Investigators around the world are searching for the molecular biologic characteristics of gliomas in an effort to improve therapy. For example, work by the Cancer Genome Atlas Research Network ⁵ and others⁶ suggests that at least three molecular subclasses of GBM exist with potential therapeutic and prognostic implications. A set of nine genes (derived from a larger set of 38 genes) identifiable from paraffin blocks is also under investigation as a stratification factor in an ongoing phase III trial (Radiation Therapy Oncology Group [RTOG] 0825).⁷ Mouse modeling has also demonstrated the oncogenic importance of abnormalities in receptor signaling (e.g., epidermal growth factor receptor [EGFR] and platelet-derived growth factor receptor [PDGFR]), signal transduction cascades (e.g., RAS and AKT), and cell cycle regulation.⁸

In the early 1990s, it was recognized that deletion of the short arm of chromosome 1 (1p) and the long arm of chromosome 19 (19q) occurs in most oligodendrogliomas.⁹ Co-deletion of 1p and 19q is prognostic for longer survival,^10,11,12 although controversy continues as to whether this finding should alter therapy.¹³ It is now recognized that an unbalanced chromosomal translocation underlies 1p/19q co-deletion,¹⁴ but the specific genes involved and their mechanism of action remain elusive. More recently, mutations in the isocitrate dehydrogenase 1 and 2 (IDH1 and IDH2) genes, occurring most often in low-grade gliomas but also in a minority of WHO grade III and IV tumors,^15,16 are also prognostic for longer survival.^12,15,17,18 Methylation of the MGMT promoter (see the section on Chemotherapy) is emerging as both a predictive factor and a prognostic factor in the treatment of newly diagnosed GBMs.

It is not yet clear how to best incorporate molecular data into the treatment of individual patients. Detailed discussions of the biologic characteristics of glioma and their clinical relevance are beyond the scope of this chapter and can be found elsewhere.¹⁹

Pathology and Pathways of Spread

The WHO classification system ²⁰ is derived in part from the correlation between pathologic findings and survival rates observed by Bailey and Cushing²¹ and published in the early 1900s. In current parlance, low grade refers to WHO grade I to II tumors and high grade to WHO grade III to IV tumors (Fig. 25-1). However, the WHO grade I gliomas (e.g., juvenile pilocytic astrocytomas) are biologically different tumors from the others, infrequently occur in adults, and may be amenable to surgical cure. Anaplastic in the context of gliomas refers to WHO grade III tumors such as anaplastic astrocytomas, anaplastic oligodendrogliomas, and anaplastic oligoastrocytomas. GBM is a WHO grade IV astrocytoma, although the most recent WHO classification scheme does account for other rare subtypes, including GBM with oligodendroglial components.

Figure 25-1 Highly cellular astrocytic glioma with foci of microvascular proliferation (asterisk) and pseudopallisading necrosis (diamond) consistent with glioblastoma.

WHO grade II to IV gliomas are characterized by a tendency to directly infiltrate adjacent brain tissue. Lesions with direct access to the corpus callosum may extend across the midline and configure themselves in a classic “butterfly pattern” (Fig. 25-2). MRI underestimates the extent of invasive disease, and the very diffusely infiltrative nature of these tumors makes complete removal of all tumor cells impossible. This “misleading appearance of enucleability” was described over 80 years ago.²¹ Leptomeningeal spread occurs occasionally (Fig. 25-3). Hematogenous and lymphatic spread are exceedingly uncommon.²²

Figure 25-2 “Butterfly” glioblastoma (long arrow); a separate satellite lesion in the right parietal lobe (short arrow) is also apparent. The contrast enhancement with central necrosis is classic for a glioblastoma, which was histologically confirmed by biopsy.

Figure 25-3 T1 contrast-enhanced sagittal view from an MRI showing rare leptomeningeal spread from a glioblastoma surrounding the spinal cord.

Clinical Manifestations, Patient Evaluation, and Staging

No specific symptom constellation is pathognomonic of high-grade glioma. Typically, patients present with some combination of headaches, neurologic deficits, nausea, and vomiting, depending on tumor size and location. It is not unusual for patients to present with or subsequently develop seizures. Tissue for pathologic diagnosis can be obtained via stereotactic biopsy, open biopsy, or gross resection in the context of craniotomy. More complete resection improves diagnostic accuracy, provides additional tissue for molecular analyses, and is increasingly thought to improve overall survival (OS).

The typical imaging appearance of a GBM is a ring-enhancing or heterogeneously enhancing lesion. The differential diagnosis may include stroke, brain metastasis, primary central nervous system lymphoma, demyelination, and even infectious or other inflammatory diseases. If a brain metastasis is suspected, it is prudent to perform an appropriate extracranial evaluation to identify the primary malignant tumor. If there is a high index of suspicion for primary central nervous system lymphoma, such as in multifocal, periventricular, and/or homogenously enhancing lesions,²³ corticosteroids should be withheld preoperatively unless herniation is imminent because their use confounds the histologic diagnosis.

Nearly all GBMs demonstrate contrast-enhancement. However, up to 10% of anaplastic gliomas will not manifest such a pattern on imaging. Background uptake in the brain (an organ with an inherent avidity for glucose) significantly compromises the use of glucose-based positron emission tomography (PET) as a diagnostic tool for high-grade gliomas.

An algorithm for the evaluation and management of patients with GBM is shown in Figure 25-4.

Figure 25-4 Algorithm for diagnosis and treatment of glioblastoma. See text for details. BCNU, carmustine; MRI, magnetic resonance imaging; PE, physical examination; TMZ, temozolomide.

Primary Therapy

Prognostic and Predictive Factors

Historically, all high-grade gliomas were lumped together in most clinical trials, thereby confounding results due to maldistribution of patients with differing prognoses. In 1993, Curran and colleagues ²⁴ published a landmark paper describing a prognostic classification scheme based on clinical variables. Data from three RTOG trials that enrolled nearly 1600 high-grade glioma patients from 1974 to 1989 were used. This recursive partitioning analysis (RPA) methodology builds decision trees to model predictors by examining all possible “cut points” for all variables included in the model. Patients were segmented into six distinct groups with different survival outcomes. The key variables included patient age, performance status, histologic tumor type (i.e., anaplastic astrocytoma vs. GBM), mental status, symptom duration antecedent to diagnosis, extent of resection, neurologic function, and radiotherapy dose. Median survival ranged from 4.6 months for class VI patients to almost 5 years for class I patients, underscoring survival heterogeneity. The European Organization for Research and Treatment of Cancer (EORTC) demonstrated the prior RTOG RPA classification remained valid among patients in a more recent phase III trial.²⁵

One of the more controversial factors in the setting of high-grade gliomas has been the extent of surgical resection. Bailey and Cushing observed longer survival following resection in their 1926 publication,²¹ as did others in the 1960s.²⁶ Numerous series since then also support more complete resection as a prognostic factor.²⁷^–³¹ One of the largest involved over 400 patients at the M.D. Anderson Cancer Center and demonstrated improved median survival (13 months vs. 8.8 months; p < .0001) following at least 98% resection, as defined by postoperative MRI scans.³² One small randomized study of 30 patients older than 65 years demonstrated improved survival rates following resection versus biopsy alone.³⁰

Surgery

Patients can often undergo a craniotomy where the goal is the safe removal of the largest possible volume of tumor to establish a diagnosis and relieve mass effect. High-grade gliomas are not surgically curable, because of their extensive infiltration.

If biopsy rather than resection is pursued, choices include stereotactic options with CT or MRI guidance or open craniotomy and biopsy. Some have employed metabolic imaging such as magnetic resonance spectroscopy (MRS) to better select the biopsy sites most likely to contain the most aggressive portions of the tumor, but this is not widely used. Usually, stereotactic biopsy can be performed using either frame-based or frameless neuronavigation systems. The pathologist can review the frozen section to make an immediate preliminary diagnosis and confirm tissue adequacy.

When craniotomy is contemplated, meticulous planning is invested in the scalp incision and flap design. Special care must be taken to preserve the vascular supply of the scalp. The bony opening is devised to be sufficiently large to facilitate resection without needlessly exposing adjacent brain to the risk of injury. After opening the scalp, burr holes are placed and connected with a craniotome. The bone flap is removed and the dura mater is opened. As a rule, the bone flap is reattached at the end of the resection, although some surgeons perform craniectomies for posterior fossa and temporal lobe lesions.

Although tumors on the brain surface are immediately visible after exposure, subcortical lesions are harder to discern. Frameless image-guided neuronavigation systems are employed to localize subcortical tumors along with intraoperative ultrasound and MRI. Tumors that are situated near “eloquent” areas of cortex, such as those harboring motor and speech function, are mapped intraoperatively via electrical cortical stimulation to achieve maximal tumor debulking without operative morbidity. Resection of gliomas in the dominant hemisphere is often carried out under local anesthesia with supplementary sedation to allow speech mapping.

Occasionally, tumors may be removed en bloc via circumferential dissection, but more frequently, resection is effected in piecemeal fashion. A cavitational ultrasonic surgical aspirator (CUSA) facilitates removal of a firm, adherent, or calcified tumor.

Patients are routinely monitored in an intensive care unit following a craniotomy. The first MRI is obtained within 24 to 48 hours after surgery before postoperative changes set in. The extent of resection can be determined in this manner.

External Beam Irradiation

External beam irradiation (EBRT) has historically been the cornerstone of the therapeutic approach to GBM (and anaplastic astrocytoma) for the past half century, and its use in brain tumors was already described by the 1920s.³³ In the 1970s and early 1980s, categorical level I data became available from several studies,³⁴ including prospective phase III trials conducted by the Brain Tumor Study Group (BTSG)^35,³⁶ (Table 25-1).

TABLE 25-1 Positive Phase III Trials Evaluating the Role of Irradiation, Chemotherapy, or Chemoradiation in the Treatment of Malignant Gliomas

The radiotherapeutic approach to high-grade glioma has evolved. Initially, large opposed lateral fields were employed to cover the entire brain volume. In 1989, Shapiro and colleagues ³⁷ published data from Brain Tumor Cooperative Group trial 80-01, in which the randomization was altered during the trial to compare partial brain irradiation with whole-brain radiotherapy. No difference in OS or change in the patterns of failure was seen. Accordingly, whole-brain irradiation is generally not advocated, except perhaps in the scenario of a widespread intracranial process such as gliomatosis cerebri.³⁸

Several lines of evidence have influenced the trend to treat the gross tumor volume along with a margin of approximately 2 cm. In a classic paper published in 1980, Hochberg and Pruitt ³⁹ used CT scans to determine that nearly 90% of GBM recurrences occurred within 2 cm of the primary tumor site (although this may be changing with the use of bevacizumab).⁴⁰ Wallner and associates⁴¹ assessed the patterns of recurrence in 32 patients with unifocal malignant gliomas who were treated with primary surgery and postoperative irradiation. Nearly 80% of patients manifested recurrence or progression within 2 cm of the original tumor. Even when 80 Gy of partial brain irradiation was used in a prospective phase I trial,⁴² 90% of patients failed within the high-dose region.

It has been demonstrated on biopsy and autopsy studies that the abnormality detected on T2 or fluid-attenuated inversion-recovery (FLAIR) images harbors microscopic tumor extension. Accordingly, 45 to 50 Gy delivered, in 1.8- to 2-Gy fractions, to the T2/FLAIR abnormality seen on the image, followed by a boost to raise the total dose to 60 Gy based on the T1-enhancing abnormality, is generally prescribed. The MRI abnormalities, however, remain quite nonspecific in terms of histopathologic confirmation, and even when novel strategies such as MRS are used for radiotherapy planning, there can be overestimation or underestimation of the target.⁴³

Rationale for Current Total Irradiation Dose

A pooled analysis of three successive randomized trials conducted by the Brain Tumor Study Group (BTSG 66-01, 69-01, 72-01, respectively) generated data to support doses in excess of 50 Gy.⁴⁴ A stepwise improvement in survival was observed with doses ranging from less than 45 Gy to 60 Gy, consistent with dose response. A comparison of 70 Gy versus 60 Gy demonstrated no survival or local control advantage for the 70-Gy dose.^45,⁴⁶ These results established 60 Gy as the standard of care.

Dose escalation has remained an important investigational option because there is still a pattern of failure characterized by local progression or recurrence. There are now multiple techniques for dose escalation, including three-dimensional conformal irradiation, radiosurgery, and brachytherapy, but these have not yielded higher rates of disease control or survival. Increasing the radiotherapy dose or reirradiation with radioprotective or antiangiogenic agents may be a useful strategy in the future.⁴⁷

Altered Fractionation

The RTOG has systematically and rigorously studied hyperfractionation for high-grade gliomas. In trial 83-02, patients were randomized to one of four dose arms (64.8, 72, 76.8, or 81.6 Gy) using twice-daily fractions of 1.2 Gy. Initial results suggested the superiority of 72 Gy,⁴⁸ but a subsequent phase III trial demonstrated no improvement.⁴⁹

Prados and colleagues ⁵⁰ used an elegant randomization to assess not only a hyperfractionation schedule of 1.6 Gy twice daily to a total dose of 70.4 Gy, but also to determine the activity of difluoromethylornithine (DFMO), a compound that inhibits sublethal and potentially lethal damage repair. Unfortunately, neither intervention improved survival.

Stereotactic Irradiation

Two provocative small-scale experiences prompted the design of phase III trials to formally evaluate radiosurgery for high-grade gliomas. Loeffler and colleagues ⁵¹ reported on 37 patients treated with 59.4 Gy fractionated radiotherapy followed by a radiosurgical boost to a median dose of 12 Gy. After a median follow-up period of 19 months, a 76% survival rate was reported. Sarkaria and associates⁵² described 115 high-grade glioma patients who received conformal radiation therapy and a stereotactic boost. The median survival time was 96 weeks. These results called into question whether they represented a benefit from radiosurgery or simply selection bias.

RTOG 93-05 compared conformal irradiation plus carmustine with or without a radiosurgical boost for newly diagnosed GBM. No differences were observed in terms of OS (median, approximately 13 months in each arm) or quality of life.⁵³

Radiosurgery may still have a role in the treatment of recurrent disease, particularly if a focal region of recurrence can be defined.⁴⁷

Brachytherapy

Brachytherapy refers to the use of implanted radioactive material at the site of the tumor. Like radiosurgery, it offers a mechanism for focal dose escalation. Both permanent and temporary radioactive implants have been placed in the brain.

Gutin and associates ⁵⁴ suggested a potential survival benefit in a phase II trial. These findings were not reproduced in subsequent randomized trials.^55,⁵⁶

Interest in this modality was rekindled when the GliaSite radiation therapy system (Proxima Therapeutics, Alpharetta, Georgia) received approval by the Food and Drug Administration (FDA) in 2001. This intracavitary device is implanted at the time of tumor debulking, and a solution of iodine-125 (¹²⁵I) is injected into an expandable closed-catheter balloon. A retrospective study suggested reasonable safety and promising efficacy.⁵⁷ A phase I study was conducted.⁵⁸ However, the implant induces changes in imaging that complicate determination of disease progression.⁵⁹

Chemotherapy: Concurrent with EBRT, Maintenance Chemotherapy, Other

Early randomized trials of chemotherapy were individually negative, but meta-analysis of these trials showed that 15% to 20% of patients treated with radiation therapy (RT) and nitrosoureas survived at least 18 months versus 5% treated with radiotherapy alone.^35,^36,^60,⁶¹ Nitrosoureas, especially carmustine, were the most commonly used drugs, although procarbazine was also used extensively.⁶² The combination of procarbazine, lomustine (CCNU), and vincristine, also called PCV,⁶³ had no clear benefit (yet much greater toxicity) versus carmustine for anaplastic astrocytoma,⁶⁴ and this regimen has been largely abandoned.

Intratumoral delivery of chemotherapy for residual postoperative disease is most commonly in the form of carmustine-eluting (Gliadel) wafers. Patients undergoing wafer implantation during surgery for recurrent GBM survived approximately 2 months longer than patients without wafers in one study (p = .02).⁶⁵ Treatment of newly diagnosed disease also yielded a 2-month prolongation of average survival.^66,⁶⁷ However, this was not statistically significant when the analysis was restricted to patients with GBM histology. Of note, wafer delivery of carmustine versus systemic administration has not been compared for safety or efficacy. Gliadel does, however, carry an FDA label for implantation during resection of recurrent GBM and newly diagnosed malignant glioma. Attempts to treat residual visible or microscopic disease with other local chemotherapies delivered through implanted catheters and using convection-enhanced migration of drug have generally failed.⁶⁸

Currently, the most widely used chemotherapeutic agent is temozolomide. Whether it is more effective than nitrosoureas has not been investigated, but it is unquestionably better tolerated with significant myelosuppression in less than 20% of patients.^69,⁷⁰ Temozolomide was first approved for recurrent anaplastic astrocytomas following a phase II study.⁷¹ A randomized study also demonstrated superior efficacy to procarbazine in recurrent GBM.⁷²

Temozolomide for newly diagnosed GBM has been studied both when given before RT ⁷³ and when combined with RT in various dosing schedules.^74,⁷⁵ Its role for newly diagnosed GBM was established by Stupp and associates^69,⁷⁰ on the basis of the EORTC 26981/22981 and NCIC CE.3 trial. In this phase III multicenter study, 573 newly diagnosed GBM patients received RT alone or RT with concurrent temozolomide followed by six adjuvant cycles of temozolomide. The patients who received the combined-modality regimen had significantly longer OS and progression-free survival (PFS) rates without significantly more toxicity (see Table 25-1). The 5-year OS was 10% among those receiving temozolomide versus 2% among those receiving radiotherapy alone (p <.0001).⁶⁹ Patients in the most favorable RPA class had a 5-year OS rate of 28% following combined therapy.⁶⁹ In a companion paper, Hegi and colleagues⁷⁶ reported that methylation of the promoter for the O₆-methylguanine DNA methyltransferase (MGMT) gene, which encodes the DNA repair enzyme O₆-alkylguanine DNA alkyltransferase (AGT or AGAT, but now commonly also referred to as MGMT), was correlated with prolonged survival.

MGMT repairs DNA damage induced by temozolomide, and methylation of the MGMT promoter silences expression of the protein, thereby accentuating the antineoplastic effects of temozolomide. However, the mechanism by which MGMT promoter methylation leads to an improved outcome is more complex. For example, patients with tumors harboring methylated MGMT survived longer following treatment with RT alone than patients who did not have tumors harboring methylated MGMT treated identically.⁶⁹ Others reported similar findings in GBM⁷⁷ and other malignant gliomas.⁷⁸ Moreover, MGMT protein expression by immunostaining does not predict outcome.⁷⁹ It remains unclear whether MGMT analysis should alter treatment, although several alternate dosing schedules of temozolomide are under investigation^80,⁸¹ as a means to overcome this and other potential resistance mechanisms.

Anaplastic Gliomas

Anaplastic astrocytomas, anaplastic oligodendrogliomas, and anaplastic oligoastrocytomas represent the most common WHO grade III tumors.^1,²⁰ In anaplastic gliomas, resection appears to improve survival relative to biopsy, as it does for GBM.

It is generally accepted that radiotherapy should be administered postoperatively for astrocytomas. In a German study (NOA-04), survival rates were equivalent whether chemotherapy or radiotherapy was used first among patients with anaplastic astrocytomas, oligodendrogliomas, and mixed tumors.¹² However, time to progression following RT was longer than after chemotherapy, and initial radiation therapy achieved more complete and partial responses than initial chemotherapy, suggesting the superiority of radiotherapy.⁸²

Combs and associates ⁸³ reviewed the outcome of 191 patients with grade III astrocytic tumors treated at the University of Heidelberg with either RT alone or RT in combination with temozolomide during a 20-year period (from 1988 to 2007). In this retrospective study, no significant advantage in rates of OS or PFS could be attributed to the combination. RTOG trial 9813 randomized patients with anaplastic astrocytomas (or oligoastrocytomas) to radiotherapy with concurrent nitrosourea (carmustine or lomustine) or with temozolomide, and results are pending.

The ongoing CATNON (Concurrent vs. Adjuvant Temozolomide for NON 1p19q co-deleted anaplastic gliomas, also called EORTC 26053-22054) Intergroup trial randomizes patients to receive postoperative RT alone, concurrent temozolomide and RT without adjuvant temozolomide, RT (without concurrent temozolomide) followed by adjuvant temozolomide, or RT with both concurrent and adjuvant temozolomide (Fig. 25-5). Although technically this study allows any WHO grade III glioma without 1p/19q co-deletion, the majority of the tumors will be anaplastic astrocytomas.⁸⁴

Figure 25-5 Schema from the CATNON study (Intergroup phase III, EORTC 26-53, RTOG 0834). AG, anaplastic glioma; R, randomization; RT, radiotherapy; TMZ, temozolomide.

The role of radiotherapy for newly diagnosed anaplastic oligodendroglioma is most controversial because of the tumor’s reported sensitivity to chemotherapy,⁸⁵ especially tumors with 1p/19q co-deletion.⁸⁶ In spite of this, two separate randomized studies that compared radiation therapy with radiation therapy plus chemotherapy (RTOG 94-02¹⁰ and EORTC 26951¹¹) failed to demonstrate a survival advantage from the addition of chemotherapy. PCV chemotherapy prolonged PFS, and those who failed irradiation were frequently treated with PCV, possibly confounding the results. Both trials provided validation of the prognostic value of the allelic loss of heterozygosity of the 1p and 19q chromosomes.

Neither of these trials used chemotherapy alone, which is recommended by 42% of clinicians (typically with temozolomide) for co-deleted tumors.¹³ The ongoing CODEL (for 1p/19q CO-DELeted tumors) trial randomizes patients to RT alone, RT with concurrent and adjuvant temozolomide, or temozolomide alone to prospectively address this issue (Fig. 25-6).

Figure 25-6 Schema from the CODEL trial (NCCTG NO0557/RTOG 0670).

The CATNON and CODEL study designs are premised on the belief that the 1p/19q chromosomal status may be more important than the histologic subtype for WHO grade III lesions. However, scientific advances in the interim may outpace the questions posed by these trials as it becomes apparent that MGMT promoter methylation ⁷⁸ and IDH gene mutation^12,¹⁷ are also powerful markers of outcome in anaplastic gliomas.

Special Topics

Toxicities of Radiation Therapy

Acute and late effects of irradiation are discussed in a subsequent section on Irradiation Techniques and Toxicities.41,83–86,87,88 This includes a discussion of brain necrosis and the risk of normal tissue damage based on the most sensitive 5% of the population.

Elderly Patients with GBM

Over 44% of GBMs occur in patients over 65 years of age,¹ but prospective trials often exclude elderly patients.⁶⁹ Because these patients often have a worse outcome than younger patients, there has been some uncertainty with regard to the aggressiveness of treatment.

A Finnish randomized study of 30 patients over age 65 years with imaging consistent with malignant gliomas demonstrated longer median survival following craniotomy than following biopsy (5.7 months vs. 2.1 months; p = .035).³⁰ The Association of French-Speaking Neuro-Oncologists⁸⁷ conducted a randomized trial comparing RT alone (50 Gy in conventional fractionation) with supportive care among patients over 70 years of age. The trial was discontinued after 85 patients were registered when it became apparent that radiotherapy was associated with a statistically advantageous outcome (median survival time, 29 weeks vs. 17 weeks; p = .002).

Efforts to abbreviate the duration of RT have been conducted. A randomized study in patients over age 60 years demonstrated that 40 Gy in 15 fractions was not inferior to 60 Gy in 30 fractions.⁸⁸ A phase II study of 32 patients over age 70 years with newly diagnosed GBM demonstrated a response rate of 31% to temozolomide, with a median PFS time of 5 months and a median survival time of 6.4 months, results that were comparable to those of radiation therapy alone.^89,⁹⁰

The EORTC is currently conducting a trial to assess the benefit of combined temozolomide and RT in elderly patients with GBM. Intravenous temozolomide recently became available as an option for elderly or otherwise impaired patients who cannot swallow capsules or comply with an oral regimen.⁹¹

Pseudoprogression

Pseudoprogression confounds interpretation of imaging performed in the first several months following completion of radiotherapy. Descriptions of “pseudoprogression” appeared as early as 1979, when Hoffman and colleagues ⁹² described patients treated with radiotherapy and carmustine. Among patients thought to have experienced disease progression immediately following irradiation, nearly half were shown to have improvement or at least stabilization on subsequent brain imaging.

A report in 2004 described that approximately one third of patients with gliomas stabilized or improved with no change in management.⁹³ Chamberlain and associates⁹⁴ reported histologically proven treatment injury rather than disease progression in approximately 50% of patients with symptomatic resectable lesions felt to represent worsening disease following concurrent radiotherapy and temozolomide. The incidence of pseudoprogression has been reported to be as high as 75%.^81,^95,⁹⁶ Pseudoprogression rather than “true” progression may also correlate with MGMT promoter methylation,⁹⁵ although this has not yet been prospectively validated.

Multiple imaging techniques have been explored to delineate radiographic pseudoprogression from true progression.^95,⁹⁷^–¹⁰¹ However, at this time, histologic analysis is the only validated method of distinguishing the two diagnoses.¹⁰²

It is presumed that the incidence of pseudoprogression is higher following concurrent radiotherapy and temozolomide than that following radiotherapy alone, although supportive evidence is equivocal.¹⁰³ One approach to dealing with this issue involves using the MRI study done following radiation therapy as a new baseline rather as a means of designating disease progression, unless there is surgical documentation of recurrent disease or clear worsening outside of the radiation therapy port. This remains an area of active study.¹⁰⁴

Recurrent Disease

In brief, efforts to treat recurrent disease with single agents with single molecular targets have generally been unsuccessful, at least in part because of the molecular complexity of the disease.¹⁰⁵ Some studies suggest the key may lie in identifying the patients most likely to respond based on molecular profiling of tumor tissue.¹⁰⁶

An exception is the use of agents that target vascular endothelial growth factor (VEGF) or VEGF receptors (VEGFRs). Bevacizumab, a monoclonal antibody against VEGF, is the most widely studied of these antiangiogenic strategies.

The “BRAIN” trial was a noncomparative phase II study for recurrent GBM that randomized 167 patients to either bevacizumab alone or bevacizumab combined with irinotecan,¹⁰⁷ the latter regimen adopted from activity of that combination in colon cancer.¹⁰⁸ The 6-month progression-free survival rate was 43% with bevacizumab and 50% with bevacizumab plus irinotecan.¹⁰⁷ These results are far superior to those of prior available therapies.

To date, however, there are no data clearly demonstrating a survival advantage for the use of bevacizumab, and major potential toxicities include thromboembolic disease and hypertension. In addition, treatment following progression on bevacizumab may have little to no efficacy,^40,¹⁰⁹ and the timing of bevacizumab for either first or later recurrence remains controversial. The potential toxicity, as well as the fear of inducing a more invasive tumor phenotype, also tempers enthusiasm for use of this agent in newly diagnosed disease until randomized data become available. RTOG 0825 and the AVAglio trial, as well as antiangiogenic approaches with the anti-integrin cilengitide (CENTRIC) and other agents such as the VEGFR tyrosine kinase inhibitor cedirinib, are currently addressing this issue (Table 25-2).

TABLE 25-2 Major Ongoing Multicenter/International Phase III Trials

Trial Name	Disease	Notes
CATNON	Anaplastic gliomas without 1p/19q co-deletion	Efficacy of radiation therapy alone or with concurrent and/or adjuvant temozolomide
CODEL	Anaplastic gliomas without 1p/19q co-deletion	Efficacy of radiation therapy with or without concurrent and adjuvant temozolomide, or temozolomide alone
RTOG 0525	GBM	Efficacy of intensive adjuvant temozolomide dosing following radiation therapy
RTOG 0825	GBM	Efficacy of bevacizumab in newly diagnosed disease
AVAglio	GBM	Efficacy of bevacizumab in newly diagnosed disease
CENTRIC	GBM	Efficacy of cilengitide for newly diagnosed disease
REGAL	GBM	Efficacy of cedirinib in recurrent disease

CATNON, Concurrent vs. Adjuvant Temozolomide for NON 1p/19q co-deleted anaplastic gliomas; CODEL, 1p/19q CO-DELeted tumors; GBM, glioblastoma; RTOG, Radiation Therapy Oncology Group.

Reirradiation

Focal radiotherapy approaches are often employed with limited volume recurrences. In a retrospective analysis of 95 patients with recurrent gliomas treated with the GliaSite device, the median survival time was 36 weeks.^110,¹¹¹ However, whether this is a function of true benefit or patient selection has not been determined.

Fractionated radiotherapy to treat larger-volume recurrent disease has also been employed. Although there has been speculation from animal studies that neural tissue will recover from previous irradiation to a large extent once some time has elapsed (e.g., 1 to 3 years),¹¹² no firm data quantify the degree to which one can assume that a “dose discount” exists. It is most likely that the damage from reirradiation is underestimated because the majority of patients do not live long enough to express such damage. A small study showed good short-term tolerance to intensity-modulated radiation therapy (IMRT) delivered in six daily fractions of 5 Gy.¹¹³ A recent single-arm trial from the Memorial Sloan-Kettering Cancer Center demonstrated reasonable safety and efficacy of combined bevacizumab and reirradiation using IMRT to small recurrent malignant gliomas.⁴⁷ No radionecrosis was observed and survival appeared to be prolonged relative to historic controls, suggesting that bevacizumab may not only treat radionecrosis¹¹⁴ but may protect against it. Another concept was recently reported by investigators from Jefferson Medical College¹¹⁵ who treated 147 patients with recurrent high-grade glioma to 35 Gy in 3.5-Gy fractions with stereotactic radiation therapy. This hypofractionated stereotactic radiation therapy (HSRT) approach was associated with a median survival time in excess of 10 months.

Irradiation Techniques and Toxicities

Whole-brain radiation therapy has been replaced with partial brain techniques by consensus. Although the dose computation component of treatment planning is still based on CT imaging, effective image registration with MRI has made this the modality of choice for contouring. The notion of a dedicated MRI simulator has also been proposed as a valuable adjunct in the radiotherapeutic management of high-grade gliomas.¹¹⁶ However, treatment plans based only on MRI are not able to take into account tissue density variations, an inability that may lead to slightly inaccurate dose calculations.

Patients are usually simulated after the surgeon has removed the craniotomy staples (generally, 10 to 14 days after the operation). An immobilization mask is fashioned to reduce motion during and between fractions. The planning CT scan is extended to encompass the head and neck region to allow sufficient anatomic areas for proper image fusion and generation of high-quality digitally reconstructed radiographs (DRRs) and to permit the introduction of noncoplanar beams; ideally, the slice thickness should match that of the fusion postoperative MRI.

For high-grade gliomas, especially GBM, the T1 contrast-enhanced sequences are used to define the gross tumor volume (GTV) and the T2 or FLAIR sequences plus a margin define the microscopic disease extent, or clinical target volume (CTV), which reflects the bulk of microscopic infiltration. To arrive at a planning target volume (PTV), both organ motion and setup error must be taken into account. Organ motion in the brain is quite minimal during therapy (e.g., <1 mm). The PTV may be further modified to exclude normal tissue in areas where gliomas are unlikely to infiltrate. Bokstein and colleagues ¹¹⁷ recently demonstrated that when anatomic barriers such as the temporal bone can serve as a border to impede tumor spread, failure is likely to be seen in less than 5% of cases even when the customary 2-cm margins are not added to the abnormalities seen on MRI (Fig. 25-7).

Figure 25-7 Lesion that could be treated without a full complement of margins because of anatomic barrier of the temporal bone.

In general, there are two major schools of thought (with numerous institutional variations based on these) that provide guidance for the prescription of the radiation regimen. The RTOG approach is a biphasic technique that includes an initial PTV (PTV 1) followed by a second PTV (PTV 2) that represents the cone down. In the lexicon of the RTOG, the PTV 1 includes the T2 or FLAIR CTV with a margin and is treated to 46 Gy in 2-Gy fractions. The PTV 2 includes the T1-enhancing GTV with a margin and is treated to an additional 14 Gy. The EORTC recommends a single-phase technique using one treatment volume throughout the course of therapy. Table 25-3 shows the partial brain volumes advocated by several cooperative groups for the successive phases of partial brain irradiation.

TABLE 25-3 Radiotherapy Volume Used in Current Clinical Trials

RT Dose	Block Edge Dosimetry Margin*	Source for RT Dose
46 Gy	T2 + 2 cm	RTOG
14 Gy	T1 + 2.5 cm	RTOG
60 Gy†	T1 + 2-3 cm	EORTC
50 Gy	T2 + 2 cm	NCCTG
10 Gy	T1 + 2 cm	NCCTG

EORTC, European Organization for Research and Treatment of Cancer; Gy, Gray; NCCTG, North Central Cancer Treatment Group; RT, irradiation dose; RTOG, Radiation Therapy Oncology Group; 3D, three-dimensional.

* Both RTOG and EORTC prescribe dose to block edge (geometric) margin, whereas NCCTG prescribes dose to dosimetry (3D) margin.

† CENTRIC studies: either RTOG or EORTC volumes and RT doses are used, at individual physician discretion.

With the advent of functional imaging tools (e.g., functional MRI [fMRI]) it may be possible to specifically modify irradiation doses to functional brain areas. Figure 25-8 displays a treatment plan wherein the region governing motor control (e.g., finger tapping) is delineated to enable an accounting for dose deposition. In this case, this region in the right hemisphere (i.e., governing tapping by the left upper extremity) is included in the high-dose region but the contralateral side is well spared.

Figure 25-8 Treatment planning based on a functional MRI. Note delineation of the clinical target volume (CTV) as well as left and right somatomotor areas (SMA) controlling hand movements of the right and left upper extremities, respectively.

Table 25-4 summarizes the tolerance of various organs at risk according to the new QUANTEC (Quantitative Analysis of Normal Tissue Effects in the Clinic) guidelines. Lawrence and colleagues¹¹⁸ postulated that the original estimates of Emami and colleagues,¹¹⁹ suggesting a 5% risk of chronic brain damage at 5 years when one-third of the brain is irradiated to 60 Gy, were overly conservative. Instead, they hypothesized (Fig. 25-9) that the dose correlated with a 5% risk of damage at 5 years following conventionally fractionated irradiation to the partial brain is 72 Gy.

TABLE 25-4 Selected Organs at Risk for Treatment Planning of Malignant Glioma

Organ	Dose Limit	Comments
Brain parenchyma¹¹⁸	72 Gy	Brain appears to be more sensitive to fraction size >2 Gy and to twice-daily radiation therapy
Optic apparatus¹³²	60 Gy	12 Gy for single-fraction radiosurgery
Brainstem¹³³	54 Gy	59 Gy if small volumes (i.e., 1-10 mL)
Parotid gland¹³⁴	20 Gy	Severe xerostomia vis-à-vis sparing of one parotid gland
Parotid gland¹³⁴	25 Gy	Severe xerostomia vis-à-vis sparing of both parotid glands

Figure 25-9 QUANTEC data for central nervous system tolerance based on the study in Lawrence and colleagues. Relationship between biologically effective dose (BED) and radiation necrosis after fractionated radiotherapy. The Lawrence figure was done with a nonlinear least-squares algorithm using MatLab software (MathWorks, Natick, Mass.). The nonlinear function chosen was the probit model (similar functional form to the Lyman model). Dotted lines represent 95% confidence levels; each dot represents data from a specific study (Lawrence Table 2); n = patient numbers as shown. A, Fraction size less than 2.5 Gy. B, Fraction size 2.5 Gy or larger (data too scattered to allow plotting of “best-fit” line). C, Twice-daily radiotherapy.

Redrawn with permission from Lawrence YR, Li XA, Naqa I, et al: Radiation dose-volume effects in the brain, Int J Radiat Oncol Biol Phys 76:S20-S27, 2010.

There may sometimes be a tendency to overlook structures that, if damaged, would have led to noncatastrophic sequelae. For instance, although it is true that radiation-induced cataracts are easily repairable,¹²⁰ avoidance of entrance and exit dose to the eye may be a relatively simple means of preventing not only cataracts but also conjunctivitis and a dry eye. Similarly, when one contours the ear canals, there is now a greater awareness of the risks of developing otitis externa as well as otitis media.

There are reports of cytopenias arising from cranial irradiation even among patients who have not received chemotherapy.¹²¹ The hypothesized mechanism of this consequence is either irradiation of circulating blood (e.g., within the cavernous sinus), injury to marrow in the cranium, or the use of vertex beams that exit the spinal axis. The latter mechanism has also been invoked to explain radiation-induced fatigue by some investigators.¹²²

For tumors located in the temporal lobes, the exit dose to the parotid gland may bring about xerostomia. Tumors in such locations may be best treated with IMRT because one of the success stories of the application of IMRT to the head and neck region is indeed the elimination of this bothersome side effect.

Overall, the aim is to achieve the treatment plan that most closely approximates the defined volumes and thereby produces the most conformal plans. Relatively low energy beams (e.g., 6 MV) are typically employed. In contrast to the irradiation of varying organs containing primary tumors, it is impossible to recommend a single idealized beam arrangement for high-grade gliomas.

Miscellaneous Technique Issues

Despite the absence of level I evidence, IMRT is increasingly used, sometimes to escalate the dose and at other times to spare surrounding tissues or to explore the concomitant boost concept. However, clinical data supporting the value of IMRT in high-grade gliomas are essentially nonexistent.

Guidelines for tumor delineation and dose selection for WHO grade III gliomas are less well developed. In general, the FLAIR and T2 compartments are regarded as containing microscopic disease and therefore constitute the CTV. The typical dose is 59.4 or 60 Gy in 1.8- to 2-Gy fractions.

Toxicities of Radiation Therapy

Acute radiation morbidity includes fatigue, erythema, alopecia, headache, and, rarely, nausea with or without vomiting; these are generally not severe and are usually self-limiting.⁴¹ Some have cautioned that the combination of cranial radiotherapy and phenytoin as well as other anticonvulsants could give rise to the Stevens-Johnson syndrome,^123,¹²⁴ but this dermatologic emergency is an exceedingly rare event.

Late effects of radiation (e.g., somnolence and, especially, cognitive impairments) are more worrisome and may become manifest many years later.¹²⁵ The impact of partial brain irradiation on neurocognitive decline continues to be a hotly debated topic. The confounding factor is always the extent to which there is baseline cognitive impairment or decreased mentation secondary to uncontrolled tumor.

Brain necrosis is a serious and uncommon late toxicity, and recently bevacizumab has been explored as a treatment.¹¹⁴ In a small trial, all patients showed improvement of MRI abnormalities as well as a reduction in corticosteroid requirements.¹²⁶

We currently estimate the risk of normal tissue damage based on the most sensitive 5% of the population. Accordingly, we bias our recommendations in a manner that is not germane to most individuals, and preliminary efforts at predicting the likelihood of toxicities based on risk are ongoing.127,128

Treatment Algorithm, Controversies, Challenges, and Future Possibilities

Treatment Algorithm (Fig. 25-4)

For adults up to age 70 years with newly diagnosed GBM, maximal surgical resection (with or without Gliadel wafer implantation) followed by concurrent EBRT (approximately 60 Gy in 30 fractions) with temozolomide and followed by at least 6 months of adjuvant temozolomide is the standard of care. For suspected progression that occurs within 3 months of completing radiotherapy, the possibility of pseudoprogression should be strongly considered.¹²⁹ Relapsed disease may be treated with re-resection, second-line chemotherapy, or experimental therapy in a clinical trial. Reirradiation is also a possibility but is used infrequently. Bevacizumab is approved for relapsed disease.¹⁰⁷

For newly diagnosed anaplastic gliomas, radiotherapy is typically administered as part of the initial therapy. Some advocate chemotherapy first, especially in oligoendrogliomas.^12,¹³

Controversies

Treatment for elderly patients (variably defined as older than 60, 65, or 70 years of age, depending on the study) remains controversial. Patients with a good performance status and few comorbidities are often treated according to the same algorithm used for younger patients. However, abbreviated radiotherapy courses appear to be non-inferior ⁸⁸ and are often employed. Whether chemoradiation using a rapid course of radiotherapy is superior to radiotherapy alone is under investigation.

Alternate temozolomide dosing schedules in the adjuvant setting (following radiotherapy) are also being studied to determine their value, with particular attention to overcoming MGMT-mediated resistance (e.g., RTOG 0525). Bevacizumab is now an accepted treatment for relapsed GBM. Ongoing trials will address the safety and efficacy of treatment with bevacizumab or other antiangiogenic agents with radiotherapy and temozolomide for newly diagnosed disease.

Treatment of anaplastic gliomas is an area of extreme controversy, with astrocytomas often treated as GBMs and a more variable approach used for oligodendrogliomas. Ongoing trials will attempt to define a standard algorithm based on the 1p/19q deletion status.

Challenges

Local control of high-grade gliomas remains a vexing problem for radiation oncologists. In addition, the increased recognition of pseudoprogression as an entity and the possibility of pseudoresponse to antiangiogenic therapy such as bevacizumab ¹³⁰ complicate matters further. A revised set of consensus criteria were recently developed in part to address some of these issues.^104,129 Certainly, correlative clinical follow-up to provide the proper perspective on the health status of the patient will never be abandoned, because imaging will always represent an imperfect surrogate for survival. Although new developments in diagnostic imaging continue to hold out promise for resolution of the diagnostic dilemmas faced by the neuro-oncology team, to date even the most sophisticated imaging studies (e.g., FDG PET, MRS) have not provided a consistently reliable solution to these and other vexing problems.

Future Possibilities

For most tumors that have demonstrated radiosensitivity but defy control, the reflex recommendation of the radiation oncologist is usually dose escalation. Ostensibly, dose escalation as a solution to the problem of local failure of high-grade gliomas has been put to rest by virtue of trials discussed earlier. However, it is conceivable that the notion of dose escalation now needs to be revisited in the backdrop of potential radioprotection offered by bevacizumab,⁴⁷ both for newly diagnosed and recurrent disease.

Many new strategies to improve outcome in high-grade gliomas have been proposed. The RTOG awaits analysis of trial 0525, which randomized newly diagnosed GBM patients between the conventional Stupp regimen and the same regimen with dose-intensive temozolomide. The brisk accrual of large numbers of glioma patients to recent studies such as RTOG 0525 and 0825 is a testimony to the renewed interest in studying high-grade gliomas. The success of antiangiogenic therapy has also reinvigorated study of this disease.¹⁰⁷

The 5-year survival rate from the EORTC-NCIC study of patients with GBM treated with radiotherapy and temozolomide was approximately 10%, and for patients with favorable prognostic factors it approached 30%.⁶⁹ Moreover, a patient who lived more than 20 years following the diagnosis of GBM was described, perhaps the longest documented survivor.¹³¹ He had been treated with surgery and partial brain irradiation (59 Gy of 6-MV photons in conventional fractionation delivered via the shrinking-field technique). The authors speculated that the outcome may have stemmed from the fact that he had a favorable molecular profile (e.g., methylated MGMT promoter, PTEN positive, and TP53 positive, which the authors termed “triple positive,” similar to the nomenclature of breast cancer). Whether this explained the long survival time is unclear. More importantly, these observations prove that one may strive to create and sustain hope for patients diagnosed with high-grade glioma.

Critical References

6 Phillips HS, Kharbanda S, Chen R, et al. Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell. 2006;9:157-173.

7 Colman H, Zhang L, Sulman EP, et al. A multigene predictor of outcome in glioblastoma. Neuro Oncol. 2010;12:49-57.

9 Reifenberger J, Reifenberger G, Liu L, et al. Molecular genetic analysis of oligodendroglial tumors shows preferential allelic deletions on 19q and 1p. Am J Pathol. 1994;145:1175-1190.

10 Cairncross G, Berkey B, Shaw E, et al. Phase III trial of chemotherapy plus radiotherapy compared with radiotherapy alone for pure and mixed anaplastic oligodendroglioma. Intergroup Radiation Therapy Oncology Group Trial 9402. J Clin Oncol. 2006;24:2707-2714.

11 van den Bent MJ, Carpentier AF, Brandes AA, et al. Adjuvant procarbazine, lomustine, and vincristine improves progression-free survival but not overall survival in newly diagnosed anaplastic oligodendrogliomas and oligoastrocytomas. A randomized European Organisation for Research and Treatment of Cancer phase III trial. J Clin Oncol. 2006;24:2715-2722.

14 Jenkins RB, Blair H, Ballman KV, et al. A t(1;19)(q10;p10) mediates the combined deletions of 1p and 19q and predicts a better prognosis of patients with oligodendroglioma. Cancer Res. 2006;66:9852-9861.

15 Yan H, Parsons DW, Jin G, et al. IDH1 and IDH2 mutations in gliomas. N Engl J Med. 2009;360:765-773.

24 Curran WJJr, Scott CB, Horton J, et al. Recursive partitioning analysis of prognostic factors in three Radiation Therapy Oncology Group malignant glioma trials. J Natl Cancer Inst. 1993;85:704-710.

34 Shapiro WR, Young DF. Treatment of malignant glioma. A controlled study of chemotherapy and irradiation. Arch Neurol. 33, 1976. 494–450

35 Walker MD, Alexander EJr, 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.

36 Walker MD, Green SB, Byar DP, et al. Randomized comparisons of radiotherapy and nitrosoureas for the treatment of malignant glioma after surgery. N Engl J Med. 1980;303:1323-1329.