignant brain tumors by intratumoral implantation of retroviral vector-producing cells. Nat Med 1997;3:1354–1361.
444. Packer RJ, Raffel C, Villablanca JG, et al. Treatment of progressive or recurrent pediatric malignant supratentorial brain tumors with herpes simplex virus thymidine kinase gene vector-producer cells followed by intravenous ganciclovir administration. J Neurosurg 2000;92:249–254.
445. Glick RP, Lichtor T, Mogharbel A, et al. Intracerebral versus subcutaneous immunization with allogeneic fibroblasts genetically engineered to secrete interleukin-2 in the treatment of central nervous system glioma and melanoma. Neurosurgery 1997;41:898–906.
446. Sampson JH, Ashley DM, Archer GE, et al. Characterization of a spontaneous murine astrocytoma and abrogation of its tumorigenicity by cytokine secretion. Neurosurgery 1997;41:1365–1372.
447. Benedetti S, Bruzzone MG, Pollo B, et al. Eradication of rat malignant gliomas by retroviral-mediated, in vivo delivery of the interleukin 4 gene. Cancer Res 1999;59:645–652.
448. Andreansky S, He B, van Cott J, et al. Treatment of intracranial gliomas in immunocompetent mice using herpes simplex viruses that express murine interleukins. Gene Ther 1998;5:121–130.
449. Bozik M, Okada H, Pollack I. Mechanisms and relative potency of cytokine-modified brain tumor vaccines for eliciting CNS anti tumor immunoreactivity. Proc Soc Neurooncol 1997;2:12.

450. Okada H, Pollack IF, Lieberman F, et al. Gene therapy of malignant gliomas: a pilot study of vaccination with irradiated autologous glioma and dendritic cells admixed with IL-4 transduced fibroblasts to elicit an immune response. Hum Gene Ther 2001;12:575–595.
451. Okada H, Lieberman FS, Edington HD, et al. Autologous glioma cell vaccine admixed with interleukin-4 gene transfected fibroblasts in the treatment of recurrent glioblastoma: preliminary observations in a patient with a favorable response to therapy. J Neurooncol 2003;64:13–20.
452. Ashley DM, Faiola B, Nair S, et al. Bone marrow-generated dendritic cells pulsed with tumor extracts or tumor RNA induce antitumor immunity against central nervous system tumors. J Exp Med 1997;186:1177–1182.
453. Okada H, Tahara H, Shurin MR, et al. Bone marrow-derived dendritic cells pulsed with a tumor-specific peptide elicit effective anti-tumor immunity against intracranial neoplasms. Int J Cancer 1998;78:196–201.
454. Duffner PK, Horowitz ME, Krischer JP, et al. Postoperative chemotherapy and delayed radiation in children less than three years of age with malignant brain tumors. N Engl J Med 1993;328:1725–1731.
455. Duffner PK, Horowitz ME, Krischer JP, et al. The treatment of malignant brain tumors in infants and very young children: an update of the Pediatric Oncology Group experience. Neurooncology 1999;1:152–161.
456. Lashford LS, Campbell RH, Gattamaneni HR, et al. An intensive multiagent chemotherapy regimen for brain tumours occurring in very young children. Arch Dis Child 1996;74:219–223.
457. Marec-Berard P, Jouvet A, Thiesse P, et al. Supratentorial embryonal tumors in children under 5 years of age: an SFOP study of treatment with postoperative chemotherapy alone. Med Pediatr Oncol 2002;38:83–90.
458. Duffner PK, Cohen ME, Myers MH, et al. Survival of children with brain tumors: SEER Program, 1973–1980. Neurology 1986;36:597–601.
459. Mason WP, Goldman S, Yates AJ, et al. Survival following intensive chemotherapy with bone marrow reconstitution for children with recurrent intracranial ependymoma—a report of the Children’s Cancer Group. J Neurooncol 1998;37:135–143.
460. White L, Kellie S, Gray E, et al. Postoperative chemotherapy in children less than 4 years of age with malignant brain tumors: promising initial response to a VETOPEC-based regimen. A Study of the Australian and New Zealand Children’s Cancer Study Group (ANZCCSG). J Pediatr Hematol Oncol 1998;20:125–130.
461. Burger PC, Yu IT, Tihan T, et al. Atypical teratoid/rhabdoid tumor of the central nervous system: a highly malignant tumor of infancy and childhood frequently mistaken for medulloblastoma: a Pediatric Oncology Group study. Am J Surg Pathol 1998;22:1083–1092.
462. Hilden JM, Meerbaum S, Burger P, et al. Central nervous system atypical teratoid/rhabdoid tumor: results of therapy in children enrolled in a registry. J Clin Oncol 2004;22:2877–2884.
463. Cohen BH, Packer RJ, Siegel KR, et al. Brain tumors in children under 2 years: treatment, survival and long-term prognosis. Pediatr Neurosurg 1993;19:171–179.
464. Albright AL, Price RA, Guthkelch AN. Diencephalic gliomas of children. A clinicopathologic study. Contemp Neurosurg 1985;55:2789–2793.
465. Haddad SF, Menezes AH, Bell WE, et al. Brain tumors occurring before 1 year of age: a retrospective reviews of 22 cases in an 11-year period (1977–1987). Neurosurgery 1991;29: 8–13.
466. Reed UC, Rosemberg S, Gherpelli JL, et al. Brain tumors in the first two years of life: a review of forty cases. Pediatr Neurosurg 1993;19:180–185.
467. Tomita T, McLone DG. Brain tumors during the first twenty-four months of life. Neurosurgery 1985;17:913–919.
468. Deutsch M. Radiotherapy for primary brain tumors in very young children. Cancer 1982; 50:2785–2789.
469. Saran FH, Driever PH, Thilmann C, et al. Survival of very young children with medulloblastoma (primitive neuroectodermal tumor of the posterior fossa) treated with craniospinal irradiation. Int J Radiat Oncol Biol Phys 1998;42:959–967.
470. Bloom HJ, Glees J, Bell J, et al. The treatment and long-term prognosis of children with intracranial tumors: a study of 610 cases, 1950–1981. Int J Radiat Oncol Biol Phys 1990;18: 723–745.
471. Kiltie AE, Lashford LS, Gattamaneni HR. Survival and late effects in medulloblastoma patients treated with craniospinal irradiation under three years old. Med Pediatr Oncol 1997;28:348–354.
472. Lannering B, Marky I, Nordborg C. Brain tumors in childhood and adolescence in west Sweden 1970–1984. Epidemiology and survival. Cancer 1990;66:604–609.
473. Mulhern RK, Kepner JL, Thomas PR, et al. Neuropsychologic functioning of survivors of childhood medulloblastoma randomized to receive conventional or reduced-dose craniospinal irradiation: a Pediatric Oncology Group study. J Clin Oncol 1998;16:1723–1728.
474. Mulhern RK, Hancock J, Fairclough D, et al. Neuropsychological status of children treated for brain tumors: a critical review and integrative analysis. Med Pediatr Oncol 1992;20: 181–191.
475. Clayton PE, Price DA, Shalet SM. Growth hormone state after completion of treatment with growth hormone. Arch Dis Child 1987;62:222–226.
476. Shalet SM. Growth and hormonal status of children treated for brain tumours. Childs Brain 1982;9:284–293.
477. Geyer JR, Zeltzer PM, Boyett JM, et al. Survival of infants with primitive neuroectodermal tumors or malignant ependymomas of the CNS treated with eight drugs in 1 day: a report from the Childrens Cancer Group. J Clin Oncol 1994;12:1607–1615.
478. Walter AW, Mulhern RK, Gajjar A, et al. Survival and neurodevelopmental outcome of young children with medulloblastoma at St Jude Children’s Research Hospital. J Clin Oncol 1999;17:3720–3728.
479. Goldwein JW, Radcliffe J, Johnson J, et al. Updated results of a pilot study of low dose craniospinal irradiation plus chemotherapy for children under five with cerebellar primitive neuroectodermal tumors (medulloblastoma). Int J Radiat Oncol Biol Phys 1996;34:899–904.
480. Xu W, Janss A, Packer RJ, et al. Endocrine outcome in children with medulloblastoma treated with 18 Gy of craniospinal radiation therapy. Neurooncology 2004;6:113–118.
481. Duffner PK, Krischer JP, Sanford RA, et al. Prognostic factors in infants and very young children with intracranial ependymomas. Pediatr Neurosurg 1998;28:215–222.
482. Merchant TE, Mulhern RK, Krasin MJ, et al. Preliminary results from a phase II trial of conformal radiation therapy and evaluation of radiation-related CNS effects for pediatric patients with localized ependymoma. J Clin Oncol 2004;22:3156–3162.
483. Ater JL, van Eys J, Woo SY, et al. MOPP chemotherapy without irradiation as primary postsurgical therapy for brain tumors in infants and young children. J Neurooncol 1997;32: 243–252.
484. Strother D, Kepner J, Aronin P, et al. Dose-intensive chemotherapy with ependymoma results of Pediatric Oncology Group study 9233. Proc Am Soc Clin Oncol 2000;19:585a.
485. Sanford RA, Horowitz ME, Kun LE, et al. Preoperative chemotherapy to facilitate the total resection of pediatric brain tumors. Concepts Pediatr Neurosurg 1989;9:139–152.
486. Biegel JA, Tan L, Zhang F, et al. Alterations of the hSNF5/INI1 gene in central nervous system atypical teratoid/rhabdoid tumors and renal and extrarenal rhabdoid tumors. Clin Cancer Res 2002;8:3461–3467.
487. McNeil DE, Cote TR, Clegg L, et al. Incidence and trends in pediatric malignancies medulloblastoma/primitive neuroectodermal tumor: a SEER update. Surveillance Epidemiology and End Results. Med Pediatr Oncol 2002;39:190–194.
488. Rickert CH, Paulus W. Epidemiology of central nervous system tumors in childhood and adolescence based on the new WHO classification. Childs Nerv Syst 2001;17:503–511.
489. Russell DS, Rubinstein LJ. Pathology of tumors of the nervous system, 5th ed. Baltimore: Williams & Wilkins, 1989.
490. Giangaspero F, Rigobello L, Badiali M, et al. Large-cell medulloblastomas. A distinct variant with highly aggressive behavior. Am J Surg Pathol 1992;16:687–693.
491. Brown HG, Kepner JL, Perlman EJ, et al. “Large cell/anaplastic” medulloblastomas: a Pediatric Oncology Group Study. J Neuropathol Exp Neurol 2000;59:857–865.
492. Eberhart C, Kratz J, Schuster A, et al. Comparative genomic hybridization detects an increased number of chromosomal alterations in large cell/anaplastic medulloblastomas. Brain Pathol 2002;12:36–44.
493. Allen JC, Epstein F. Medulloblastoma and other primary malignant neuroectodermal tumors of the CNS. The effect of patients’ age and extent of disease on prognosis. J Neurosurg 1982;57:446–451.
494. Deutsch M, Reigel DH. The value of myelography in the management of childhood medulloblastoma. Cancer 1980;45:2194–2197.
495. Kleinman GM, Hochberg FH, Richardson EP Jr. Systemic metastases from medulloblastoma: report of two cases and review of the literature. Cancer 1981;48:2296–2309.
496. Campbell AN, Chan HS, Becker LE, et al. Extracranial metastases in childhood primary intracranial tumors. A report of 21 cases and review of the literature. Cancer 1984;53:974–981.
497. Tarbell NJ, Loeffler JS, Silver B, et al. The change in patterns of relapse in medulloblastoma. Cancer 1991;68:1600–1604.
498. Evans AE, Jenkin RD, Sposto R, et al. The treatment of medulloblastoma. Results of a prospective randomized trial of radiation therapy with and without CCNU, vincristine, and prednisone. J Neurosurg 1990;72:572–582.
499. Eberhart CG, Cohen KJ, Tihan T, et al. Medulloblastomas with systemic metastases: evaluation of tumor histopathology and clinical behavior in 23 patients. J Pediatr Hematol Oncol 2003;25:198–203.
500. Segal RA, Goumnerova LC, Kwon YK, et al. Expression of the neurotrophin receptor TrkC is linked to a favorable outcome in medulloblastoma. Proc Natl Acad Sci U S A 1994;91: 12867–12871.
501. Grotzer MA, Janss AJ, Fung K, et al. TrkC expression predicts good clinical outcome in primitive neuroectodermal brain tumors. J Clin Oncol 2000;18:1027–1035.
502. Kim JY, Sutton ME, Lu DJ, et al. Activation of neurotrophin-3 receptor TrkC induces apoptosis in medulloblastomas. Cancer Res 1999;59:711–719.
503. Gajjar A, Hernan R, Kocak M, et al. Clinical, histopathologic, and molecular markers of prognosis: toward a new disease risk stratification system for medulloblastoma. J Clin Oncol 2004;22:984–993.
504. Duffner PK, Kun LE, Burger PC, et al. Postoperative chemotherapy and delayed radiation in infants and very young children with choroid plexus carcinomas. The Pediatric Oncology Group. Pediatr Neurosurg 1995;22:189–196.
505. Kortmann RD, Kuhl J, Timmermann B, et al. Postoperative neoadjuvant chemotherapy before radiotherapy as compared to immediate radiotherapy followed by maintenance chemotherapy in the treatment of medulloblastoma in childhood: results of the German prospective randomized trial HIT ’91. Int J Radiat Oncol Biol Phys 2000;46:269–279.
506. Packer RJ, Sutton LN, Elterman R, et al. Outcome for children with medulloblastoma treated with radiation and cisplatin, CCNU, and vincristine chemotherapy. J Neurosurg 1994;81:690–698.
507. Bouffet E, Gentet JC, Doz F, et al. Metastatic medulloblastoma: the experience of the French Cooperative M7 Group. Eur J Cancer 1994;30A:1478–1483.
508. Albright AL, Wisoff JH, Zeltzer PM, et al. Effects of medulloblastoma resections on outcome in children: a report from the Children’s Cancer Group. Neurosurgery 1996;38:265–271.
509. Packer RJ, Goldwein J, Nicholson HS, et al. Treatment of children with medulloblastomas with reduced-dose craniospinal radiation therapy and adjuvant chemotherapy: a Children’s Cancer Group Study. J Clin Oncol 1999;17:2127–2136.
510. Zeltzer PM, Boyett JM, Finlay JL, et al. Metastasis stage, adjuvant treatment, and residual tumor are prognostic factors for medulloblastoma in children: conclusions from the Children’s Cancer Group 921 randomized phase III study. J Clin Oncol 1999;17:832–845.
511. Gajjar A, Fouladi M, Walter AW, et al. Comparison of lumbar and shunt cerebrospinal fluid specimens for cytologic detection of leptomeningeal disease in pediatric patients with brain tumors. J Clin Oncol 1999;17:1825–1828.
512. Bailey CC, Gnekow A, Wellek S, et al. Prospective randomised trial of chemotherapy given before radiotherapy in childhood medulloblastoma. International Society of Paediatric Oncology (SIOP) and the (German) Society of Paediatric Oncology (GPO): SIOP II. Med Pediatr Oncol 1995;25:166–178.

513. Jenkin D, Goddard K, Armstrong D, et al. Posterior fossa medulloblastoma in childhood: treatment results and a proposal for a new staging system. Int J Radiat Oncol Biol Phys 1990;19:265–274.
514. Krischer JP, Ragab AH, Kun L, et al. Nitrogen mustard, vincristine, procarbazine, and prednisone as adjuvant chemotherapy in the treatment of medulloblastoma. A Pediatric Oncology Group study. J Neurosurg 1991;74:905–909.
515. McLaurin RL. Disadvantages of the preoperative shunt in posterior fossa tumors. Clin Neurosurg 1983;30:286–292.
516. Albright AL. The value of precraniotomy shunts in children with posterior fossa tumors. Clin Nuerosurg 1983;30:278–285.
517. Berger MS, Baumeister B, Geyer JR, et al. The risks of metastases from shunting in children with primary central nervous system tumors. J Neurosurg 1991;74:872–877.
518. Tomita T, Yasue M, Engelhard HH, et al. Flow cytometric DNA analysis of medulloblastoma. Prognostic implication of aneuploidy. Cancer 1988;61:744–749.
519. Jung HL, Wang KC, Kim SK, et al. Loss of heterozygosity analysis of chromosome 17p13.1–13.3 and its correlation with clinical outcome in medulloblastomas. J Neurooncol 2004;67:41–46.
520. Scheurlen WG, Schwabe GC, Joos S, et al. Molecular analysis of childhood primitive neuroectodermal tumors defines markers associated with poor outcome. J Clin Oncol 1998;16:2478–2485.
521. Pingoud-Meier C, Lang D, Janss AJ, et al. Loss of caspase-8 protein expression correlates with unfavorable survival outcome in childhood medulloblastoma. Clin Cancer Res 2003;9:6401–6409.
522. Fernandez-Teijeiro A, Betensky RA, Sturla LM, et al. Combining gene expression profiles and clinical parameters for risk stratification in medulloblastomas. J Clin Oncol 2004;22: 994–998.
523. Ellison DW, Clifford SC, Gajjar A, et al. What’s new in neuro-oncology? Recent advances in medulloblastoma. Eur J Paediatr Neurol 2003;7:53–66.
524. Wisoff JH, Epstein FJ. Pseudobulbar palsy after posterior fossa operation in children. Neurosurgery 1984;15:707–709.
525. Pollack IF, Lunsford LD, Slamovits TL, et al. Stereotaxic intracavitary irradiation for cystic craniopharyngiomas. J Neurosurg 1988;68:227–233.
526. Bouffet E, Bernard JL, Frappaz D, et al. M4 protocol for cerebellar medulloblastoma: supratentorial radiotherapy may not be avoided. Int J Radiat Oncol Biol Phys 1992;24:79–85.
527. Thomas PR, Deutsch M, Kepner JL, et al. Low-stage medulloblastoma: final analysis of trial comparing standard-dose with reduced-dose neuraxis irradiation. J Clin Oncol 2000;18:3004–3011.
528. Wolden SL, Dunkel IJ, Souweidane MM, et al. Patterns of failure using a conformal radiation therapy tumor bed boost for medulloblastoma. J Clin Oncol 2003;21:3079–3083.
529. Merchant TE, Kun LE, Krasin MJ, et al. Multi-institution prospective trial of reduced-dose craniospinal irradiation (23.4 Gy) followed by conformal posterior fossa (36 Gy) and primary site irradiation (55.8 Gy) and dose-intensive chemotherapy for average-risk medulloblastoma. Proceedings of the 45th Annual ASTRO Meeting 2003;57:S194–S195.
530. Douglas JG, Barker JL, Ellenbogen RG, et al. Concurrent chemotherapy and reduced-dose cranial spinal irradiation followed by conformal posterior fossa tumor bed boost for average-risk medulloblastoma: efficacy and patterns of failure. Int J Radiat Oncol Biol Phys 2004;58:1161–1164.
531. Miralbell R, Bleher A, Huguenin P, et al. Pediatric medulloblastoma: radiation treatment technique and patterns of failure. Int J Radiat Oncol Biol Phys 1997;37:523–529.
532. Allen JC, Helson L, Jereb B. Preradiation chemotherapy for newly diagnosed childhood brain tumors. A modified Phase II trial. Cancer 1983;52:2001–2006.
533. Loeffler JS, Kretschmar CS, Sallan SE, et al. Pre-radiation chemotherapy for infants and poor prognosis children with medulloblastoma. Int J Radiat Oncol Biol Phys 1988;15:177–181.
534. Pendergrass TW, Milstein JM, Geyer JR, et al. Eight drugs in one day chemotherapy for brain tumors: experience in 107 children and rationale for preradiation chemotherapy. J Clin Oncol 1987;5:1221–1231.
535. Kovnar EH, Kellie SJ, Horowitz ME, et al. Preirradiation cisplatin and etoposide in the treatment of high-risk medulloblastoma and other malignant embryonal tumors of the central nervous system: a phase II study. J Clin Oncol 1990;8:330–336.
536. Mosijczuk AD, Nigro MA, Thomas PR, et al. Preradiation chemotherapy in advanced medulloblastoma. A Pediatric Oncology Group pilot study. Cancer 1993;72:2755–2762.
537. Heideman RL, Kovnar EH, Kellie SJ, et al. Preirradiation chemotherapy with carboplatin and etoposide in newly diagnosed embryonal pediatric CNS tumors. J Clin Oncol 1995;13:2247–2254.
538. Taylor RE, Bailey CC, Robinson K, et al. Results of a randomized study of preradiation chemotherapy versus radiotherapy alone for nonmetastatic medulloblastoma: The International Society of Paediatric Oncology/United Kingdom Children’s Cancer Study Group PNET-3 Study. J Clin Oncol 2003;21:1581–1591.
539. Hartsell WF, Gajjar A, Heideman RL, et al. Patterns of failure in children with medulloblastoma: effects of preirradiation chemotherapy. Int J Radiat Oncol Biol Phys 1997;39: 15–24.
540. Kun LE. Medulloblastoma—challenges in radiation therapy and the addition of chemotherapy. Int J Radiat oncol Biol Phys 2000;46:261–263.
541. Packer RJ, Sutton LN, Goldwein JW, et al. Improved survival with the use of adjuvant chemotherapy in the treatment of medulloblastoma. J Neurosurg 1991;74:433–440.
542. Prados MD, Wara W, Edwards MS, et al. Treatment of high-risk medulloblastoma and other primitive neuroectodermal tumors with reduced dose craniospinal radiation therapy and multi-agent nitrosourea-based chemotherapy. Pediatr Neurosurg 1996;25:174–181.
543. Gaffney CC, Sloane JP, Bradley NJ, et al. Primitive neuroectodermal tumours of the cerebrum. Pathology and treatment. J Neurooncol 1985;3:23–33.
544. Yang HJ, Nam DH, Wang KC, et al. Supratentorial primitive neuroectodermal tumor in children: clinical features, treatment outcome and prognostic factors. Childs Nerv Syst 1999;15:377–383.
545. Pollack IF. Brain tumors in children. N Engl J Med 1994;331:1500–1507.
546. Bruno LA, Rorke LB, Norris DG. Primitive neuroectodermal tumors of infancy and childhood. In: Humphrey GB, Dehner LP, Grindey GB, eds. Pediatric oncology. Boston: Nijhoff, 1981:265–267.
547. Duffner PK, Cohen ME, Heffner RR, et al. Primitive neuroectodermal tumors of childhood. An approach to therapy. J Neurosurg 1981;55:376–381.
548. Dirks PB, Harris L, Hoffman HJ, et al. Supratentorial primitive neuroectodermal tumors in children. J Neurooncol 1996;29:75–84.
549. Paulino AC, Melian E. Medulloblastoma and supratentorial primitive neuroectodermal tumors: an institutional experience. Cancer 1999;86:142–148.
550. Cohen BH, Zeltzer PM, Boyett JM, et al. Prognostic factors and treatment results for supratentorial primitive neuroectodermal tumors in children using radiation and chemotherapy: a Children’s Cancer Group randomized trial. J Clin Oncol 1995;13:1687–1696.
551. Tomita T, McLone DG, Yasue M. Cerebral primitive neuroectodermal tumors in childhood. J Neurooncol 1988;6:233–243.
552. Knapp J, Daisy F, van Eys J. Primitive neuroectodermal tumors of brain in childhood: literature review and the MD Anderson experience. In: Humphrey GB, Dehner LP, Grindey GB, eds. Pediatric oncology. The Hague: Martinus Nijhoff, 1981:215.
553. Parker JC Jr, Mortara RH, McCloskey JJ. Biological behavior of the primitive neuroectodermal tumors: significant supratentorial childhood gliomas. Surg Neurol 1975;4:383–388.
554. Humphrey GB, Dehner LP, Kaplan RJ. Overview on the management of primitive neuroectodermal tumors. In: Humphrey GB, Dehner LP, Grindey GB, eds. Pediatric oncology. The Hague: Martinus Nijhoff, 1981:289.
555. Wald B, Siegle SE, Isaacs H. Cerebral primitive neuroectodermal tumor (primary cerebral neuroblastoma). In: Humphrey GB, Dehner LP, Grindey GB, eds. Pediatric oncology. The Hague: Martinus Nijhoff, 1981:229.
556. Wara WM, Edwards MS, Surti NR. Primary cerebral neuroblastoma. In: Humphrey GB, Dehner LP, Grindey GB, eds. Pediatric oncology. The Hague: Martinus Nijhoff, 1981: 225.
557. Baum ES, Morgan ER, DalCanto MC. Review and experience with primitive neuroectodermal tumors of childhood. In: Humphrey GB, Dehner LP, Grindey GB, eds. Pediatric oncology. The Hague: Martinus Nijhoff, 1981:238.
558. Priest J, Dehner LP, Sung JH. Primitive neuroectodermal tumors (embryonal gliomas) of childhood: a clinicopathologic study of 12 cases. In: Humphrey GB, Dehner LP, Grindey GB, eds. Pediatric oncology. The Hague: Martinus Nijhoff, 1981:247–264.
559. Ganti SR, Silver AJ, Diefenbach P, et al. Computed tomography of primitive neuroectodermal tumors. Am J Neuroradiol 1983;4:819–821.
560. Jenkin D. Primitive neuroectodermal tumors of childhood. In: Humphrey GB, Dehner LP, Grindey GB, eds. Pediatric oncology. The Hague: Martinus Nijhoff, 1981:243.
561. Jakacki RI. Pineal and nonpineal supratentorial primitive neuroectodermal tumors. Childs Nerv Syst 1999;15:586–591.
562. Reddy AT, Janss AJ, Phillips PC, et al. Outcome for children with supratentorial primitive neuroectodermal tumors treated with surgery, radiation, and chemotherapy. Cancer 2000;88:2189–2193.
563. Kosnik EJ, Boesel CP, Bay J, et al. Primitive neuroectodermal tumors of the central nervous system in children. J Neurosurg 1978;48:741–746.
564. Berger MS, Edwards MS, Wara WM, et al. Primary cerebral neuroblastoma. Long-term follow-up review and therapeutic guidelines. J Neurosurg 1983;59:418–423.
565. Sexauer CP, Drous HF, Kaplan RJ. Supratentorial primitive neuroectodermal tumor: clinical response to vincristine, cyclophosphamide, and BCNU. In: Humphrey GB, Dehner LP, Grindey GB, eds. Pediatric oncology. The Hague: Martinus Nijhoff, 1981:235.
566. Ashley DM, Longee D, Tien R, et al. Treatment of patients with pineoblastoma with high dose cyclophosphamide. Med Pediatr Oncol 1996;26:387–392.
567. Judkins AR, Mauger J, Ht A, et al. Immunohistochemical analysis of hSNF5/INI1 in pediatric CNS neoplasms. Am J Surg Pathol 2004;28:644–650.
568. Tekautz T, Fuller C, Fouladi M, et al. Atypical teratoid/rhabdoid tumors (AT/RT): improved survival in children .3 years old with radiation therapy and high dose alkylator based chemotherapy. J Clin Oncol 2005;23:1491–1499.
569. Packer RJ, Biegel JA, Blaney S, et al. Atypical teratoid/rhabdoid tumor of the central nervous system: report on workshop. J Pediatr Hematol Oncol 2002;24:337–342.
570. Perilongo G, Massimino M, Sotti G, et al. Analyses of prognostic factors in a retrospective review of 92 children with ependymoma: Italian Pediatric Neuro-oncology Group. Med Pediatr Oncol 1997;29:79–85.
571. Sala F, Talacchi A, Mazza C, et al. Prognostic factors in childhood intracranial ependymomas: the role of age and tumor location. Pediatr Neurosurg 1998;28:135–142.
572. Kun LE, Kovnar EH, Sanford RA. Ependymomas in children. Pediatr Neurosci 1988;14: 57–63.
573. Bouffet E, Perilongo G, Canete A, et al. Intracranial ependymomas in children: a critical review of prognostic factors and a plea for cooperation. Med Pediatr Oncol 1998;30:319–329.
574. Goldwein JW, Leahy JM, Packer RJ, et al. Intracranial ependymomas in children. Int J Radiat Oncol Biol Phys 1990;19:1497–1502.
575. Bergsagel DJ, Finegold MJ, Butel JS, et al. DNA sequences similar to those of simian virus 40 in ependymomas and choroid plexus tumors of childhood. N Engl J Med 1992;326: 988–993.
576. Merchant TE, Jenkins JJ, Burger PC, et al. Influence of tumor grade on time to progression after irradiation for localized ependymoma in children. Int J Radiat Oncol Biol Phys 2002;53:52–57.
577. Ross GW, Rubinstein LJ. Lack of histopathological correlation of malignant ependymomas with postoperative survival. J Neurosurg 1989;70:31–36.
578. Rawlings CE III, Giangaspero F, Burger PC, et al. Ependymomas: a clinicopathologic study. Surg Neurol 1988;29:271–281.
579. Vanuytsel L, Brada M. The role of prophylactic spinal irradiation in localized intracranial ependymoma. Int J Radiat Oncol Biol Phys 1991;21:825–830.
580. Merchant TE, Haida T, Wang MH, et al. Anaplastic ependymoma: treatment of pediatric patients with or without craniospinal radiation therapy. J Neurosurg 1997;86:943–949.
581. Hukin J, Epstein F, Lefton D, et al. Treatment of intracranial ependymoma by surgery alone. Pediatr Neurosurg 1998;29:40–45.

582. Korshunov A, Golanov A, Sycheva R, et al. The histologic grade is a main prognostic factor for patients with intracranial ependymomas treated in the microneurosurgical era: an analysis of 258 patients. Cancer 2004;100:1230–1237.
583. McLaughlin MP, Buatti JM, Marcus RB Jr, et al. Outcome after radiotherapy of primary spinal cord glial tumors. Radiat Oncol Investig 1998;6:276–280.
584. Needle MN, Goldwein JW, Grass J, et al. Adjuvant chemotherapy for the treatment of intracranial ependymoma of childhood. Cancer 1997;80:341–347.
585. Hirose Y, Aldape K, Bollen A, et al. Chromosomal abnormalities subdivide ependymal tumors into clinically relevant groups. Am J Pathol 2001;158:1137–1143.
586. Nagib MG, O’Fallon MT. Posterior fossa lateral ependymoma in childhood. Pediatr Neurosurg 1996;24:299–305.
587. Sanford RA, Kun LE, Heideman RL, et al. Cerebellar pontine angle ependymoma in infants. Pediatr Neurosurg Pediatr Neurosurg 1997;27:84–91.
588. Figarella-Branger D, Civatte M, Bouvier-Labit C, et al. Prognostic factors in intracranial ependymomas in children. J Neurosurg 2000;93:605–613.
589. Tomita T, McLone DG, Das L, et al. Benign ependymomas of the posterior fossa in childhood. Pediatr Neurosci 1988;14:277–285.
590. Pollack IF, Claassen D, al Shboul Q, et al. Low-grade gliomas of the cerebral hemispheres in children: an analysis of 71 cases. J Neurosurg 1995;82:536–547.
591. Wallner KE, Wara WM, Sheline GE, et al. Intracranial ependymomas: results of treatment with partial or whole brain irradiation without spinal irradiation. Int J Radiat Oncol Biol Phys 1986;12:1937–1941.
592. Merchant TE, Kiehna EN, Thompson SJ, et al. Pediatric low-grade and ependymal spinal cord tumors. Pediatr Neurosurg 2000;32:30–36.
593. Nadkarni TD, Rekate HL. Pediatric intramedullary spinal cord tumors. Critical review of the literature. Childs Nerv Syst 1999;15:17–28.
594. Sexauer CL, Khan A, Burger PC, et al. Cisplatin in recurrent pediatric brain tumors. A POG Phase II study. A Pediatric Oncology Group Study. Cancer 1985;56:1497–1501.
595. Gaynon PS, Ettinger LJ, Moel D, et al. Pediatric phase I trial of carboplatin: a Childrens Cancer Study Group report. Cancer Treat Rep 1987;71:1039–1042.
596. Bouffet E, Foreman N. Chemotherapy for intracranial ependymomas. Childs Nerv Syst 1999;15:563–570.
597. Evans AE, Anderson JR, Lefkowitz-Boudreaux IB, et al. Adjuvant chemotherapy of childhood posterior fossa ependymoma: cranio-spinal irradiation with or without adjuvant CCNU, vincristine, and prednisone: a Childrens Cancer Group study. Med Pediatr Oncol 1996;27:8–14.
598. Burkhard C, Di Patre PL, Schuler D, et al. A population-based study of the incidence and survival rates in patients with pilocytic astrocytoma. J Neurosurg 2003;98:1170–1174.
599. Freeman CR, Farmer JP, Montes J. Low-grade astrocytomas in children: evolving management strategies. Int J Radiat Oncol Biol Phys 1998;41:979–987.
600. Campbell JW, Pollack IF. Cerebellar astrocytomas in children. J Neurooncol 1996;28:223–231.
601. Wallner KE, Gonzales MF, Edwards MS, et al. Treatment results of juvenile pilocytic astrocytoma. J Neurosurg 1988;69:171–176.
602. Farwell JR, Dohrmann GJ, Flannery JT. Central nervous system tumors in children. Cancer 1977;40:3123–3132.
603. Griffin TW, Beaufait D, Blasko JC. Cystic cerebellar astrocytomas in childhood. Cancer 1979;44:276–280.
604. Smoots DW, Geyer JR, Lieberman DM, et al. Predicting disease progression in childhood cerebellar astrocytoma. Childs Nerv Syst 1998;14:636–648.
605. Dohrmann GJ, Farwell JR, Flannery JT. Astrocytomas in childhood: a population-based study. Surg Neurol 1985;23:64–68.
606. Gajjar A, Sanford RA, Heideman R, et al. Low-grade astrocytoma: a decade of experience at St. Jude Children’s Research Hospital. J Cliin Oncol 1997;15:2792–2799.
607. Gajjar A, Bhargava R, Jenkins JJ, et al. Low-grade astrocytoma with neuraxis dissemination at diagnosis. J Neurosurg 1995;83:67–71.
608. Civitello LA, Packer RJ, Rorke LB, et al. Leptomeningeal dissemination of low-grade gliomas in childhood. Neurology 1988;38:562–566.
609. Pollack IF, Hurtt M, Pang D, et al. Dissemination of low grade intracranial astrocytomas in children. Cancer 1994;73:2869–2878.
610. Rutka JT, George RE, Davidson G, et al. Low-grade astrocytoma of the tectal region as an unusual cause of knee pain: case report. Neurosurgery 1991;29:608–612.
611. Kepes JJ, Rubinstein LJ, Eng LF. Pleomorphic xanthoastrocytoma: a distinctive meningocerebral glioma of young subjects with relatively favorable prognosis. A study of 12 cases. Cancer 1979;44:1839–1852.
612. Komotar RJ, Burger PC, Carson BS, et al. Pilocytic and pilomyxoid hypothalamic/chiasmatic astrocytomas. Neurosurgery 2004;54:72–79; discussion 79–80.
613. Tihan T, Fisher PG, Kepner JL, et al. Pediatric astrocytomas with monomorphous pilomyxoid features and a less favorable outcome. J Neuropathol Exp Neurol 1999;58:1061–1068.
614. Becker LE, Yates AJ. Astrocytic tumors in children. In: Finegold M, ed. Pathology of neoplasia in children and adolescents. Philadelphia: WB Saunders, 1986:373.
615. Bernell WR, Kepes JJ, Seitz EP. Late malignant recurrence of childhood cerebellar astrocytoma. Report of two cases. J Neurosurg 1972;37:470–474.
616. Scott RM, Ballantine HT Jr. Cerebellar astrocytoma: malignant recurrence after prolonged postoperative survival. Case report. J Neurosurg 1973;39:777–779.
617. Burger P. Atlas of tumor pathology: tumors of the central nervous system. Washington, DC: Armed Forces Institute of Pathology, 1993.
618. Tice H, Barnes PD, Goumnerova L, et al. Pediatric and adolescent oligodendrogliomas. AJNR Am J Neuroradiol 1993;14:1293–1300.
619. Jay V, Squire J, Becker LE, et al. Malignant transformation in a ganglioglioma with anaplastic neuronal and astrocytic components. Report of a case with flow cytometric and cytogenetic analysis. Cancer 1994;73:2862–2868.
620. Schwartz AM, Ghatak NR. Malignant transformation of benign cerebellar astrocytoma. Cancer 1990;65:333–336.
621. Hirsch JF, Sainte RC, Pierre-Kahn A, et al. Benign astrocytic and oligodendrocytic tumors of the cerebral hemispheres in children. J Neurosurg 1989;70:568–572.
622. Sgouros S, Fineron PW, Hockley AD. Cerebellar astrocytoma of childhood: long-term follow-up. Childs Nerv Syst 1995;11:89–96.
623. Laws ER Jr., Taylor WF, Clifton MB, et al. Neurosurgical management of low-grade astrocytoma of the cerebral hemispheres. J Neurosurg 1984;61:665–673.
624. Winston K, Gilles FH, Leviton A, et al. Cerebellar gliomas in children. J Natl Cancer Inst 1977;58:833–838.
625. Conway PD, Oechler HW, Kun LE, et al. Importance of histologic condition and treatment of pediatric cerebellar astrocytoma. Cancer 1991;67:2772–2775.
626. Gjerris F, Klinken L. Long-term prognosis in children with benign cerebellar astrocytoma. J Neurosurg 1978;49:179–184.
627. Hayostek CJ, Shaw EG, Scheithauer B, et al. Astrocytomas of the cerebellum. A comparative clinicopathologic study of pilocytic and diffuse astrocytomas. Cancer 1993;72:856–869.
628. Shaw EG, Daumas-Duport C, Scheithauer BW, et al. Radiation therapy in the management of low-grade supratentorial astrocytomas. J Neurosurg 1989;70:853–861.
629. Bowers DC, Gargan L, Kapur P, et al. Study of the MIB-1 labeling index as a predictor of tumor progression in pilocytic astrocytomas in children and adolescents. J Clin Oncol 2003;21:2968–2973.
630. Prados MD, Krouwer HG, Edwards MS, et al. Proliferative potential and outcome in pediatric astrocytic tumors. J Neurooncol 1992;13:277–282.
631. Pencalet P, Maixner W, Sainte-Rose C, et al. Benign cerebellar astrocytomas in children. J Neurosurg 1999;90:265–273.
632. Garcia DM, Latifi HR, Simpson JR, et al. Astrocytomas of the cerebellum in children. J Neurosurg 1989;71:661–664.
633. Gjerris F. Clinical aspects and long-term prognosis in supratentorial tumors of infancy and childhood. Acta Neurol Scand 1978;57:445–470.
634. Fazekas JT. Treatment of grades I and II brain astrocytomas. The role of radiotherapy. Int J Radiat Oncol Biol Phys 1977;2:661–666.
635. Vertosick FT Jr, Selker RG, Arena VC. Survival of patients with well-differentiated astrocytomas diagnosed in the era of computed tomography. Neurosurgery 1991;28:496–501.
636. Packer RJ, Sutton LN, Patel KM, et al. Seizure control following tumor surgery for childhood cortical low-grade gliomas. J Neurosurg 1994;80:998–1003.
637. Berger MS, Ghatan S, Haglund MM, et al. Low-grade gliomas associated with intractable epilepsy: seizure outcome utilizing electrocorticography during tumor resection. J Neurosurg 1993;79:62–69.
638. Leibel SA, Sheline GE, Wara WM, et al. The role of radiation therapy in the treatment of astrocytomas. Cancer 1975;35:1551–1557.
639. Garcia DM, Marks JE, Latifi HR, et al. Childhood cerebellar astrocytomas: is there a role for postoperative irradiation? Int J Radiat Oncol Biol Phys 1990;18:815–818.
640. Geissinger JD. Astrocytomas of the cerebellum in children. Long-term study. Arch Neurol 1971;24:125–135.
641. Sutton LN, Schut L. Cerebellar astrocytomas. In: McLaurin RL, Vennes JL, Schut L, eds. Pediatric neurosurgery: surgery of the developing nervous system. 2nd ed. Philadelphia: WB Saunders, 1989:338.
642. Sutton LN, Packer RJ, Rorke LB, et al. Cerebral gangliogliomas during childhood. Neurosurgery 1983;13:124–128.
643. VandenBerg SR. Desmoplastic infantile ganglioglioma and desmoplastic cerebral astrocytoma of infancy. Brain Pathol 1993;3:275–281.
644. Haddad SF, Moore SA, Menezes AH, et al. Ganglioglioma: 13 years of experience. Neurosurgery 1992;31:171–178.
645. Kalyan-Raman UP, Olivero WC. Ganglioglioma: a correlative clinicopathological and radiological study of ten surgically treated cases with follow-up. Neurosurgery 1987;20: 428–433.
646. Chintagumpala MM, Armstrong D, Miki S, et al. Mixed neuronal-glial tumors (gangliogliomas) in children. Pediatr Neurosurg 1996;24:306–313.
647. Cappelli C, Grill J, Raquin M, et al. Long-term follow up of 69 patients treated for optic pathway tumours before the chemotherapy era. Arch Dis Child 1998;79:334–338.
648. Merchant TE, Zhu Y, Thompson SJ, et al. Preliminary results from a Phase II trail of conformal radiation therapy for pediatric patients with localised low-grade astrocytoma and ependymoma. Int J Radiat Oncol Biol Phys 2002;52:325–332.
649. Fouladi M, Wallace D, Langston JW, et al. Survival and functional outcome of children with hypothalamic/chiasmatic tumors. Cancer 2003;15:1084–1092.
650. Tao ML, Barnes PD, Billett AL, et al. Childhood optic chiasm gliomas: radiographic response following radiotherapy and long-term clinical outcome. Int J Radiat Oncol Biol Phys 1997;39:579–587.
651. Hoffman HJ, Humphreys RP, Drake JM, et al. Optic pathway/hypothalamic gliomas: a dilemma in management. Pediatr Neurosurg 1993;19:186–195.
652. Pollack IF, Hoffman HJ, Humphreys RP, et al. The long-term outcome after surgical treatment of dorsally exophytic brain-stem gliomas. J Neurosurg 1993;78:859–863.
653. Dunbar SF, Tarbell NJ, Kooy HM, et al. Stereotactic radiotherapy for pediatric and adult brain tumors: preliminary report. Int J Radiat Oncol Biol Phys 1994;30:531–539.
654. Grabb PA, Lunsford LD, Albright AL, et al. Stereotactic radiosurgery for glial neoplasms of childhood. Neurosurgery 1996;38:696–701; discussion 701–702.
655. Kirsch DG, Tarbell NJ. New technologies in radiation therapy for pediatric brain tumors: the rationale for proton radiation therapy. Pediatr Blood Cancer 2004;42:461–464.
656. Dirks PB, Jay V, Becker LE, et al. Development of anaplastic changes in low-grade astrocytomas of childhood. Neurosurgery 1994;34:68–78.
657. Friedman HS, Krischer JP, Burger P, et al. Treatment of children with progressive or recurrent brain tumors with carboplatin or iproplatin: a Pediatric Oncology Group randomized phase II study. J Clin Oncol 1992;10:249–256.
658. Castello MA, Schiavetti A, Varrasso G, et al. Chemotherapy in low-grade astrocytoma management. Childs Nerv Syst 1998;14:6–9.
659. McCowage G, Tien R, McLendon R, et al. Successful treatment of childhood pilocytic astrocytomas metastatic to the leptomeninges with high-dose cyclophosphamide. Med Pediatr Oncol 1996;27:32–39.
660. Gajjar A, Heideman RL, Kovnar EH, et al. Response of pediatric low grade gliomas to chemotherapy. Pediatr Neurosurg 1993;19:113–118.

661. Packer RJ, Ater J, Allen J, et al. Carboplatin and vincristine chemotherapy for children with newly diagnosed progressive low-grade gliomas. J Neurosurg 1997;86:747–754.
662. Pons MA, Finlay JL, Walker RW, et al. Chemotherapy with vincristine (VCR) and etoposide (VP-16) in children with low-grade astrocytoma. J Neurooncol 1992;14:151–158.
663. Longee DC, Friedman HS, Albright RE Jr, et al. Treatment of patients with recurrent gliomas with cyclophosphamide and vincristine. J Neurosurg 1990;72:583–588.
664. Brown MT, Friedman HS, Oakes WJ, et al. Chemotherapy for pilocytic astrocytomas. Cancer 1993;71:3165–3172.
665. Chamberlain MC. Recurrent cerebellar gliomas: salvage therapy with oral etoposide. J Child Neurol 1997;12:200–204.
666. Prados MD, Edwards MS, Rabbitt J, et al. Treatment of pediatric low-grade gliomas with a nitrosourea-based multiagent chemotherapy regimen. J Neurooncol 1997;32:235–241.
667. Kuo DJ, Weiner HL, Wisoff J, et al. Temozolomide is active in childhood, progressive, unresectable, low-grade gliomas. J Pediatr Hematol Oncol 2003;25:372–378.
668. Quinn JA, Reardon DA, Friedman AH, et al. Phase II trial of temozolomide in patients with progressive low-grade glioma. J Clin Oncol 2003;21:646–651.
669. Weiss L, Sagerman RH, King GA, et al. Controversy in the management of optic nerve glioma. Cancer 1987;59:1000–1004.
670. Alshail E, Rutka JT, Becker LE, et al. Optic chiasmatic-hypothalamic glioma. Brain Pathol 1997;7:799–806.
671. Lewis RA, Gerson LP, Axelson KA, et al. von Recklinghausen neurofibromatosis. II. Incidence of optic gliomata. Ophthalmology 1984;91:929–935.
672. Housepian EM, Chi TL. Neurofibromatosis and optic pathways gliomas. J Neurooncol 1993;15:51–55.
673. Listernick R, Darling C, Greenwald M, et al. Optic pathway tumors in children: the effect of neurofibromatosis type 1 on clinical manifestations and natural history. J Pediatr 1995;127:718–722.
674. Rush JA, Younge BR, Campbell RJ, et al. Optic glioma. Long-term follow-up of 85 histopathologically verified cases. Ophthalmology 1982;89:1213–1219.
675. Danoff BF, Kramer S, Thompson N. The radiotherapeutic management of optic nerve gliomas in children. Int J Radiat Oncol Biol Phys 1980;6:45–50.
676. Oxenhandler DC, Sayers MP. The dilemma of childhood optic gliomas. J Neurosurg 1978;48:34–41.
677. Tenny RT, Laws ER Jr, Younge BR, et al. The neurosurgical management of optic glioma. Results in 104 patients. J Neurosurg 1982;57:452–458.
678. Packer RJ, Savino PJ, Bilaniuk LT, et al. Chiasmatic gliomas of childhood. A reappraisal of natural history and effectiveness of cranial irradiation. Childs Brain 1983;10:393–403.
679. Fletcher WA, Imes RK, Hoyt WF. Chiasmal gliomas: appearance and long-term changes demonstrated by computerized tomography. J Neurosurg 1986;65:154–159.
680. Chan MY, Foong AP, Heisey DM, et al. Potential prognostic factors of relapse-free survival in childhood optic pathway glioma: a multivariate analysis. Pediatr Neurosurg 1998;29:23–28.
681. Alvord EC Jr, Lofton S. Gliomas of the optic nerve or chiasm. Outcome by patients’ age, tumor site, and treatment. J Neurosurg 1988;68:85–98.
682. Berger MS, Deliganis AV, Dobbins J, et al. The effect of extent of resection on recurrence in patients with low grade cerebral hemisphere gliomas. Cancer 1994;74:1784–1791.
683. Imes RK, Hoyt WF. Childhood chiasmal gliomas: update on the fate of patients in the 1969 San Francisco Study. Br J Ophthalmol 1986;70:179–182.
684. Jenkin D, Angyalfi S, Becker L, et al. Optic glioma in children: surveillance, resection, or irradiation? Int J Radiat Oncol Biol Phys 1993;25:215–225.
685. Janss AJ, Grundy R, Cnaan A, et al. Optic pathway and hypothalamic/chiasmatic gliomas in children younger than age 5 years with a 6-year follow-up. Cancer 1995;75:1051–1059.
686. Guillamo JS, Creange A, Kalifa C, et al. Prognostic factors of CNS tumours in Neurofibromatosis 1 (NF1): a retrospective study of 104 patients. Brain 2003;126:152–160.
687. Medlock MD, Madsen JR, Barnes PD, et al. Optic chiasm astrocytomas of childhood. 1. Long-term follow-up. Pediatr Neurosurg 1997;27:121–128.
688. Thiagalingam S, Flaherty M, Billson F, et al. Neurofibromatosis type 1 and optic pathway gliomas: follow-up of 54 patients. Ophthalmology 2004;111:568–577.
689. Wong JY, Uhl V, Wara WM, et al. Optic gliomas. A reanalysis of the University of California, San Francisco experience. Cancer 1987;60:1847–1855.
690. Allen JC. Initial management of children with hypothalamic and thalamic tumors and the modifying role of neurofibromatosis-1. Pediatr Neurosurg 2000;32:154–162.
691. Wisoff JH, Abbott R, Epstein F. Surgical management of exophytic chiasmatic-hypothalamic tumors of childhood. J Neurosurg 1990;73:661–667.
692. Khafaga Y, Hassounah M, Kandil A, et al. Optic gliomas: a retrospective analysis of 50 cases. Int J Radiat Oncol Biol Phys 2003;56:807–812.
693. Packer RJ, Bilaniuk LT, Cohen BH, et al. Intracranial visual pathway gliomas in children with neurofibromatosis. Neurofibromatosis 1988;1:212–222.
694. Dosoretz DE, Blitzer PH, Wang CC, et al. Management of glioma of the optic nerve and/or chiasm: an analysis of 20 cases. Cancer 1980;45:1467–1471.
695. Sutton LN, Molloy PT, Sernyak H, et al. Long-term outcome of hypothalamic/chiasmatic astrocytomas in children treated with conservative surgery. J Neurosurg 1995;83:583–589.
696. Bataini JP, Delanian S, Ponvert D. Chiasmal gliomas: results of irradiation management in 57 patients and review of literature. Int J Radiat Oncol Biol Phys 1991;21:615–623.
697. Packer RJ. Chemotherapy: low-grade gliomas of the hypothalamus and thalamus. Pediatr Neurosurg 2000;32:259–263.
698. Grill J, Laithier V, Rodriguez D, et al. When do children with optic pathway tumours need treatment? An oncological perspective in 106 patients treated in a single centre. Eur J Pediatr 2000;159:692–696.
699. Kestle JR, Hoffman HJ, Mock AR. Moyamoya phenomenon after radiation for optic glioma. J Neurosurg 1993;79:32–35.
700. Loeffler JS, Smith AR, Suit HD. The potential role of proton beams in radiation oncology. Semin Oncol 1997;24:686–695.
701. Packer RJ, Sutton LN, Bilaniuk LT, et al. Treatment of chiasmatic/hypothalamic gliomas of childhood with chemotherapy: an update. Ann Neurol 1988;23:79–85.
702. Mahoney DH Jr., Cohen ME, Friedman HS, et al. Carboplatin is effective therapy for young children with progressive optic pathway tumors: a Pediatric Oncology Group phase II study. Neurooncology 2000;2:213–220.
703. Petronio J, Edwards MS, Prados M, et al. Management of chiasmal and hypothalamic gliomas of infancy and childhood with chemotherapy. J Neurosurg 1991;74:701–708.
704. Kato T, Sawamura Y, Tada M, et al. Cisplatin/vincristine chemotherapy for hypothalamic/visual pathway astrocytomas in young children. J Neurooncol 1998;37:263–270.
705. Walter AW, Gajjar A, Reardon DA, et al. Tamoxifen and carboplatin for children with low-grade gliomas: a pilot study at St. Jude Children’s Research Hospital. J Pediatr Hematol Oncol 2000;22:247–251.
706. Chamberlain MC, Grafe MR. Recurrent chiasmatic-hypothalamic glioma treated with oral etoposide. J Clin Oncol 1995;13:2072–2076.
707. Heideman RL, Kuttesch J Jr, Gajjar AJ, et al. Supratentorial malignant gliomas in childhood: a single institution perspective. Cancer 1997;80:497–504.
708. Sposto R, Ertel IJ, Jenkin RD, et al. The effectiveness of chemotherapy for treatment of high grade astrocytoma in children: results of a randomized trial. A report from the Childrens Cancer Study Group. J Neurooncol 1989;7:165–177.
709. Dropcho EJ, Wisoff JH, Walker RW, et al. Supratentorial malignant gliomas in childhood: a review of fifty cases. Ann Neurol 1987;22:355–364.
710. Marchese MJ, Chang CH. Malignant astrocytic gliomas in children. Cancer 1990;65:2771–2778.
711. Kandt RS, Shinnar S, D’Souza BJ, et al. Cerebrospinal metastases in malignant childhood astrocytomas. J Neurooncol 1984;2:123–128.
712. Winger MJ, Macdonald DR, Cairncross JG. Supratentorial anaplastic gliomas in adults. The prognostic importance of extent of resection and prior low-grade glioma. J Neurosurg 1989;71:487–493.
713. Wood JR, Green SB, Shapiro WR. The prognostic importance of tumor size in malignant gliomas: a computed tomographic scan study by the Brain Tumor Cooperative Group. J Clin Oncol 1988;6:338–343.
714. Burger PC, Heinz ER, Shibata T, et al. Topographic anatomy and CT correlations in the untreated glioblastoma multiforme. J Neurosurg 1988;68:698–704.
715. Burger PC, Dubois PJ, Schold SC Jr., et al. Computerized tomographic and pathologic studies of the untreated, quiescent, and recurrent glioblastoma multiforme. J Neurosurg 1983;58:159–169.
716. Halperin EC, Bentel G, Heinz ER, et al. Radiation therapy treatment planning in supratentorial glioblastoma multiforme: an analysis based on post mortem topographic anatomy with CT correlations. Int J Radiat Oncol Biol Phys 1989;17:1347–1350.
717. Curran WJ Jr., Scott CB, Horton J, et al. Does extent of surgery influence outcome for astrocytoma with atypical or anaplastic foci (AAF)? A report from three Radiation Therapy Oncology Group (RTOG) trials. J Neurooncol 1992;12:219–227.
718. Phuphanich S, Edwards MS, Levin VA, et al. Supratentorial malignant gliomas of childhood. Results of treatment with radiation therapy and chemotherapy. J Neurosurg 1984; 60:495–499.
719. Pollack IF, Boyett JM, Yates AJ, et al. The influence of central review on outcome associations in childhood malignant gliomas: results from the CCG-945 experience. Neuro-oncology 2003;5:197–207.
720. Pollack IF, Hamilton RL, Burnham J, et al. Impact of proliferation index on outcome in childhood malignant gliomas: results in a multi-institutional cohort. Neurosurgery 2002; 50:1238–1244; discussion 1244–1245.
721. Raffel C, Frederick L, O’Fallon JR, et al. Analysis of oncogene and tumor suppressor gene alterations in pediatric malignant astrocytomas reveals reduced survival for patients with PTEN mutations. Clin Cancer Res 1999;5:4085–4090.
722. Bredel M, Pollack IF, Campbell JW, et al. Basic fibroblast growth factor expression as a predictor of prognosis in pediatric high-grade gliomas. Clin Cancer Res 1997;3:2157–2164.
723. Pirzkall A, Li X, Oh J, et al. 3D MRSI for resected high-grade gliomas before RT: tumor extent according to metabolic activity in relation to MRI. Int J Radiat Oncol Biol Phys 2004;59:126–137.
724. Tralins KS, Douglas JG, Stelzer KJ, et al. Volumetric analysis of 18F-FDG PET in glioblastoma multiforme: prognostic information and possible role in definition of target volumes in radiation dose escalation. J Nucl Med 2002;43:1667–1673.
725. Florell RC, Macdonald DR, Irish WD, et al. Selection bias, survival, and brachytherapy for glioma. J Neurosurg 1992;76:179–183.
726. Fontanesi J, Heideman RL, Muhlbauer M, et al. High-activity 125I interstitial irradiation in the treatment of pediatric central nervous system tumors: a pilot study. Pediatr Neurosurg 1995;22:289–297; discussion 98.
727. Loeffler JS, Alexander E 3rd, Shea WM, et al. Radiosurgery as part of the initial management of patients with malignant gliomas. J Clin Oncol 1992;10:1379–1385.
728. Cairncross JG, Macdonald DR, Ramsay DA. Aggressive oligodendroglioma: a chemosensitive tumor. Neurosurgery 1992;31:78–82.
729. Kyritsis AP, Yung WK, Bruner J, et al. The treatment of anaplastic oligodendrogliomas and mixed gliomas. Neurosurgery 1993;32:365–370.
730. Ino Y, Betensky RA, Zlatescu MC, et al. Molecular subtypes of anaplastic oligodendroglioma: implications for patient management at diagnosis. Clin Cancer Res 2001;7:839–845.
731. Pollack IF, Finkelstein SD, Burnham J, et al. Association between chromosome 1p and 19q loss and outcome in pediatric malignant gliomas: results from the CCG-945 cohort. Pediatr Neurosurg 2003;39:114–121.
732. Blaney SM, Phillips PC, Packer RJ, et al. Phase II evaluation of topotecan for pediatric central nervous system tumors. Cancer 1996;78:527–531.
733. Prados MD, Warnick RE, Mack EE, et al. Intravenous carboplatin for recurrent gliomas. A dose-escalating phase II trial. Am J Clin Oncol 1996;19:609–612.
734. Heideman RL, Douglass EC, Langston JA, et al. A phase II study of every other day high-dose ifosfamide in pediatric brain tumors: a Pediatric Oncology Group Study. J Neurooncol 1995;25:77–84.
735. Chastagner P, Sommelet-Olive D, Kalifa C, et al. Phase II study of ifosfamide in childhood brain tumors: a report by the French Society of Pediatric Oncology (SFOP). Med Pediatr Oncol 1993;21:49–53.

736. Walker MD, Alexander E Jr., Hunt WE, et al. Evaluation of BCNU and/or radiotherapy in the treatment of anaplastic gliomas. A cooperative clinical trial. J Neurosurg 1978;49:333–343.
737. Lashford LS, Thiesse P, Jouvet A, et al. Temozolomide in malignant gliomas of childhood: a United Kingdom Children’s Cancer Study Group and French Society for Pediatric Oncology Intergroup Study. J Clin Oncol 2002;20:4684–4691.
738. Wagner S, Erdlenbruch B, Langler A, et al. Oral topotecan in children with recurrent or progressive high-grade glioma: a Phase I/II study by the German Society for Pediatric Oncology and Hematology. Cancer 2004;100:1750–1757.
739. Gajjar A, Chintagumpala MM, Bowers DC, et al. Effect of intrapatient dosage escalation of irinotecan on its pharmacokinetics in pediatric patients who have high-grade gliomas and receive enzyme-inducing anticonvulsant therapy. Cancer 2003;97:2374–2380.
740. Pollack IF, Boyett JM, Finlay JL. Chemotherapy for high-grade gliomas of childhood. Childs Nerv Syst 1999;15:529–544.
741. Grovas AC, Boyett JM, Lindsley K, et al. Regimen-related toxicity of myeloablative chemotherapy with BCNU, thiotepa, and etoposide followed by autologous stem cell rescue for children with newly diagnosed glioblastoma multiforme: report from the Children’s Cancer Group. Med Pediatr Oncol 1999;33:83–87.
742. Bouffet E, Mottolese C, Jouvet A, et al. Etoposide and thiotepa followed by ABMT (autologous bone marrow transplantation) in children and young adults with high-grade gliomas. Eur J Cancer 1997;33:91–95.
743. Chintagumpala M. Responses to procarbazine in newly-diagnosed patients with high-grade gliomas: a Pediatric Oncology Group Study. Proc Am Soc Clin Oncol, 2000;18: 558a.
744. Chintagumpala M, Stewart C, Burger P, et al. Responses to topotecan in newly-diagnosed patients with high-grade gliomas—a Pediatric Oncology Group (POG) study. The 9th International Symposium on Pediatric Neuro-Oncology. June 11–14, 2000; San Francisco, CA. Abstract.
745. Pollack IF, Shultz B, Mulvihill JJ. The management of brainstem gliomas in patients with neurofibromatosis 1. Neurology 1996;46:1652–1660.
746. Molloy PT, Bilaniuk LT, Vaughan SN, et al. Brainstem tumors in patients with neurofibromatosis type 1: a distinct clinical entity. Neurology 1995;45:1897–1902.
747. Robertson PL, Muraszko KM, Brunberg JA, et al. Pediatric midbrain tumors: a benign subgroup of brainstem gliomas. Pediatr Neurosurg 1995;22:65–73.
748. Vandertop WP, Hoffman HJ, Drake JM, et al. Focal midbrain tumors in children. Neurosurgery 1992;31:186–194.
749. Barkovich AJ, Krischer J, Kun LE, et al. Brain stem gliomas: a classification system based on magnetic resonance imaging. Pediatr Neurosurg 1990;16:73–83.
750. May PL, Blaser SI, Hoffman HJ, et al. Benign intrinsic tectal “tumors” in children. J Neurosurg 1991;74:867–871.
751. Fischbein NJ, Prados MD, Wara W, et al. Radiologic classification of brain stem tumors: correlation of magnetic resonance imaging appearance with clinical outcome. Pediatr Neurosurg 1996;24:9–23.
752. Boydston WR, Sanford RA, Muhlbauer MS, et al. Gliomas of the tectum and periaqueductal region of the mesencephalon. Pediatr Neurosurg 1991;17:234–238.
753. Lassman LP, Arjona VE. Pontine gliomas of childhood. Lancet 1967;1:913–915.
754. Epstein FJ, Farmer JP. Brain-stem glioma growth patterns. J Neurosurg 1993;78:408–412.
755. Abbott R, Shiminski-Maher T, Wisoff JH, et al. Intrinsic tumors of the medulla: surgical complications. Pediatr Neurosurg 1991;17:239–244.
756. Epstein F, McCleary EL. Intrinsic brain-stem tumors of childhood: surgical indications. J Neurosurg 1986;64:11–15.
757. Packer RJ, Nicholson HS, Johnson DL, et al. Dilemmas in the management of childhood brain tumors: brainstem gliomas. Pediatr Neurosurg 1991;17:37–43.
758. Robertson PL, Allen JC, Abbott IR, et al. Cervicomedullary tumors in children: a distinct subset of brainstem gliomas. Neurology 1994;44:1798–1803.
759. Freeman CR, Farmer JP. Pediatric brain stem gliomas: a review. Int J Radiat oncol Biol Phys 1998;40:265–271.
760. Sanford RA, Bebin J, Smith RW. Pencil gliomas of the aqueduct of Sylvius. Report of two cases. J Neurosurg 1982;57:690–696.
761. Hoffman HJ, Becker L, Craven MA. A clinically and pathologically distinct group of benign brain stem gliomas. Neurosurgery 1980;7:243–248.
762. Khatib ZA, Heideman RL, Kovnar EH, et al. Predominance of pilocytic histology in dorsally exophytic brain stem tumors. Pediatr Neurosurg 1994;20:2–10.
763. Stroink AR, Hoffman HJ, Hendrick EB, et al. Diagnosis and management of pediatric brain-stem gliomas. J Neurosurg 1986;65:745–750.
764. Soffer D, Sahar A. Cystic glioma of the brain stem with prolonged survival. Neurosurgery 1982;10:499–502.
765. Edwards MS, Wara WM, Ciricillo SF, et al. Focal brain-stem astrocytomas causing symptoms of involvement of the facial nerve nucleus: long-term survival in six pediatric cases. J Neurosurg 1994;80:20–25.
766. Albright AL, Price RA, Guthkelch AN. Brain stem gliomas of children. A clinicopathological study. Cancer 1983;52:2313–2319.
767. Littman P, Jarrett P, Bilaniuk LT, et al. Pediatric brain stem gliomas. Cancer 1980;45:2787–2792.
768. Pierre-Kahn A, Hirsch JF, Vinchon M, et al. Surgical management of brain-stem tumors in children: results and statistical analysis of 75 cases. J Neurosurg 1993;79:845–852.
769. Silbergeld D, Berger M, Griffin B, et al. Brainstem glioma with multiple intraspinal metastases during life: case report and review of the literature. Pediatr Neurosci 1988;14:103–107.
770. Packer RJ, Allen J, Nielsen S, et al. Brainstem glioma: clinical manifestations of meningeal gliomatosis. Ann Neurol 1983;14:177–182.
771. Zagzag D, Miller DC, Knopp E, et al. Primitive neuroectodermal tumors of the brainstem: investigation of seven cases. Pediatrics 2000;106:1045–1053.
772. Freeman CR, Kepner J, Kun LE, et al. A detrimental effect of a combined chemotherapy-radiotherapy approach in children with diffuse intrinsic brain stem gliomas? Int J Radiat Oncol Biol Phys 2000;47:561–564.
773. Packer RJ, Prados M, Phillips P, et al. Treatment of children with newly diagnosed brain stem gliomas with intravenous recombinant beta-interferon and hyperfractionated radiation therapy: a Childrens Cancer Group phase I/II study. Cancer 1996;77:2150–2156.
774. Pollack IF, Pang D, Albright AL. The long-term outcome in children with late-onset aqueductal stenosis resulting from benign intrinsic tectal tumors. J Neurosurg 1994;80:681–688.
775. Squires LA, Allen JC, Abbott R, et al. Focal tectal tumors: management and prognosis. Neurology 1994;44:953–956.
776. Epstein F, Wisoff J. Intra-axial tumors of the cervicomedullary junction. J Neurosurg 1987;67:483–487.
777. Epstein F, Wisoff JH. Intrinsic brainstem tumors in childhood: surgical indications. J Neurooncol 1988;6:309–317.
778. Freeman CR, Krischer JP, Sanford RA, et al. Final results of a study of escalating doses of hyperfractionated radiotherapy in brain stem tumors in children: a Pediatric Oncology Group study. Int J Radiat Oncol Biol Phys 1993;27:197–206.
779. Packer RJ, Boyett JM, Zimmerman RA, et al. Hyperfractionated radiation therapy (72 Gy) for children with brain stem gliomas. A Childrens Cancer Group Phase I/II Trial. Cancer 1993;72:1414–1421.
780. Shrieve DC, Wara WM, Edwards MS, et al. Hyperfractionated radiation therapy for gliomas of the brainstem in children and in adults. Int J Radiat Oncol Biol Phys 1992;24:599–610.
781. Packer RJ, Boyett JM, Zimmerman RA, et al. Outcome of children with brain stem gliomas after treatment with 7800 cGy of hyperfractionated radiotherapy. A Childrens Cancer Group Phase I/II Trial. Cancer 1994;74:1827–1834.
782. Prados MD, Wara WM, Edwards MS, et al. The treatment of brain stem and thalamic gliomas with 78 Gy of hyperfractionated radiation therapy. Int J Radiat Oncol Biol Phys 1995;32:85–91.
783. Bernhard EJ, Mitchell JB, Deen D, et al. Re-evaluating gadolinium(III) texaphyrin as a radiosensitizing agent. Cancer Res 2000;60:86–91.
784. Sanghavi SN, Needle MN, Krailo MD, et al. A phase I study of topotecan as a radiosensitizer for brainstem glioma of childhood: first report of the Children’s Cancer Group-0952. Neuro-oncology 2003;5:8–13.
785. Chakravarti A, Dicker A, Mehta M. The contribution of epidermal growth factor receptor (EGFR) signaling pathway to radioresistance in human gliomas: a review of preclinical and correlative clinical data. Int J Radiat Oncol Biol Phys 2004;58:927–931.
786. Farmer JP, Montes JL, Freeman CR, et al. Brainstem gliomas. A 10-year institutional review. Pediatr Neurosurg 2001;34:206–214.
787. Kun LE. Tumors of the posterior fossa and the spinal canal. In: Halperin EC, Constine LS, Tarbell NJ, et al., eds. Pediatric radiation oncology. Philadelphia: Lippincott, Williams & Wilkins, 2004;107–109.
788. Packer RJ, Lange B, Ater J, et al. Carboplatin and vincristine for recurrent and newly diagnosed low-grade gliomas of childhood. J Clin Oncol 1993;11:850–856.
789. Fulton DS, Levin VA, Wara WM, et al. Chemotherapy of pediatric brain-stem tumors. J Neurosurg 1981;54:721–725.
790. Levin VA, Edwards MS, Wara WM, et al. 5-Fluorouracil and 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea (CCNU) followed by hydroxyurea, misonidazole, and irradiation for brain stem gliomas: a pilot study of the Brain Tumor Research Center and the Childrens Cancer Group. Neurosurgery 1984;14:679–681.
791. Kretschmar CS, Tarbell NJ, Barnes PD, et al. Pre-irradiation chemotherapy and hyperfractionated radiation therapy 66 Gy for children with brain stem tumors. A phase II study of the Pediatric Oncology Group, Protocol 8833. Cancer 1993;72:1404–1413.
792. Jennings MT, Sposto R, Boyett JM, et al. Preradiation chemotherapy in primary high-risk brainstem tumors: phase II study CCG-9941 of the Children’s Cancer Group. J Clin Oncol 2002;20:3431–3437.
793. Jenkin RD, Boesel C, Ertel I, et al. Brain-stem tumors in childhood: a prospective randomized trial of irradiation with and without adjuvant CCNU, VCR, and prednisone. A report of the Childrens Cancer Study Group. J Neurosurg 1987;66:227–233.
794. Chamberlain MC. Recurrent brainstem gliomas treated with oral VP-16. J Neurooncol 1993;15:133–139.
795. Packer RJ. Brain stem gliomas: therapeutic options at time of recurrence. Pediatr Neurosurg 1996;24:211–216.
796. Packer RJ, Cohen BH, Cooney K. Intracranial germ cell tumors. Oncologist 2000;5:312–320.
797. Kretschmar CS. Germ cell tumors of the brain in children: a review of current literature and new advances in therapy. Cancer Invest 1997;15:187–198.
798. Wara WM, Jenkin RD, Evans A, et al. Tumors of the pineal and suprasellar region: Childrens Cancer Study Group treatment results 1960–1975: a report from Childrens Cancer Study Group. Cancer 1979;43:698–701.
799. Edwards MS, Hudgins RJ, Wilson CB, et al. Pineal region tumors in children. J Neurosurg 1988;68:689–697.
800. Jenkin D, Berry M, Chan H, et al. Pineal region germinomas in childhood treatment considerations. Int Radiat Oncol Biol Phys 1990;18:541–545.
801. Abay EO, Laws ER Jr., Grado GL, et al. Pineal tumors in children and adolescents. Treatment by CSF shunting and radiotherapy. J Neurosurg 1981;55:889–895.
802. Jennings MT, Gelman R, Hochberg F. Intracranial germ-cell tumors: natural history and pathogenesis. J Neurosurg 1985;63:155–167.
803. D’Andrea AD, Packer RJ, Rorke LB, et al. Pineocytomas of childhood. A reappraisal of natural history and response to therapy. Cancer 1987;59:1353–1357.
804. Disclafani A, Hudgins RJ, Edwards MS, et al. Pineocytomas. Cancer 1989;63:302–304.
805. Gururangan S, Heideman RL, Kovnar EH, et al. Peritoneal metastases in two patients with pineoblastoma and ventriculo-peritoneal shunts. Med Pediatr Oncol 1994;22:417–420.
806. Matsutani M, Sano K, Takakura K, et al. Primary intracranial germ cell tumors: a clinical analysis of 153 histologically verified cases. J Neurosurg 1997;86:446–455.
807. Balmaceda C, Heller G, Rosenblum M, et al. Chemotherapy without irradiation—a novel approach for newly diagnosed CNS germ cell tumors: results of an international cooperative trial. The First International Central Nervous System Germ Cell Tumor Study. J Clin Oncol 1996;14:2908–2915.

808. Hoffman HJ, Otsubo H, Hendrick EB, et al. Intracranial germ-cell tumors in children. J Neurosurg 1991;74:545–551.
809. Jakacki RI, Zeltzer PM, Boyett JM, et al. Survival and prognostic factors following radiation and/or chemotherapy for primitive neuroectodermal tumors of the pineal region in infants and children: a report of the Childrens Cancer Group. J Clin Oncol 1995;13: 1377–1383.
810. Regis J, Bouillot P, Rouby-Volot F, et al. Pineal region tumors and the role of stereotactic biopsy: review of the mortality, morbidity, and diagnostic rates in 370 cases. Neurosurgery 1996;39:907–912.
811. Knappe UJ, Bentele K, Horstmann M, et al. Treatment and long-term outcome of pineal nongerminomatous germ cell tumors. Pediatr Neurosurg 1998;28:241–245.
812. Ushio Y, Kochi M, Kuratsu J, et al. Preliminary observations for a new treatment in children with primary intracranial yolk sac tumor or embryonal carcinoma. Report of five cases. J Neurosurg 1999;90:133–137.
813. Yoshida J, Sugita K, Kobayashi T, et al. Prognosis of intracranial germ cell tumours: effectiveness of chemotherapy with cisplatin and etoposide (CDDP and VP-16). Acta Neurochir (Wien) 1993;120:111–117.
814. Chapman PH, Linggood RM. The management of pineal area tumors: a recent reappraisal. Cancer 1980;46:1253–1257.
815. Paulino AC, Wen BC, Mohideen MN. Controversies in the management of intracranial germinomas. Oncology (Huntingt) 1999;13:513–521; discussion 521–522, 528–533.
816. Sung DI, Harisliadis L, Chang CH. Midline pineal tumors and suprasellar germinomas: highly curable by irradiation. Radiology 1978;128:745–751.
817. Shibamoto Y, Abe M, Yamashita J, et al. Treatment results of intracranial germinoma as a function of the irradiated volume. Int J Radiat Oncol Biol Phys 1988;15:285–290.
818. Linstadt D, Wara WM, Edwards MS, et al. Radiotherapy of primary intracranial germinomas: the case against routine craniospinal irradiation. Int J Radiat Oncol Biol Phys 1988;15:291–297.
819. Shibamoto Y, Sasai K, Oya N, et al. Intracranial germinoma: radiation therapy with tumor volume-based dose selection. Radiology 2001;218:452–456.
820. Aoyama H, Shirato H, Kakuto Y, et al. Pathologically-proven intracranial germinoma treated with radiation therapy. Radiother Oncol 1998;47:201–205.
821. Calaminus G, Bamberg M, Baranzelli MC, et al. Intracranial germ cell tumors: a comprehensive update of the European data. Neuropediatrics 1994;25:26–32.
822. Merchant TE, Sherwood SH, Mulhern RK, et al. CNS germinoma: disease control and long-term functional outcome for 12 children treated with craniospinal irradiation. Int J Radiat Oncol Biol Phys 2000;46:1171–1176.
823. Bamberg M, Kortmann RD, Calaminus G, et al. Radiation therapy for intracranial germinoma: results of the German cooperative prospective trials MAKEI 83/86/89. J Clin Oncol 1999;17:2585–2592.
824. Baranzelli MC, Patte C, Bouffet E, et al. Nonmetastatic intracranial germinoma: the experience of the French Society of Pediatric Oncology. Cancer 1997;80:1792–1797.
825. Shibamoto Y, Oda Y, Yamashita J, et al. The role of cerebrospinal fluid cytology in radiotherapy planning for intracranial germinoma. Int J Radiat Oncol Biol Phys 1994;29:1089–1094.
826. Bouffet E, Baranzelli MC, Patte C, et al. Combined treatment modality for intracranial germinomas: results of a multicentre SFOP experience. Societe Francaise d’Oncologie Pediatrique. Br J Cancer 1999;79:1199–1204.
827. Shibamoto Y, Takahashi M, Abe M. Reduction of the radiation dose for intracranial germinoma: a prospective study. Br J Cancer 1994;70:984–989.
828. Hardenbergh PH, Golden J, Billet A, et al. Intracranial germinoma: the case for lower dose radiation therapy. Int J Radiat Oncol Biol Phys 1997;39:419–426.
829. Haas-Kogan DA, Missett BT, Wara WM, et al. Radiation therapy for intracranial germ cell tumors. Int J Radiat Oncol Biol Phys 2003;56:511–518.
830. Jereb B, Zupancic N, Petric J. Intracranial germinoma: report of seven cases. Pediatr Hematol Oncol 1990;7:183–188.
831. Patel SR, Buckner JC, Smithson WA, et al. Cisplatin-based chemotherapy in primary central nervous system germ cell tumors. J Neurooncol 1992;12:47–52.
832. Sawamura Y, Shirato H, Ikeda J, et al. Induction chemotherapy followed by reduced-volume radiation therapy for newly diagnosed central nervous system germinoma. J Neurosurg 1998;88:66–72.
833. Paulino AC. Induction chemotherapy and involved-field radiotherapy for intracranial germinoma. J Clin Oncol 2002;20:2911; author reply 2911–2912.
834. Wolden SL, Wara WM, Larson DA, et al. Radiation therapy for primary intracranial germ-cell tumors. Int J Radiat Oncol Biol Phys 1995;32:943–949.
835. Robertson PL, DaRosso RC, Allen JC. Improved prognosis of intracranial non-germinoma germ cell tumors with multimodality therapy. J Neurooncol 1997;32:71–80.
836. Chamberlain MC, Levin VA. Chemotherapeutic treatment of the diencephalic syndrome. A case report. Cancer 1989;63:1681–1684.
837. Kobayashi T, Yoshida J, Ishiyama J, et al. Combination chemotherapy with cisplatin and etoposide for malignant intracranial germ-cell tumors. An experimental and clinical study. J Neurosurg 1989;70:676–681.
838. Allen JC, Bosl G, Walker R. Chemotherapy trials in recurrent primary intracranial germ cell tumors. J Neurooncol 1985;3:147–152.
839. Schild SE, Haddock MG, Scheithauer BW, et al. Nongerminomatous germ cell tumors of the brain. Int J Radiat Oncol Biol Phys 1996;36:557–563.
840. Allen JC, Kim JH, Packer RJ. Neoadjuvant chemotherapy for newly diagnosed germ-cell tumors of the central nervous system. J Neurosurg 1987;67:65–70.
841. Allen JC, DaRosso RC, Donahue B, et al. A phase II trial of preirradiation carboplatin in newly diagnosed germinoma of the central nervous system. Cancer 1994;74:940–944.
842. Pinkerton CR, Broadbent V, Horwich A, et al. ‘JEB’—a carboplatin based regimen for malignant germ cell tumours in children. Br J Cancer 1990;62:257–262.
843. Buckner JC, Peethambaram PP, Smithson WA, et al. Phase II trial of primary chemotherapy followed by reduced-dose radiation for CNS germ cell tumors. J Clin Oncol 1999;17: 933–940.
844. Bryant WP, O’Marcaigh AS, Ledger GA, et al. Aqueous vasopressin infusion during chemotherapy in patients with diabetes insipidus. Cancer 1994;74:2589–2592.
845. Chang TK, Wong TT, Hwang B. Combination chemotherapy with vinblastine, bleomycin, cisplatin, and etoposide (VBPE) in children with primary intracranial germ cell tumors. Med Pediatr Oncol 1995;24:368–372.
846. Kida Y, Kobayashi T, Yoshida J, et al. Chemotherapy with cisplatin for AFP-secreting germ-cell tumors of the central nervous system. J Neurosurg 1986;65:470–475.
847. Itoyama Y, Kochi M, Kuratsu J, et al. Treatment of intracranial nongerminomatous malignant germ cell tumors producing alpha-fetoprotein. Neurosurgery 1995;36:459–464.
848. Bunin GR, Surawicz TS, Witman PA, et al. The descriptive epidemiology of craniopharyngioma. J Neurosurg 1998;89:547–551.
849. Weiss M, Sutton L, Marcial V, et al. The role of radiation therapy in the management of childhood craniopharyngioma. Int J Radiat Oncol Biol Phys 1989;17:1313–1321.
850. Luse SA, Kernohan JW. Squamous-cell nests of the pituitary gland. Cancer 1955;8:623–628.
851. Baskin DS, Wilson CB: Surgical management of craniopharyngiomas. A review of 74 cases. J Neurosurg 1986;65:22–27.
852. Pang D. Surgical management of craniopharyngiomas. New York: Raven Press, 1993.
853. Curtis J, Daneman D, Hoffman HJ, et al. The endocrine outcome after surgical removal of craniopharyngiomas. Pediatr Neurosurg 1994;21[Suppl 1]:24–27.
854. Richmond IL, Wara WM, Wilson CB. Role of radiation therapy in the management of craniopharyngiomas in children. Neurosurgery 1980;6:513–517.
855. Danoff BF, Cowchock FS, Kramer S. Childhood craniopharyngioma: survival, local control, endocrine and neurologic function following radiotherapy. Int J Radiat oncol Biol Phys 1983;9:171–175.
856. Yasargil MG, Curcic M, Kis M, et al. Total removal of craniopharyngiomas. Approaches and long-term results in 144 patients. J Neurosurg 1990;73:3–11.
857. Hoffman HJ, de Silva M, Humphreys RP, et al. Aggressive surgical management of craniopharyngiomas in children. J Neurosurg 1992;76:47–52.
858. Hoffman H, Chuang S, Ehrlich R. The microsurgical removal of craniopharyngiomas in childhood. Concepts Pediatr Neurosurg 1985;6:52.
859. Duff JM, Meyer FB, Ilstrup DM. Long-term outcomes for surgically resected craniopharyngiomas. Neurosurgery 2000;46:291–305.
860. Bruce DA, Schut L, Rorke LB. Craniopharyngiomas in a capsule? Concepts Pediatr Neurosurg 1981;1:29.
861. Carmel PW. Radical removal of craniopharyngioma: 1971 to 1991. J Neurosurg 1993;72: 351a.
862. Hetelekidis S, Barnes PD, Tao ML, et al. 20-year experience in childhood craniopharyngioma. Int J Radiat Oncol Biol Phys 1993;27:189–195.
863. Manaka S, Teramoto A, Takakura K. The efficacy of radiotherapy for craniopharyngioma. J Neurosurg 1985;62:648–656.
864. Cabezudo JM, Vaquero J, Areitio E, et al. Craniopharyngiomas: a critical approach to treatment. J Neurosurg 1981;55:371–375.
865. Fischer EG, Welch K, Shillito J Jr., et al. Craniopharyngiomas in children. Long-term effects of conservative surgical procedures combined with radiation therapy. J Neurosurg 1990;73:534–540.
866. Lyen KR, Grant DB. Endocrine function, morbidity, and mortality after surgery for craniopharyngioma. Arch Dis Child 1982;57:837–841.
867. Laws ER Jr. Transsphenoidal microsurgery in the management of craniopharyngioma. J Neurosurg 1980;52:661–666.
868. Rajan B, Ashley S, Gorman C, et al. Craniopharyngioma—a long-term results following limited surgery and radiotherapy. Radiother Oncol 1993;26:1–10.
869. Merchant TE, Kiehna EN, Sanford RA, et al. Craniopharyngioma: the St. Jude Children’s Research Hospital experience 1984–2001. Int J Radiat Oncol Biol Phys 2002;53:533–542.
870. Regine WF, Kramer S. Pediatric craniopharyngiomas: long term results of combined treatment with surgery and radiation. Int J Radiat Oncol Biol Phys 1992;24:611–617.
871. Scott RM, Hetelekidis S, Barnes PD, et al. Surgery, radiation, and combination therapy in the treatment of childhood craniopharyngioma—a 20-year experience. Pediatr Neurosurg 1994;21[Suppl 1]:75–81.
872. Habrand JL, Ganry O, Couanet D, et al. The role of radiation therapy in the management of craniopharyngioma: a 25-year experience and review of the literature. Int J Radiat Oncol Biol Phys 1999;44:255–263.
873. Brada M, Thomas DG. Craniopharyngioma revisited. Int J Radiat Oncol Biol Phys 1993;27:471–475.
874. Varlotto JM, Flickinger JC, Kondziolka D, et al. External beam irradiation of craniopharyngiomas: long-term analysis of tumor control and morbidity. Int J Radiat Oncol Biol Phys 2002;54:492–499.
875. Kalapurakal JA, Goldman S, Hsieh YC, et al. Clinical outcome in children with craniopharyngioma treated with primary surgery and radiotherapy deferred until relapse. Med Pediatr Oncol 2003;40:214–218.
876. Kalapurakal JA, Goldman S, Hsieh YC, et al. Clinical outcome in children with recurrent craniopharyngioma after primary surgery. Cancer J 2000;6:388–393.
877. Backlund EO. Treatment of craniopharyngiomas: the multimodality approach. Pediatr Neurosurg 1994;21[Suppl 1]:82–89.
878. Lunsford LD, Pollock BE, Kondziolka DS, et al. Stereotactic options in the management of craniopharyngioma. Pediatr Neurosurg 1994;21[Suppl 1]:90–97.
879. Bremer AM, Nguyen TQ, Balsys R. Therapeutic benefits of combination chemotherapy with vincristine, BCNU, and procarbazine on recurrent cystic craniopharyngioma. A case report. J Neurooncol 1984;2:47–51.
880. Takahashi H, Nakazawa S, Shimura T. Evaluation of postoperative intratumoral injection of bleomycin for craniopharyngioma in children. J Neurosurg 1985;62:120–127.
881. Jakacki RI, Cohen BH, Jamison C, et al. Phase II evaluation of interferon-alpha-2a for progressive or recurrent craniopharyngiomas. J Neurosurg 2000;92:255–260.
882. Allen J, Wisoff J, Helson L, et al. Choroid plexus carcinoma—responses to chemotherapy alone in newly diagnosed young children. J Neurooncol 1992;12:69–74.

883. Ellenbogen RG, Winston KR, Kupsky WJ. Tumors of the choroid plexus in children. Neurosurgery 1989;25:327–335.
884. Chow E, Reardon DA, Shah AB, et al. Pediatric choroid plexus neoplasms. Int J Radiat Oncol Biol Phys 1999;44:249–254.
885. Pierga JY, Kalifa C, Terrier-Lacombe MJ, et al. Carcinoma of the choroid plexus: a pediatric experience. Med Pediatr Oncol 1993;21:480–487.
886. Greenberg ML. Chemotherapy of choroid plexus carcinoma. Childs Nerv Syst 1999;15: 571–577.
887. Carpenter DB, Michelsen WJ, Hays AP. Carcinoma of the choroid plexus. Case report. J Neurosurg 1982;56:722–727.
888. Laurence KM. The biology of choroid plexus papilloma in infancy and childhood. Acta Neurochir (Wien)1979;50:79–90.
889. Tomita T, McLone DG, Flannery AM. Choroid plexus papillomas of neonates, infants and children. Pediatr Neurosci 1988;14:23–30.
890. McGirr SJ, Ebersold MJ, Scheithauer BW, et al. Choroid plexus papillomas: long-term follow-up results in a surgically treated series. J Neurosurg 1988;69:843–849.
891. St Clair SK, Humphreys RP, Pillay PK, et al. Current management of choroid plexus carcinoma in children. Pediatr Neurosurg 1991;17:225–233.
892. Johnson DL. Management of choroid plexus tumors in children. Pediatr Neurosci 1989;15:195–206.
893. Milhorat TH, Hammock MK, Davis DA, et al. Choroid plexus papilloma. I. Proof of cerebrospinal fluid overproduction. Childs Brain 1976;2:273–289.
894. Lowis SP, Pizer BL, Coakham H, et al. Chemotherapy for spinal cord astrocytoma: can natural history be modified? Childs Nerv Syst 1998;14:317–321.
895. Lonjon M, Goh KY, Epstein FJ. Intramedullary spinal cord ependymomas in children: treatment, results and follow-up. Pediatr Neurosurg 1998;29:178–183.
896. Allen JC, Aviner S, Yates AJ, et al. Treatment of high-grade spinal cord astrocytoma of childhood with “8-in-1” chemotherapy and radiotherapy: a pilot study of CCG-945. Children’s Cancer Group. J Neurosurg 1998;88:215–220.
897. Mottl H, Koutecky J. Treatment of spinal cord tumors in children. Med Pediatr oncol 1997;29:293–295.
898. Bouffet E, Pierre-Kahn A, Marchal JC, et al. Prognostic factors in pediatric spinal cord astrocytoma. Cancer 1998;83:2391–2399.
899. Peschel RE, Kapp DS, Cardinale F, et al. Ependymomas of the spinal cord. Int J Radiat Oncol Biol Phys 1983;9:1093–1096.
900. Reimer R, Onofrio BM. Astrocytomas of the spinal cord in children and adolescents. J Neurosurg 1985;63:669–675.
901. Hardison HH, Packer RJ, Rorke LB, et al. Outcome of children with primary intramedullary spinal cord tumors. Childs Nerv Syst 1987;3:89–92.
902. Cohen AR, Wisoff JH, Allen JC, et al. Malignant astrocytomas of the spinal cord. J Neurosurg 1989;70:50–54.
903. Cooper PR. Outcome after operative treatment of intramedullary spinal cord tumors in adults: intermediate and long-term results in 51 patients. Neurosurgery 1989;25:855–859.
904. Epstein F, Epstein N. Surgical treatment of spinal cord astrocytomas of childhood. A series of 19 patients. J Neurosurg 1982;57:685–689.
905. O’Sullivan C, Jenkin RD, Doherty MA, et al. Spinal cord tumors in children: long-term results of combined surgical and radiation treatment. J Neurosurg 1994;81:507–512.
906. Muszynski CA, Constantini S, Epstein FJ. Intraspinal intramedullary neoplasms. In: Albright LA, Pollack IF, Adelson PD, eds. Principles and practice of pediatric neurosurgery. New York: Thieme, 1999:697–709.
907. Cristante L, Herrmann HD. Surgical management of intramedullary spinal cord tumors: functional outcome and sources of morbidity. Neurosurgery 1994;35:69–74.
908. Albright AL. Pediatric intramedullary spinal cord tumors. Childs Nerv Syst 1999;15:436–438.
909. Linstadt DE, Wara WM, Leibel SA, et al. Postoperative radiotherapy of primary spinal cord tumors. Int J Radiat Oncol Biol Phys 1989;16:1397–1403.
910. Chun HC, Schmidt-Ullrich RK, Wolfson A, et al. External beam radiotherapy for primary spinal cord tumors. J Neurooncol 1990;9:211–217.
911. Merchant TE, Nguyen D, Thompson SJ, et al. High-grade pediatric spinal cord tumors. Pediatr Neurosurg 1999;30:1–5.
912. Doireau V, Grill J, Zerah M, et al. Chemotherapy for unresectable and recurrent intramedullary glial tumours in children. Brain Tumours Subcommittee of the French Society of Paediatric Oncology (SFOP). Br J Cancer 1999;81:835–840.
913. Linet MS, Ries LA, Smith MA, et al. Cancer surveillance series: recent trends in childhood cancer incidence and mortality in the United States. J Natl Cancer Inst 1999;91:1051–1058.
914. Mulhern RK, Horowitz ME, Kovnar EH, et al. Neurodevelopmental status of infants and young children treated for brain tumors with preirradiation chemotherapy. J Clin Oncol 1989;7:1660–1666.
915. Kao GD, Goldwein JW, Schultz DJ, et al. The impact of perioperative factors on subsequent intelligence quotient deficits in children treated for medulloblastoma/posterior fossa primitive neuroectodermal tumors. Cancer 1994;74:965–971.
916. Duffner PK, Cohen ME. Long-term consequences of CNS treatment for childhood cancer, Part II: Clinical consequences. Pediatr Neurol 1991;7:237–242.
917. Packer RJ, Sutton LN, Atkins TE, et al. A prospective study of cognitive function in children receiving whole-brain radiotherapy and chemotherapy: 2-year results. J Neurosurg 1989;70:707–713.
918. Anderson VA, Godber T, Smibert E, et al. Cognitive and academic outcome following cranial irradiation and chemotherapy in children: a longitudinal study. Br J Cancer 2000;82: 255–262.
919. Johnson DL, McCabe MA, Nicholson HS, et al. Quality of long-term survival in young children with medulloblastoma. J Neurosurg 1994;80:1004–1010.
920. Jannoun L, Bloom HJ. Long-term psychological effects in children treated for intracranial tumors. Int J Radiat Oncol Biol Phys 1990;18:747–753.
921. Silber JH, Radcliffe J, Peckham V, et al. Whole-brain irradiation and decline in intelligence: the influence of dose and age on IQ score. J Clin Oncol 1992;10:1390–1396.
922. Palmer SL, Goloubeva O, Reddick WE, et al. Patterns of intellectual development among survivors of pediatric medulloblastoma: a longitudinal analysis. J Clin Oncol 2001;19: 2302–2308.
923. Palmer SL, Gajjar A, Reddick WE, et al. Predicting intellectual outcome among children treated with 35–40 Gy craniospinal irradiation for medulloblastoma. Neuropsychology 2003;17:548–555.
924. Hoppe-Hirsch E, Renier D, Lellouch-Tubiana A, et al. Medulloblastoma in childhood: progressive intellectual deterioration. Childs Nerv Syst 1990;6:60–65.
925. Dennis M, Spiegler BJ, Hoffman HJ, et al. Brain tumors in children and adolescents—I. Effects on working, associative and serial-order memory of IQ, age at tumor onset and age of tumor. Neuropsychologia 1991;29:813–827.
926. Thompson SJ, Leigh L, Christensen R, et al. Immediate neurocognitive effects of methylphenidate on learning-impaired survivors of childhood cancer. J Clin Oncol 2001;19: 1802–1808.
927. Shalet SM, Gibson B, Swindell R, et al. Effect of spinal irradiation on growth. Arch Dis Child 1987;62:461–464.
928. Duffner PK, Cohen ME, Thomas PR, et al. The long-term effects of cranial irradiation on the central nervous system. Cancer 1985;56:1841–1846.
929. Olshan JS, Gubernick J, Packer RJ, et al. The effects of adjuvant chemotherapy on growth in children with medulloblastoma. Cancer 1992;70:2013–2017.
930. Packer RJ, Boyett JM, Janss AJ, et al. Growth hormone replacement therapy in children with medulloblastoma: use and effect on tumor control. J Clin Oncol 2001;19:480–487.
931. Moshang T Jr., Rundle AC, Graves DA, et al. Brain tumor recurrence in children treated with growth hormone: the National Cooperative Growth Study experience. J Pediatr 1996;128:S4–S7.
932. Swerdlow AJ, Reddingius RE, Higgins CD, et al. Growth hormone treatment of children with brain tumors and risk of tumor recurrence. J Clin Endocrinol Metab 2000;85:4444–4449.
933. Livesey EA, Hindmarsh PC, Brook CG, et al. Endocrine disorders following treatment of childhood brain tumours. Br J Cancer 1990;61:622–625.
934. Belza MG, Donaldson SS, Steinberg GK, et al. Medulloblastoma: freedom from relapse longer than 8 years—a therapeutic cure? J Neurosurg 1991;75:575–582.
935. Kramer ED, Vezina LG, Packer RJ, et al. Staging and surveillance of children with central nervous system neoplasms: recommendations of the Neurology and Tumor Imaging Committees of the Children’s Cancer Group. Pediatr Neurosurg 1994;20:254–262.
936. Torres CF, Rebsamen S, Silber JH, et al. Surveillance scanning of children with medulloblastoma. N Engl J Med 1994;330:892–895.
937. Shaw DW, Geyer JR, Berger MS, et al. Asymptomatic recurrence detection with surveillance scanning in children with medulloblastoma. J Clin Oncol 1997;15:1811–1813.
938. Duffner PK, Cohen ME, Thomas P. Late effects of treatment on the intelligence of children with posterior fossa tumors. Cancer 1983;51:233–237.
939. Jenkin D, Danjoux C, Greenberg M. Subsequent quality of life for children irradiated for a brain tumor before age four years. Med Pediatr Oncol 1998;31:506–511.
940. Varley JM, Evans DG, Birch JM. Li-Fraumeni syndrome—a molecular and clinical review. Br J Cancer 1997;76:1–14.
941. Lasser DM, DeVivo DC, Garvin J, et al. Turcot’s syndrome: evidence for linkage to the adenomatous polyposis coli (APC) locus. Neurology 1994;44:1083–1086.
942. Grill J, Le Deley MC, Gambarelli D, et al. Postoperative chemotherapy without irradiation for ependymoma in children under 5 years of age: a multicenter trial of the French Society of Pediatric Oncology. J Clin Oncol 2001;19:1288–1296.