Epidemiologic Statistics

Malignant gliomas make up 35% to 45% of primary brain tumors, and of these, nearly 85% are glioblastoma multiforme (GBM).⁴ The incidence of anaplastic astrocytoma (AA) peaks in children younger than 10 years of age and then remains constant in each subsequent decade of life. In contrast, GBM is uncommon before the age of 20 years, whereas the rate of occurrence increases dramatically after the age of 40.³ The incidence of malignant gliomas has increased at least twofold among the elderly during the previous two decades.⁵

Current knowledge of the molecular biology and genetics of the astrocytic neoplasms has been well summarized.⁶ Two genetic pathways, a progression pathway and a de novo pathway, that lead to GBM development have been described (Fig. 21-1). Although primary and secondary GBMs are histologically indistinguishable, it is well accepted that distinct molecular aberrations and pathways lead to the formation of primary (de novo) versus secondary GBMs. Primary GBMs are thought to arise in a de novo manner, without any malignant precursor lesion. Amplification of the epidermal growth factor receptor (EGFR) is thought to be a critical event in the molecular cause of primary GBMs. Secondary GBMs, on the other hand, are thought to arise as a result of the malignant transformation of lower grade gliomas, with overexpression of platelet-derived growth factor, fibroblast growth factor–2, and cyclin-dependent kinase 4, as well as p53 mutations and loss of Rb playing major roles in such transformations. Loss of PTEN has been implicated in both pathways, although it is much more common in the pathogenesis of primary GBM.

FIGURE 21-1 • Genetics of gliomas pathway.

Several hereditary syndromes are associated with an increased risk of brain tumors (Fig. 21-2). Neurofibromatosis-1 is the most common familial tumor syndrome associated with brain tumors.⁷ With the exception of Turcot syndrome, all are inherited in an autosomal dominant pattern. In addition to the association between hereditary syndromes and brain tumors, certain families of brain tumor patients have aggregations of brain tumors and extraneural malignancies. Germ-line p53 mutations have been frequently identified in glioma patients with multifocal lesions, a different second primary malignancy, and a family history of cancer.

FIGURE 21-2 • Hereditary syndromes associated with brain tumors.

Although some environmental factors have been linked with brain tumor development, they do not appear to be responsible for most brain tumors.⁸ Apart from a predisposing genetic syndrome that is present in less than 5% of patients with brain tumors, the only other firmly established cause is ionizing radiation. Radiation-induced gliomas have been reported, and 25% of such cases have arisen in children with acute lymphoblastic leukemia who received prophylactic cranial irradiation and chemotherapy.

Anatomy

The brain is divided into supratentorial and infratentorial compartments by the tentorium cerebelli. The supratentorial compartment includes the cerebral hemispheres and the sellar, pineal, and diencephalic regions, whereas the infratentorial compartment includes the midbrain, pons, medulla, and cerebellum. The cerebral hemispheres are connected by the corpus callosum and are divided into the frontal, parietal, occipital, and temporal lobes. The frontal lobes are concerned with behavior organization, planning and association, and speech; the parietal lobes with motor, sensory, and complex intellectual functions; the occipital lobes with vision; and the temporal lobes with behavior, memory, speech, emotion, and auditory-visual pathways.⁹ Most malignant gliomas arise in the cerebral hemispheres, and the lobar distribution is directly related to the amount of white matter present in each region.

The brain is housed within the bony calvarium, made up of the frontal, ethmoid, parietal, sphenoid, temporal, and occipital bones. The posterior margin of the lesser wings of the sphenoid marks the anterior aspect of the lateral cerebral (Sylvian) fissure, the junction of the frontal and temporal lobes of the brain. The base of the skull can be approximated on the surface of the patient by a line, called the orbitomeatal or Reid baseline, which extends from the infraorbital rim to the external auditory meatus.⁹

Pathologic Conditions

Neuropathologists have not agreed on a uniform classification system for brain tumors. However, the fourth edition of the World Health Organization (WHO) classification of tumors of the CNS, published in 2007, is based on the consensus of an international working group of pathologists and geneticists.¹⁰ It is the standard for the definition of brain tumors for clinical oncology and cancer research communities worldwide. According to this classification, CNS tumors are classified as follows: (1) tumors of neuroepithelial tissue, (2) germ cell tumors, (3) tumors of cranial and paraspinal nerves, (4) tumors of the sellar region, (5) tumors of the meninges, (6) lymphomas and hemopoietic neoplasms, and (7) metastatic tumors (Table 21-1).

Table 21-1 World Health Organization Classification of Primary Central Nervous System Tumors

Reliance on a pathologic classification of brain tumors is a requisite for treatment. Indeed, histopathologic examination is more important than anatomic staging in determining the clinical behavior and prognosis of these tumors. Tumor classification is based on histologic evidence of differentiation; however, expression of histologic features may vary widely within tumors of a given type, particularly diffuse gliomas. This regional heterogeneity makes interpretation of small biopsy samples difficult; this is especially true of astrocytomas. Unfortunately, these difficulties can have significant implications on the diagnosis and grading of these tumors.

Histologic grading is a means of predicting the biologic behavior of a neoplasm. In the clinical setting, tumor grade is a key factor influencing the choice of therapies, particularly determining the use of adjuvant radiation and specific chemotherapy protocols. The WHO classification of tumors of the nervous system includes a grading scheme that ranges across a wide variety of neoplasms rather than a strict histologic grading system.¹⁰ Grade I applies to lesions with low proliferative potential and the possibility of cure following surgical resection alone. Neoplasms designated grade II are generally infiltrative in nature and, despite low-level proliferative activity, often recur. Some type II tumors tend to progress to higher grades of malignancy, for example, low-grade diffuse astrocytomas that transform to AA and glioblastoma. Similar transformation occurs in oligodendroglioma and oligoastrocytoma. The designation WHO grade III is generally reserved for lesions with histologic evidence of malignancy, including nuclear atypia and brisk mitotic activity. The designation WHO grade IV is assigned to cytologically malignant, mitotically active, necrosis-prone neoplasms. Examples of grade IV neoplasms include glioblastoma, medulloblastoma, and lymphoma.

Astrocytic gliomas, which are the most common primary brain tumors, arise from astrocytes, the supporting cells of the brain and spinal cord. The cytoplasmic processes that extend from the astrocytes contain a characteristic filamentous protein, glial fibrillary acidic protein (GFAP), which provides an immunohistochemical marker for these tumors. The WHO defines diffusely infiltrative astrocytic tumors with cytologic atypia alone as grade II (diffuse astrocytoma), those also showing anaplasia and mitotic activity as grade III (AA), and tumors additionally showing microvascular proliferation or necrosis as WHO grade IV (Fig. 21-3). Necrosis may be of any type; perinecrotic palisading need not be present.

FIGURE 21-3 • Pathologic conditions of malignant glioma.

A, Anaplastic astrocytoma showing hypercellularity and pleomorphic nuclei. Tumor necrosis is absent. (Hematoxylin and eosin; original magnification ×250.) B, Glioblastoma multiforme with the hallmark features of necrosis with peripheral pseudopalisading of neoplastic nuclei. (Hematoxylin and eosin; original magnification ×250.)

Clinical Presentation

The presenting symptoms and signs of brain tumors are divided into those associated with a mass effect and increased intracranial pressure and those that are focal. The most common presenting symptom is headache, and approximately 20% of patients with supratentorial tumors present with seizures. Alterations in personality, mood, mental capacity, and concentration are frequently seen early in the clinical course.⁴ The focal neurologic symptoms and signs observed with supratentorial brain tumors are summarized in Table 21-2.

Table 21-2 Focal Symptoms

Routes of Spread

Malignant gliomas form as an expansile mass that conforms to the barriers of the cortical convolutions, deep nuclear structures, and adjacent myelinated nerve fibers.¹¹ Microscopic examination typically shows a gradient of infiltrating tumor cells that decreases with the distance from the periphery of the mass. As they enlarge, malignant gliomas extend directly into adjacent lobes, infiltrate throughout the ipsilateral cerebral hemisphere, and disseminate along anatomically defined nerve fiber pathways. In some cases, individual cells, facilitated by the accompanying edema, may infiltrate for long distances from the main tumor mass. Multicentric gliomas are found in less than 5% of patients. Dissemination by seeding through the cerebrospinal fluid pathways occurs in approximately 10% of cases, but is usually a late event, often appearing at a time when its effects are inconsequential compared with those of the recurrent primary tumor mass. Metastases rarely arise outside the CNS.

Diagnostic Studies

Magnetic resonance imaging (MRI) is the modality of choice for diagnosis and evaluation of intracranial neoplasms. It is the ideal modality for initial preoperative diagnosis, including tumor extent, treatment planning, and image-guided therapies because of its multiplanar capability, anatomic detail, and superior resolution.^12,13 The ability to characterize tissue allows for improved assessment of fatty, hemorrhagic, cystic, necrotic, and vascular components, as well as mass effect and location. In malignant gliomas, on T1-weighted sequence enhanced with gadolinium contrast, the tumor appears as an irregular, ringlike configuration that may surround a central area of necrosis (Fig. 21-4). On T2-weighted fluid-attenuation image recovery (FLAIR) images, the changes that reflect edema extend beyond the boundaries of contrast enhancement that may harbor microscopic extension. In fact, tumor may even be found in normal-appearing brain beyond the T2-weighted abnormality.

FIGURE 21-4 • Axial magnetic resonance image of high-grade astrocytoma of right frontal lobe with fluid-attenuation image recovery sequence and T-weighted image with contrast.

Imaging studies are performed within the first 48 hours after surgical resection to determine the presence of residual tumor and to provide a baseline for subsequent treatment. Enhancement resulting from surgical trauma may be indistinguishable from residual tumor even after a complete surgical resection. Postsurgical enhancement develops as early as the fifth postoperative day, peaks after 2 weeks, and may persist for months.¹⁴ Corticosteroids, which act to re-establish the blood-brain barrier, also have a profound effect on the area of enhancement, and diminution in the area of enhancement may be due to corticosteroids alone. To determine a radiographic response to therapy, patients should be on the same or a lower dose of corticosteroids than the dose at the time of the pretreatment scan.

Lacking with conventional MRI is an assessment of physiologic and functional information about the tumor. The gadolinium enhancement, as seen on T1-weighted MRI, reflects regions where there has been a breakdown of the blood-brain barrier. This may not be a reliable indicator of high-grade tumor because of the presence of nonenhancing tumor or contrast-enhancing necrosis.¹⁵ Similarly, T2-defined volume either overestimates or underestimates the microscopic or nonenhancing disease in a majority of patients. New MRI techniques, such as MR perfusion, MR diffusion, MR spectroscopy (MRS), and functional MRI (fMRI), can provide further physiologic characterization of the tumor related to hemodynamics, cellularity, and metabolism, respectively, and may be used to assess response to therapy.¹⁶

MRS is a noninvasive method to evaluate the malignancy of gliomas based on metabolite levels such as N-acetylaspartate (NAA), choline (Cho) compounds, total creatine (Cr), and lactate (Lac), or metabolite ratios such as Cho/NAA and Cho/Cr (Fig. 21-5). Spectroscopy also holds great promise in aiding target definition for radiotherapy treatment planning and evaluation of response to therapy and tumor recurrence.¹⁷ With the advantage of measuring tumor regional variations in abnormalities of metabolite levels, MRS has recently been used as an in vivo molecular imaging technique that assists in targeting and predicts response to radiation therapy for patients with brain gliomas.

FIGURE 21-5 • Single voxel spectroscopy on a patient with high-grade brainstem glioma.

In addition to MRS, brain fMRI has gained potential in clinical application recently as it can image the eloquent cortices, such as the motor cortex, the Broca area, the Wernicke area, and the visual cortex, which are difficult to identify on anatomic MRI. Incorporation of fMRI information into the radiotherapy treatment process can possibly allow a radiation oncologist to properly plan and deliver an adequate radiation dose while avoiding damage to the adjacent functional cortices in the treatment of gliomas.¹⁸

Dynamic MRI techniques can be used to measure important features of tumor vascularity in vivo, including the density and permeability of capillaries. Increased vessel numbers lead to higher regional cerebral (i.e., tumor) blood volume, which can be measured relative to healthy brain tissue by perfusion MRI techniques. Thus, relative cerebral blood volume provides analysis of capillary density in gliomas (Fig. 21-6). In adult astrocytomas, relative cerebral blood volumes correlate well with histopathologic grade and survival. A relative cerebral blood volume value of more than 1.75 is associated with the presence of high-grade tumor and with reduced survival in patients with gliomas.¹⁹

FIGURE 21-6 • Perfusion imaging in a patient with recurrent glioblastoma. A-F represent images taken before and 10 months after therapy, demonstrating decrease in contrast enhancement in T1 images (B and D); rCBV in perfusion imaging (B and E) and K^trans in diffusion imaging (C and F) sequences.

Staging

There is no accepted staging system for gliomas. The American Joint Committee on Cancer (AJCC) proposed a staging scheme for primary brain tumors based on tumor (T) and metastasis (M) (dissemination within and rarely outside the CNS) as well as grade (G). Because this system was not generally adapted to clinical use, it was removed from the 1997 AJCC staging manual and since then has not been used.

Prognostic Factors

A major contribution of the cooperative group brain tumor studies has been the identification of pretreatment characteristics that affect the outcome in patients with malignant gliomas. A nonparametric recursive partitioning technique was applied to an analysis of 1578 patients accrued to three successive Radiation Therapy Oncology Group (RTOG) trials.²⁰ Age, histologic appearance, Karnofsky performance status (KPS), mental status, duration of symptoms, neurologic functional class, extent of surgery, and radiation dose were identified as significant partitioning covariates. As shown in Table 21-3, six patient classes were defined with median survival times ranging from 4.6 to 58.6 months and 2-year survival rates of 4% to 76%. These classes were used to define favorable (classes I and II, 12% of the patients evaluated), intermediate (classes III and IV, 43%), and poor (classes V and VI, 45%) prognosis subgroups. A subsequent reanalysis of the data in glioblastoma patients showed no statistical difference between class V and VI with a median survival time of 7.5 months.²¹ This information is important for interpreting correctly the results of studies comparing different treatment regimens and for assessing the potential of new therapeutic methodologies.

Table 21-3 Malignant Glioma Recursive Partitioning Analysis Classification

Standard Therapeutic Approaches

Radiation Therapy

Randomized trials conducted by the Brain Tumor Cooperative Group (BTCG) and the Scandinavian Glioblastoma Study Group (SGSG) provided seminal evidence that external-beam irradiation favorably affects the outcome of malignant gliomas. BTCG trials 6901 and 7201 demonstrated a significant survival advantage for irradiated patients who received 50 to 60 Gy to the whole brain (in single daily fractions of 1.7-2.0 Gy, 5 days per week) either alone or with chemotherapy compared with those treated with either resection and supportive care only (P = 0.001) or with chemotherapy alone (P < 0.01).^25,26 The median survival of patients receiving 60 Gy was 2.3 times longer than that observed for nonirradiated patients. Nearly 30% of irradiated patients in the SGSG trial maintained a full or partial working capacity, whereas the nonirradiated patients progressively deteriorated, and none regained their original performance level.²⁷ Randomized trials of postoperative radiotherapy in gliomas is shown in Fig. 21-7. These studies were so convincing that virtually all patients with malignant gliomas receive adjuvant radiation therapy.

FIGURE 21-7 • Pooled results of postoperative radiotherapy versus no radiotherapy in gliomas.

Techniques of Radiotherapy

Most malignant gliomas are unifocal at the time of initial presentation, and after treatment the majority recur at or within 1 to 2 cm of their original location.³² Thus, limited treatment portals are used for malignant gliomas. Intracranial metastases that appear after partial brain irradiation do not affect the ultimate outcome, because they are nearly always accompanied by relapse at the primary tumor site. Whole-brain radiation therapy (WBRT) is commonly recommended for patients with multifocal tumors, but even for these lesions, relapses occur most frequently in the sites of known disease.³²

The conventional dose schedule is 59.4 to 60 Gy, given in single daily fractions of 1.8 to 2.0 Gy, 5 days per week. With this scheme, approximately 25% of patients with GBM and 50% of those with AA exhibit a significant radiographic response. Rapid fractionation schemes (such as 30-40 Gy in 10-15 fractions of 2.67-3.0 Gy each) may be appropriate for some elderly or poor-performance-status patients with GBM who have relatively short survival expectancies.

The gross target volume (GTV) represents a three-dimensional (3-D) reconstruction of the location of the tumor, determined by merging data from contrast-enhanced CT scans and MRI studies. The integration of MRI into CT-based treatment planning provides complementary information for accurately defining the GTV. Little agreement exists as to the definition of the clinical target volume (CTV) or the planning target volume (PTV). A shrinking field approach is used in RTOG trials. The initial PTV (PTV1) encompasses the enhancing lesion (GTV) and edema (CTV) with a 2-cm margin. After 46 Gy of a conventionally fractionated treatment course, the PTV (PTV2) is reduced to include only the enhancing lesion with a 2.5 cm margin. Using the extent of edema to define the CTV has some limitations. The identification of peritumoral edema is subjective, and its volume may vary with corticosteroid dosage. An analysis of the patterns of failure using the RTOG-defined PTVs demonstrated that nearly all relapses occurred within the reduced PTV2 at the site of the enhancing tumor.³³ Based on these data, we define the CTV by adding a margin of approximately 2.5 cm around the T1 contrast or a 1.5-cm margin around the FLAIR series. The PTV is determined by adding an additional 0.3- to 0.5-cm margin to the CTV to account for treatment uncertainties.

Patients are fixed in a custom-designed immobilization device, and are simulated and treated in the supine or prone position, depending on tumor location. Field arrangements and beam energy are selected after consideration of the location of the tumor within the brain and the geometry of the PTV. Lateral opposed fields are used only when extensive bilateral tumor involvement is present. This technique is suboptimal for most lesions because a substantial volume of normal brain tissue will be irradiated to the same dose as the tumor. Three-dimensional conformal radiation therapy (3D-CRT) is ideally suited for the treatment of brain tumors. Three-dimensional treatment plans are designed directly from CT (and MRI) image data sets, and the target volume and surrounding normal tissues are displayed using the beam’s-eye view technique. The use of multiple 3-D–planned nonaxial coplanar and noncoplanar fields reduces the volume of normal brain tissue treated to high doses. This approach allows higher than traditional doses to be safely administered to tumors located proximal to critical structures.

Intensity-modulated radiation therapy (IMRT) is an advanced form of 3D-CRT that uses inverse planning and computer-controlled radiation dose deposition. The advantage of precision delivery of radiation enables dose escalation and sparing of normal tissues. The brain IMRT treatment planning consists of immobilization using an Aquaplast face mask. A CT simulation with intravenous contrast is done with the acquisition of 1.5- to 3-mm thickness images. The previously obtained MRI data is superimposed on the CT images using the fusion program. GTV, CTV, and PTV are defined as in a 3D-CRT plan. All normal structures are outlined in three dimensions. User-defined constraints, including maximum and minimum dose constraints both for the PTV and the normal structures, are prescribed. As in 3D-CRT, noncoplanar beams are used. The plan consists of an intensity profile that is created for each beam. This is translated to leaf motion of the dynamic multileaf collimator using the inverse planning algorithm (Fig. 21-8). We have seen that the use of IMRT decreases the dose to the critical structures like optic nerves and brainstem when compared with the 3-D conformal plan.³⁴ There is evidence that use of IMRT decreases the dose to the cochlea in children with medulloblastoma when used for the posterior fossa boost.

FIGURE 21-8 • Intensity-modulated radiation therapy treatment plan for a patient with high-grade glioma of right parietal region.

Special Issues With Radiation Management

Volume of Radiation

In the BTCG 8001 trial, patients were randomized to receive either 6000 cGy WBRT or 4300 cGy WBRT with 1720 cGy boost. No difference in survival or recurrence was noted.³⁵ In a Japanese trial, patients were randomized to 4000 cGy WBRT with 1800 cGy boost or 5600 cGy local field radiation alone.³⁶ Again, the 2-year survival rate was no different in this study (43% versus 39%), indicating that large-volume irradiation is unnecessary in high-grade gliomas.

Dose of Radiation

Radiation dose is another important consideration. In an MRC phase III trial, 444 patients were randomized to either 45 Gy in 20 fractions or 60 Gy in 30 fractions (1 : 2 randomization). A benefit of 60 Gy was seen in terms of 1-year survival (29% versus 39%) and the median survival (9 versus 11 months).³⁷ In the RTOG phase III trial, patients received either 60 Gy WBRT or 60 Gy WBRT with a 10 Gy boost (Fig. 21-9). There was no benefit with the boost in terms of the median survival (9.3 versus 8.2 months).³⁸ Researchers at the University of Michigan reported dose escalation to 80 Gy and 90 Gy with 3D-CRT and showed no difference in patterns of relapse or survival, indicating the absence of benefit with dose escalation beyond 60 Gy in malignant gliomas.³⁹

FIGURE 21-9 • Overall survival as a function of radiation dose.

Radiation Sensitizers

Multiple randomized trials conducted by the RTOG, the Brain Tumor Study Group (BTSG), and others have failed to show any survival benefit with misonidazole or other agents when compared with radiation therapy alone. Difluoromethylornithine, an inhibitor of polyamine synthesis, has not shown any benefit in a phase III clinical trial.⁴⁰ Efaproxiral (RSR13), a synthetic allosteric modifier of hemoglobin, showed a median survival of 12.3 months when used along with 60-Gy radiation therapy in a phase II study, and is being presently tested in a phase III trial.⁴¹

Two halogenated pyrimidine analogues, bromodeoxyuridine (BUdR) and iododeoxyuridine, have been tested and in randomized trials have failed to show any benefit. Northern California Oncology Group data had shown a median survival of 252 weeks with BUdR with concomitant radiation and chemotherapy in AA compared with 82 weeks without BUdR.⁴² However, an RTOG trial of 60 Gy in 30 fractions combined with PCV chemotherapy with or without BUdR in 189 patients of AA had to be prematurely closed because of lack of benefit (1-year survival 68% versus 82%) and higher toxicity.⁴³

Altered Fractionation

Again, multiple randomized trials conducted by RTOG, EORTC, and others have failed to show any survival benefit with hyperfractionation when compared with conventional radiation therapy alone.⁴⁴ Similarly, several trials using different accelerated fractionation regimens have been reported; none have shown a survival benefit over conventional irradiation. An EORTC trial involving 60 Gy given in either conventional fractionation or with three fractions of 2 Gy given in a single day, 4 hours apart, in 340 patients showed no difference in survival or any increased toxicity.⁴⁵ Doses of 64 and 70.4 Gy in 1.6-Gy, twice-daily fractions with BCNU tested in an RTOG trial showed no benefit.⁴⁶ Shortening the treatment time has again failed to show any improvement.⁴⁷ A Johns Hopkins trial of 30 Gy in 10 fractions followed by 2 weeks rest and then another 21 Gy in 7 fractions to reduced fields in 219 patients showed a survival similar to RTOG recursive partitioning analysis (RPA) groups I through VI patients.⁴⁸

Brachytherapy

An earlier BTSG trial that randomized malignant glioma patients to 60 Gy radiation with or without iodine-125 (I-125) seed implant to 50 Gy showed a survival benefit with the addition of brachytherapy (median survival of 13 versus 16 months).⁴⁹ Since the information was presented as an abstract only and never published, it remains difficult to establish that this approach clearly improves survival. However, a randomized trial done at Princess Margaret Hospital that randomized patients to 50-Gy external-beam radiation with or without temporary I-125 seed implant to 60 Gy did not show any survival benefit (median survival of 13.2 versus 13.8 months).⁵⁰

Stereotactic Radiosurgery

Several retrospective trials have indicated that there is a possible survival advantage with the addition of a stereotactic radiosurgery (SRS) boost to high-grade gliomas. The RTOG conducted a randomized trial of conventional radiotherapy to 60 Gy and BCNU alone or preceded by a radiosurgery boost to 15 to 24 Gy in patients with GBM measuring 4 cm or less.⁵¹ However, the results were disappointing. The median survival (14 versus 13.7 months), 2-year survival (22% versus 18%), and 3-year survival (16% versus 8%), were similar with or without the boost. There was no improvement in any RPA class. No increased toxicities were seen with the addition of the SRS boost. Failures were still predominantly local (>90%).