Key points

Introduction

Malignancies arising from the central nervous system (CNS) are rare, representing 1.4% of cancer cases, with approximately 24,000 new cases estimated in 2016 leading to approximately 16,000 attributable deaths. Brain metastases, in contrast, are perhaps the most common neurologic complication of cancer, with an estimated incidence rate of up to 200,000 per year in the United States. Radiotherapy, as part of combined modality therapy, continues to evolve with the advancement of stereotactic radiosurgery (SRS) indications; addition of new technologies, such as alternating electric field therapy; and mounting advances in the complex biology of these entities.

Introduction

Malignancies arising from the central nervous system (CNS) are rare, representing 1.4% of cancer cases, with approximately 24,000 new cases estimated in 2016 leading to approximately 16,000 attributable deaths. Brain metastases, in contrast, are perhaps the most common neurologic complication of cancer, with an estimated incidence rate of up to 200,000 per year in the United States. Radiotherapy, as part of combined modality therapy, continues to evolve with the advancement of stereotactic radiosurgery (SRS) indications; addition of new technologies, such as alternating electric field therapy; and mounting advances in the complex biology of these entities.

Brain metastases

Decision-making with respect to the treatment of brain metastases is a function of many factors, including patient age, performance status, comorbid conditions, symptoms, extracranial disease extent, and histopathology of the primary tumor. Treatment options include resection, whole-brain radiotherapy (WBRT), SRS, or some combination of these modalities. Work has been done to stratify patients into various prognostic groups, identifying cut-points of three factors prognostic for overall survival: (1) Karnofsky performance status (KPS), (2) age, and (3) presence of extracranial metastases. More recently, a diagnosis-specific graded prognostic assessment (GPA) has been developed, reflective of the strong role of primary tumor histopathology. Algorithms have been proposed using the functional status, systemic disease status, and number of metastases, such that patients with the worst prognosis are identified and recommended WBRT only, in contrast to patients with more favorable factors who may be recommended surgery or SRS, possibly followed by WBRT.

WBRT is a well-established and widely used treatment of brain metastases. In a landmark trial comparing the effect of surgical resection in addition to WBRT, Patchell and colleagues randomized a total of 48 patients with a single brain metastasis to either surgical resection or a biopsy only (in patients with supratentorial disease) followed by WBRT. The radiation was delivered to a total dose of 36 Gy through two parallel-opposed lateral portals, with the treatment field including the entire brain and meninges to the level of foramen magnum. Patients receiving surgery and WBRT had a significantly longer survival than those not receiving surgery, with a median survival of 40 weeks versus 15 weeks, respectively ( P <.01). On multivariate analysis, only surgical treatment of brain metastasis was associated with a better functional status and quality of life on multivariate analysis. Patchell and colleagues later performed a multicenter randomized trial of treatment of a single brain metastasis with surgical resection followed by radiotherapy versus surgical resection alone. A total of 49 patients were randomized to postoperative WBRT and 46 patients to resection alone, with WBRT given to a total dose of 50.4 Gy via parallel-opposed lateral fields. They demonstrated a significantly decreased rate of recurrence at the operative bed and remainder of the brain with receipt of WBRT. Moreover, patients who received postoperative WBRT were significantly less likely to die of neurologic causes, although no difference in overall survival was shown.

WBRT, however, is associated with long-term permanent neurocognitive effects and adverse effects on quality of life. DeAngelis and coworkers described WBRT-induced dementia causing severe disability in 1.9% to 5.1% of patients receiving a total dose ranging from 25 Gy to 39 Gy, attributing this to large fractions of radiation (3–6 Gy per day) delivered. More recently, Chang and colleagues examined neurocognitive outcomes as assessed by a formal tool, the Hopkins Verbal Learning Test-Revised, in patients with one to three brain metastases randomized to SRS alone versus SRS with the addition of WBRT. Although 73% of patients who received SRS followed by WBRT were free from CNS recurrence at 1 year, compared with 27% who received SRS alone, the trial met early stopping rules because of a significant predicted decline in learning and memory function.

To ameliorate the neurocognitive effects of WBRT, the use of memantine, an N -methyl- d -aspartate-receptor antagonist, has been explored in the setting of a randomized, double-blinded, placebo-controlled trial. This randomized patients receiving WBRT to memantine (20 mg/d) versus placebo, for a total of 24 weeks, with the observation of significantly longer time to cognitive decline in patients receiving memantine and a strong trend toward less decline in delayed recall in the memantine arm at 24 weeks ( P = .059). The probability of cognitive function failure at 24 weeks was 53.8% in the memantine arm and 64.9% in the placebo arm. Better results were seen in the memantine arm for executive function, processing speed, and delayed recognition at 24 weeks. Memantine was well tolerated and had a toxicity profile similar to placebo.

Other efforts directed toward the preservation of neurocognitive function include the use of conformal radiotherapy, specifically through the use of intensity modulated radiation therapy to spare the hippocampal neural stem-cell compartment, for which radiation-induced damage has been implicated in causing cognitive decline. This hypothesis was examined in Radiation Therapy Oncology Group (RTOG) 0933, a single-arm, phase 2 multi-institutional trial that enrolled recursive partitioning analysis (RPA) class I-II patients with no brain metastases within a 5-mm margin around either hippocampus and examined the primary end point of Hopkins Verbal Learning Test-Revised delayed recall. Use of hippocampal-avoiding intensity modulated radiation therapy was associated with a 7% mean relative decline in this metric from baseline, significantly lower compared with the historical control, and additionally, without reported decline in patient-reported quality-of-life scores. Given these encouraging positive results, and observed benefit with the use of memantine, NRG-CC001, a randomized phase 3 trial of memantine and WBRT with or without hippocampal avoidance, launched in July of 2015 and is presently accruing.

SRS, the administration of highly conformal high dose external beam radiotherapy typically in one to five total fractions, is another frequently used modality for the treatment of brain metastases. Common treatment platforms include gamma knife, cyber knife, and linear accelerator based SRS. SRS alone, WBRT alone, and both SRS and WBRT have all been noted to have level 1 evidence to support their use in patients with multiple brain metastases less than 3 to 4 cm in size with a life expectancy of greater than 3 months.

European Organisation for Research and Treatment of Cancer trials 22952–26001 examined patients with one to three metastases who underwent surgical resection or SRS, and were then randomized to WBRT versus observation. No difference in median time to loss of functional independence (World Health Organization [WHO] performance status >2) was noted. Although intracranial progression caused death in 44% of patients in the observation arm and in 28% of patients in the WBRT arm, no difference was observed in overall survival between groups. Conversely, the addition of an SRS boost to WBRT was examined in RTOG 9508, in which patients with one to three brain metastases were randomized to receive WBRT with or without SRS. Patients in the SRS boost group were more likely to have a stable or improved KPS score at 6-month follow-up than were patients allocated WBRT alone (43% vs 27%, respectively; P = .03). In addition, the authors noted a survival benefit in the SRS group in recursive partitioning analysis class I patients. This suggests that the local control afforded by WBRT alone is suboptimal, at least in a subgroup of patients.

The results of these two trials, together with other data, including the aforementioned work by Chang and colleagues highlighting the adverse neurocognitive effects of WBRT, has led to ongoing interest in the avoidance of WBRT in favor of SRS alone in favorable subsets of patients. A recent individual patient meta-analysis pooling multiple phase 3 trials examining the addition of WBRT to SRS has generated considerable interest in this regard. Identifying a total of 364 patients, the authors demonstrated that the use of SRS alone was associated with improved survival in patients age 50 or younger, and that avoidance of WBRT in this population did not result in increased failure elsewhere in the brain.

The cost-effectiveness of WBRT has recently been investigated in the QUARTZ study, which examined patients with metastatic non–small cell lung cancer with brain metastases unsuitable for either surgical resection or SRS who were randomly assigned to WBRT and dexamethasone versus best supportive care. The authors noted a benefit in terms of mean quality-adjusted life-year of only 4.7 days with WBRT. Thus, hospice care should be strongly considered in unfavorable patient subsets, for which WBRT seems to offer minimal benefit.

Primary brain malignancies

Traditionally, characterization of brain tumors has relied on histopathologic features, such as morphology, immunohistochemistry, and other testing to describe tumor histogenesis and degree of differentiation. With recent advances in the molecular characterization of various primary CNS malignancies, the WHO classification of brain tumors has undergone a major restructuring. Although a full description of the various changes being instituted is beyond the scope of this article, highlights include restructuring of diffuse gliomas, such as glioblastoma multiforme (GBM), to include genetic prognostic factors of isocitrate dehydrogenase (IDH) mutation wild-type versus IDH-mutant and H3 K27M–mutant diffuse midline glioma. Ependymoma is further characterized by the presence of RELA gene fusion and medulloblastoma is further characterized by molecular subtype as WNT-activated versus SHH-activated. Embryonal tumors with multilayered rosettes are characterized by upregulation of the oncogenic miRNA cluster C19MC. These exciting changes herald an era when perhaps molecular characteristics may supercede traditional clinicopathologic characteristics in terms of prognostic and therapeutic significance.

High-grade glioma

GBM is the most common primary malignant brain malignancy of adulthood, and the most devastating, with a median survival of 15 months, occurring with an average incidence rate of approximately 3 per 100,000 people, and with a median age at diagnosis of 64. The exploration of the use of radiotherapy to treat GBM in the 1970s with radiotherapy demonstrated a more than doubling of median survival in comparison with best supportive care. The standard of care treatment of WHO grade 4 GBM remains maximal safe resection followed by chemoradiation with concurrent and adjuvant temozolomide delivered to the resection cavity and residual tumor. This treatment regimen was established by the seminal trial of Stupp and colleagues, which randomized 573 patients with histologically proven debulked GBM to receive focal brain radiation of 60 Gy in 30 fractions over 6 weeks, versus the same radiation with concurrent daily temozolomide followed by adjuvant temozolomide for six 28-day cycles. With a median follow-up of 28 months, the median survival was 14.6 months with radiotherapy plus temozolomide and 12.1 months with radiotherapy alone. Perhaps even more striking was a reported 2-year overall survival rate of 26.5% with radiotherapy plus temozolomide and 10.4% with radiotherapy alone, which on longer follow-up, demonstrated a 5-year overall survival of 9.8% with combined treatment versus 1.9% with radiotherapy alone. Thus, the addition of concurrent temozolomide and adjuvant temozolomide to radiotherapy led to an approximate five-fold increase in 5-year overall survival, a historic advance in the treatment of GBM.

One of the seminal findings in the Stupp and colleagues trial was the prognostic importance of methylation of the O6-methylguanine–DNA methyltransferase (MGMT) DNA repair enzyme. Epigenetic silencing via promoter methylation of MGMT, whose gene product is an enzyme that catalyzes the removal of alkyl groups from the O6 position of guanine, impairs the ability of tumor cells to repair DNA damage. Patients with MGMT promoter methylation who received radiotherapy and temozolomide had a 2-year overall survival of 46%, whereas only 23% of patients with the MGMT promoter methylation who were assigned to radiotherapy alone were alive at 2 years. However, patients without MGMT promoter methylation had 2-year overall survival of 14% receiving radiotherapy and temozolomide versus less than 2% in those receiving radiotherapy only.

Treatment of GBM in the elderly, who have a bleak prognosis, has also been examined. A 2007 trial randomized good performance status patients 70 or older to 50 Gy of focal radiotherapy versus supportive care only and demonstrated a marked survival benefit. The hazard ratio of death in the radiotherapy group was 0.42 with a P value of 0.002. Roa and colleagues randomized 100 patients age 60 or older to receive surgical resection and standard radiotherapy (60 Gy in 30 fractions for 6 weeks) versus a shortened course of radiotherapy (40 Gy in 15 fractions for 2 weeks) without temozolomide. No difference in median survival from diagnosis was noted (5.9 months and 6.1 months in the standard and shortened course, respectively). Additionally, no difference was seen between the two groups in KPS scores tracked over time. More recently, Roa and colleagues examined a further radiotherapy dose reduction in elderly and/or frail patients, comparing their previously examined 15 fraction treatment with a schedule of 25 Gy in five daily fractions for 1 week. In terms of overall survival, progression-free survival, and quality of life measures, the five-fraction regimen was noninferior. Thus, elderly patients with GBM can be reasonably afforded a 1-week course of radiotherapy without the burden of an extended course of treatment. Finally, in the elderly patient population, consideration can be given to temozolomide only. The NOA-08 study group of the German Cancer Society examined temozolomide versus radiotherapy in patients older than 65 and with a KPS of at least 60 with anaplastic astrocytoma or GBM. This demonstrated a median survival of 8.6 months in the temozolomide group and 9.6 months in the radiotherapy group, within their 25% margin of noninferiority. Examining tumor MGMT methylation status, patients with methylation who received temozolomide had a superior event-free survival compared with patients who underwent radiotherapy. Thus, temozolomide alone in this patient population is yet another option, with the selection of patients informed by MGMT status.

An exciting recent development in the treatment of GBM has been the advent of alternating electric fields technology. Specifically, the use of low-intensity, intermediate frequency (100–300 kHz) has been investigated and shown to induce growth inhibitory effects in proliferating cells and cell death in neoplastic cells by disrupting the polymerization and depolymerization of microtubules in mitosis. Stupp and colleagues examined the addition of tumor-treating fields to conventional chemoradiation in the initial treatment of GBM. This EF-14 trial randomized a total of 466 patients in a 2:1 fashion who had undergone resection and radiotherapy with temozolomide, stratified by extent of resection and MGMT status, to receiving adjuvant temozolomide and 200-kHz treatment field therapy delivered via transducer arrays applied to the shaved scalp. Examining a primary end point of progression-free survival, the trial met a prespecified stopping rule for experimental treatment success, with a median progression-free survival in the intent-to-treat population of 7.1 months in the alternating electric fields plus temozolomide group (n = 196) versus 4.0 months in the temozolomide alone group (n = 84), with an estimated hazard ratio of 0.62 for progression ( P = .001). Even more striking was a benefit in median overall survival, with a median survival of 20.5 months in the alternating electric fields plus temozolomide group and 15.6 months in the temozolomide alone group, with an estimated hazard ratio of 0.64 for death ( P = .004). Moreover, alternating electric field was well tolerated, with a chief toxicity of grade 1 to 2 skin reaction characterized by localized dermatitis underneath the area of the transducer leads from the alternating electric field device in 43% of patients and more severe skin reaction (grade 3) in 2% of patients.

Despite some logistical challenges, these results represent a paradigm shift in the locoregional treatment of GBM, and perhaps herald the addition of this exciting new technology to the therapeutic armamentarium in other malignancies. For instance, the PANOVA trial is currently ongoing, which seeks to examine the safety and preliminary efficacy of the use of tumor-treating fields with chemotherapy in the setting of unresectable pancreatic cancer.

A major advance in immuno-oncology has been the development of checkpoint inhibitors, specifically the development of targeted agents directed at programmed death cell 1 (PD-1) protein and its ligand (PD-1L), an inhibitory cell surface receptor involved in inhibiting T-cell effector function and maintaining normal tissue tolerance. The use of radiotherapy, particularly SRS, in conjunction with immune checkpoint blockade is an evolving therapeutic strategy. Preclinical data suggest concurrent PD-1 blockade potentiates not only the local effect of radiotherapy, but also possibly a secondary downstream systemic or abscopal effect in melanoma and renal cell carcinoma models. Zeng and colleagues examined the combination of anti-PD-1 immunotherapy with SRS in a mouse orthotopic GBM model where mice underwent treatment with anti-PD-1 immunotherapy only, SRS only, or SRS with anti-PD-1 immunotherapy, with a control group undergoing no treatment. The median survival in the control group of mice was 25 days, compared with 27 days in the anti-PD-1 antibody arm, 28 days in the radiation arm, and 53 days in the radiation plus anti-PD-1 ( P <.05). Perhaps even more striking was the observation that only mice in the combined modality treatment had long-term survival beyond 180 days. MEDI473 is a phase 2 trial currently recruiting patients with newly diagnosed unmethylated MGMT GBM investigating an anti-PD-L1 monoclonal antibody with standard chemoradiation.