In this review, the role of radiotherapy in the management of brain metastases is considered from a historical perspective, in the context of other treatment modalities, and with regard to different radiotherapy techniques.
Brain metastases represent the most common intracranial malignancy. The management of brain metastases is often complex, multidisciplinary, and highly individualized. In this review, the role of radiotherapy in the management of brain metastases is considered from a historical perspective, in the context of other treatment modalities, and with regard to different radiotherapy techniques. Data regarding coordination of systemic therapy with radiotherapy is reviewed, outlining historical findings and a paucity of data in the context of novel systemic cytotoxic and biologic therapies with whole brain radiotherapy and stereotactic radiosurgery. Controversial aspects of patient management are considered, including multifactorial patient, tumor, and treatment factors that inform treatment recommendations for individual patients. Current clinical controversies and research endeavors are reviewed as they relate to maximization of therapeutic efficacy, minimization of toxicity, and optimization of quality of life. General evidence-based approaches to management of brain metastases are considered and published guidelines are addressed in this review.
Brain metastases represent the most common intracranial malignancy in adults and are reported to develop in 10% to 40% of patients with a known extracranial primary malignancy.1 While the exact incidence is uncertain, present estimates suggest that approximately 180,000 patients in the United States are diagnosed with brain metastases annually.1—3 With advances in systemic therapies and the potential increased longevity of cancer patients, the absolute incidence of brain metastases may increase, as may survival following therapeutic intervention for brain metastases. Thus, more than ever before, therapeutic strategies are urgently needed that offer effective, durable control of brain metastases with a favorable toxicity profile.
The management of brain metastases has evolved over the past several decades but has continued to rely on radiotherapy and neurosurgical resection as the predominant treatment modalities. Most systemic therapies including chemotherapy have demonstrated poor penetration of the blood brain barrier and have had limited efficacy against brain metastases. As such, whole brain radiotherapy (WBRT) has been the classic treatment for brain metastases for decades, and includes irradiation of the entire cerebrum, cerebellum, and brainstem.
The techniques utilized in the delivery of WBRT are generally straightforward but beyond the scope of this review. For the purposes of this review, the treatment may be thought of as one lateral x-ray delivered from the patient’s left and one lateral x-ray delivered from the patient’s right. Throughout the 1960s to 1980s several studies evaluating different WBRT dosing and fractionation schemes were performed.4—10 In recent decades, most institutions have used WBRT regimens involving between 5 and 20 daily fractions of radiotherapy treatment, with 10 fractions being the most common. Landmark randomized controlled trials in the 1990s evaluated the role of WBRT, surgery, and various combinations of WBRT and surgery for patients with a single brain metastasis.11—13
For patients treated with WBRT, the addition of neurosurgical resection to WBRT in patients with a single brain metastasis resulted in decreased rates of local recurrence (20% recurrence with WBRT and surgery vs 52% recurrence with WBRT and no surgery) and improved overall survival (OS; 40 weeks with surgery vs 15 weeks with WBRT alone).13 Multiple studies suggest the survival benefit is most pronounced in young patients with a single brain metastasis, good performance status, and absent/controlled extracranial disease.14,15 Of note, one study evaluating the role of surgery in patients with a Karnofsky Performance Score (KPS) of 50% or greater and one cerebral metastasis undergoing WBRT failed to show a survival benefit to the addition of surgery.16
Additional factors influencing recommendation for surgical resection include the rapidity of onset of symptoms, size of a metastasis, location (eloquent vs non-eloquent location), number of metastases, and whether a histopathologic diagnosis has previously been established. In a study evaluating the role of WBRT following upfront neurosurgical intervention for patients with a single brain metastasis, the addition of WBRT resulted in no increase in OS but did demonstrate decreased rates of local failure (10% with WBRT and surgery vs 46% with surgery alone) and decreased rates of any intracranial failure (18% with WBRT and surgery vs 70% with surgery alone).12 While no OS benefit was appreciated, the addition of WBRT to surgery reduced the rate of neurologic death in patients with a single brain metastasis (14% with surgery and WBRT vs 44% with surgery alone).12
The findings of this study were supported by an additional multiinstitutional randomized trial by the European Organization for Research and Treatment of Cancer (EORTC) evaluating the role of WBRT after surgical resection or stereotactic radiosurgery (SRS) for patients with 1-3 brain metastases.17 The authors found that regardless of the local treatment modality (surgery or SRS), the addition of WBRT decreased rates of recurrence at the site of index lesion(s) (27% with WBRT and surgery vs 59% with surgery alone; 19% with WBRT with SRS vs 31% with SRS alone) and reduced the likelihood of patients developing new brain metastases (23% with WBRT and surgery vs 42% with surgery alone; 33% with WBRT and SRS vs 48% with SRS alone).17
As in the prior study of neurosurgical resection with/ without WBRT, no OS advantage was appreciated, though rates of neurologic death were reduced (44% without WBRT vs 28% with WBRT). The potential advantages of adjuvant WBRT must be weighed against potential drawbacks of WBRT, including potential neurologic toxicity and a recently reported transient reduction in health-related quality of life in patients treated with WBRT in the EORTC phase III study of SRS or surgery followed by WBRT or observation.18
At the same time that the roles of WBRT and surgery were being evaluated regarding their relative benefits in the management of brain metastases, a more recently developed radiotherapeutic technique, SRS, was being developed and refined as a mechanism for conformal, ablative high-dose radiation treatment of individual metastatic lesions, instead of irradiation of the entire brain.19
Though SRS was developed by Lars Leksell in the early 1950s, it was not widely utilized in the United States until the mid-1990s and early 2000s, when radiographic imaging techniques caught up with treatment delivery systems, allowing for improved visualization of intracranial radiotherapy targets with CT- and MRI-based imaging. Presently, several SRS platforms exist. Treatment may be delivered with: 1) a standard linear accelerator (LINACbased SRS) with adequate stereotactic capabilities; 2) the CyberKnife® system, which is a photon-only single-energy LINAC mounted on a robotic arm; 3) TomoTherapy®, which is helical delivery LINAC-based SRS; 4) the GammaKnife® system, which delivers SRS using highly focused photons from 192- 201 60Co radionuclide sources; and 5) particle-therapy SRS, most commonly utilizing protons (see Table). The intricacies of the different SRS treatment delivery techniques are beyond the scope of this review. However, all of these systems deliver multiple small photon beams from a large number of angles. The contribution of each beam (or beamlet) is very small and delivers an exceedingly low dose of radiation, but, at the point where all of the individual beams (or beamlets) converge, a very large dose of radiation is delivered. Protons may also be used in delivery of SRS, though in the United States SRS is generally delivered with photons.
Type of Radiation Delivered
Invasive Head Frame Required
Source of Radiation
Depends on technique
Stereoscopic x-rays and/or cone-beam CT
Able to perform stereotactic and nonstereotactic radiotherapy to other parts of the body; electron therapy
Alignment must be checked and verified before each isocentric treatment; requires intense physics assessment before each patient’s treatment
CyberKnife® (Single energy linear accelerator)
Small linear accelerator mounted on a robotic arm
Linear accelerator on robotic arm allows treatment delivery with nearly unlimited treatment angles
Single energy photon; no electrons; no portal imaging
Linear accelerator with helical delivery of x-rays
Capable of treating very long fields in a continuous fashion
Treatment plan does not have traditional review
Depends on technique
192-201 radioactive cobalt-60 sources
Rigid frame placement
Multiple isocenters easily treated in same session; daily quality assurance allows treatment of multiple lesions and multiple patients
Radioactive source/ regulations; if patient undergoes multiple fractions (SRT), often requires overnight stay with frame in place; treatment is limited to high cervical spine and above
No exit dose permits targeting of lesions near critical structures
Access to proton therapy is limited
Upon completion of an SRS dose-finding study,20 randomized controlled trials evaluated the role of SRS alone, WBRT alone, SRS with/without addition of WBRT (evaluating whether WBRT added any benefit to SRS), and WBRT with/without SRS (conversely evaluating whether SRS added any benefit in addition to WBRT). Patients with 1-3 brain metastases treated with SRS in addition to WBRT were noted to have better rates of local control at one year (82% for WBRT + SRS) than patients treated with WBRT alone (71% for patients treated with WBRT alone).21 The same study also noted a survival benefit in patients with favorable histology (squamous cell or nonsmall cell carcinoma of the lung) and in patients <65 years of age with KPS of ≥ 70%, controlled primary disease, and no extracranial metastases.21 Conversely, when WBRT was added to SRS in patients with 1-4 brain metastases, investigators found improved local control with the addition of WBRT (89% local control with SRS and WBRT vs 73% local control with SRS alone) as well as reduced rates of development of new brain metastases (42% developed new brain metastases with SRS and WBRT vs 64% with SRS alone).22 No difference in preservation of neurologic function was noted between patients treated with SRS alone and with SRS and WBRT.22 These studies echoed the findings of the previously discussed studies that demonstrated a survival benefit for the addition of local therapy to WBRT and a decrease of both local recurrence and in the rate of patients developing additional brain metastases when WBRT is added to surgery.
In 2010, the American Association of Neurologic Surgeons (AANS) and Congress of Neurologic Surgeons (CNS) published guidelines recommending SRS in combination with WBRT for patients with KPS ≥ 70% and 1 to 4 brain metastases all measuring ≤ 3 cm in greatest diameter.23 This recommendation was supported in 2012 by evidence-based guidelines from the American Society for Radiation Oncology (ASTRO), which advised that patients with anticipated survival of at least 3 months and a single brain metastasis ≤ 3-4 cm undergo SRS ± WBRT (or surgery + WBRT).24 In the same guidelines, patients with good prognosis and 1-3 brain metastases ≤ 3-4 cm were recommended for SRS ± WBRT or WBRT alone.24
Thus, WBRT is considered standard for most patients with > 4 brain metastases, and is also used for many patients with 1-4 metastases depending on individual patient factors. SRS is typically avoided for metastases > 4 cm in greatest diameter, and surgery and/or fractionated stereotactic radiotherapy (SRT) may be considered. While there are several reports describing the outcomes of SRS in the management of patients with > 4 brain metastases,25 SRS in this patient population remains controversial.
Patients initially managed with SRS alone have been shown to experience greater rates of local and distant intracranial failure (a new brain metastasis) than those treated with both SRS and WBRT.17,22,26 This increased intracranial failure is potentially associated with neurologic morbidity, and patients may require more salvage therapies for local and distant intracranial failure than patients who undergo initial therapeutic intervention with SRS and WBRT.27 However, the shorter overall treatment time of SRS/SRT alone compared with WBRT may allow patients to pursue additional systemic therapy without delay and may offer an opportunity to avoid potential decreases in healthrelated quality of life that patients may experience with SRS/ surgery and WBRT.18
In addition to maximizing local control and OS, optimization of neurocognitive function in patients with brain metastases becomes increasingly important. A majority of patients with brain metastases in randomized controlled trials have demonstrated impaired baseline neurocognitive function.28,29 Welzel et al noted below-normal neurocognitive function prior to WBRT, primarily noted in the verbal domain, in 62% of patients due to receive prophylactic cranial irradiation (36 Gy/18 fx), 44% of patients due to receive therapeutic cranial irradiation (40 Gy/20 fx), and 13% of controls (receiving breast radiotherapy).30 This finding may be related to the brain metastases themselves, prior systemic therapy, and/ or paraneoplastic processes. In multiple studies, intracranial failure at the site of initial metastatic involvement, or distant intracranial failure, has been associated with neurologic deterioration following SRS alone and following WBRT alone.22,29
The importance of intracranial control on neurocognition was supported by Li et al, who reported a correlation between post-treatment volumetric reduction in intracranial disease burden with survival and neurocognitive function in the domains of executive function and fine motor coordination.31 However, the same prospective study noted no statistically significant correlation between reduction in volume of intracranial metastatic burden and changes in patients’ performance on memory tasks, specifically delayed recall on Hopkins Verbal Learning Test (HVLT-R).31 As this patient cohort was noted to have earlier decline in HVLT and HVLT-R than other neurocognitive domains, the authors concluded that certain neurocognitive domains may exhibit differential radiosensitivity.31 The impact of WBRT on neurocognition had been previously reported in a retrospective review by DeAngelis et al in 1989, in which up to 5.1% of patients treated with WBRT developed severe radiation-associated dementia at a median of 14 months, though the authors estimated the risk may be as high as 19%, given their retrospective identification of only patients with severe dementia. They also note that 9 of 12 patients were treated with multiple fractions of > 300 cGy/ fx, which is a higher dose of radiotherapy per fraction than is used in many modern regimens.32 This study has limitations given that the true denominator of patients evaluated is unclear, and a proportion of these patients had a major improvement in symptoms after treatment for normal pressure hydrocephalus. An association between normal pressure hydrocephalus and radiotherapy is unclear, and patients treated with WBRT are presently not commonly diagnosed with normal pressure hydrocephalus.
The impact of WBRT on cognition, specifically verbal memory and verbal learning, has been reported in multiple studies to have a potentially deleterious effect.26,30,31 In 2009, Chang et al published results from their single institution randomized controlled trial of SRS ± WBRT in patients with 1-3 brain metastases. The study met its early stopping criteria for the primary endpoint of a >5 point reduction in HVLT-R from baseline findings, which was noted in 49% of patients with SRS + WBRT and 23 % of patients with SRS alone.26 A similar impact on a related verbal learning neurocognitive tool (the Auditory Verbal Learning Test) was noted at a similar time point of 6 to 8 weeks following WBRT in a descriptive study one year earlier by Welzel et al, who also reported the greatest decline was noted in patients with above average initial performance, while patients with initial documentation of neurocognition of average or below average levels demonstrated less pronounced decreases from baseline or improvement in neurocognitive testing.30
Taken together, these studies suggest WBRT may preferentially impact verbal learning 2 to 4 months following WBRT. Additional follow-up at subsequent time points is indicated to evaluate the long-term effects of radiotherapy. Prevention of neurocognitive deficits as a result of WBRT with memantine was recently evaluated in a randomized controlled trial (RTOG 0614), in which patients with brain metastases undergoing WBRT were randomized to receive concurrent memantine or placebo, which was continued for 6 months following WBRT administration.33
Patients receiving memantine, an inhibitor of the glutamatergic NMDA receptor, demonstrated a significantly longer time to neurocognitive decline, with reduced neurocognitive failure at 24 weeks (53.8% vs 64.9%, defined as 2 standard deviations [SD] below the patient’s baseline neurocognitive function). Also noted was superior executive functioning in the memantine group at 8 and 16 weeks, superior processing speed at 24 weeks, and superior delayed recognition at 24 weeks.33 Another potential mechanism by which neurocognitive side effects may be minimized is through hippocampal-sparing WBRT techniques, as researchers previously have noted a correlation between neurocognitive function and radiotherapy to the hippocampus.34,35
The RTOG has completed accrual of a phase II study evaluating the role of hippocampal sparing WBRT in patients with at least one brain metastasis (excluding small cell lung cancer, germ cell tumors, and hematologic malignancies), KPS ≥ 70%, with a primary objective of evaluating patients’ delayed recall (HVLT-R) at 4 months following hippocampalsparing WBRT, a time point known to be associated with a decrease in neurocognitive testing in the HVLT-R domain.
Initial results suggest a less pronounced decrease in HVLTdelayed recall at 4 months than historical controls.36 Plans for a phase III randomized control trial evaluating hippocampalsparing WBRT with conventional WBRT in patients undergoing prophylactic cranial irradiation for small cell lung cancer (RTOG 1316) are currently under way.
The role of systemic therapy in the management of brain metastases, as previously noted, has historically been limited, as most cytotoxic agents have poor penetration of the blood brain barrier. Moreover, several studies have demonstrated conflicting efficacy in improving OS and tolerability. Limited data exists regarding administration of modern cytotoxic and non-cytotoxic systemic therapies with concurrent WBRT.
Following publication of two articles that reported a 15%-25% rate of severe neurocognitive dysfunction in patients with concurrent chemotherapy and WBRT (including 10% fatal leukoencephalopathy in one study) treated for primary central nervous system lymphoma, concurrent chemotherapy with whole brain radiotherapy was generally cautioned against.37,38
A delay between chemotherapy administration and WBRT may reduce the likelihood and potential severity of toxicity of systemic therapy combined with WBRT. Theoretically, a delay may allow for repair of a potential radiotherapy-associated disruption in the blood-brain barrier that may contribute to neurocognitive toxicity. Several conflicting studies have been published regarding the tolerability and efficacy of several systemic therapies administered concurrently with WBRT.39- 47 Concurrent systemic therapy and radiotherapy has recently been revisited,48—58 but remains controversial. The interval between systemic therapy and radiotherapy—and whether a waiting period is required with more recently developed systemic therapies—is not well understood and is likely to be an area of active research in the future.
Patients with progressive or persistent disease at the site of previously treated brain metastases may be considered on an individual basis.59 Management options in these scenarios include surgical resection, SRS ± WBRT, repeat WBRT, and/or supportive care. These options are considered in the context of symptomatology, time to recurrence, overall prognosis, location, size, and nature of prior therapeutic interventions. New brain metastases at previously untreated regions may be managed as per recommendations for index lesions as discussed above.
Several prognostic tools have been developed to aid in identifying patient populations in whom aggressive therapy may be warranted.60,61 Most recently, a nomogram has been developed that may aid in estimating survival in patients with brain metastases.62 This nomogram is based on site and histology, status of primary disease, metastatic spread, age, KPS, and number of brain lesions. These factors are included in aiding prediction of 6-month survival probability, 12-month survival probability, and predicted median survival (days).62 The nomogram may assist in further identifying patients in whom various brain metastasis therapies may provide the greatest benefit.
For patients with newly diagnosed brain metastases, corticosteroid administration is recommended for associated neurologic symptoms.63 Corticosteroid use in patients with asymptomatic brain metastases is not well delineated. Surgical resection is advised for patients with oligometastatic disease with significant symptoms, patients requiring histopathologic diagnosis, and in patients with large metastases, recognizing there are conflicting size cut-offs at which point surgical intervention is recommended over SRS/SRT.
Anti-epileptic agents are recommended only for patients with prior seizure activity. For all patients undergoing craniotomy for resection of a parenchymal brain metastasis, the perioperative risk of seizure is approximately 3%-5% (3% rate of clinically significant seizure activity), and was not shown to be reduced with perioperative phenytoin in a recent study.64 The same study reported an overall rate of seizure activity in 13%-15% of all patients with brain metastases who had undergone prior resection for metastatic disease.64 While literature reviews and meta-analyses demonstrate no benefit or insufficient evidence to support prophylactic use of antiepileptic medications in patients undergoing neurosurgical resection for brain tumors,65,66 single-institution data suggest newer agents including levetiracetam are tolerable and may be effective in preventing perioperative seizures in this patient population when compared with phenytoin.67,68 Patients who have undergone neurosurgical intervention and have been prophylactically placed on anti-epileptic regimens should be tapered off approximately 1-2 weeks after surgery.69 Postoperative radiotherapy is considered standard, using WBRT and/or SRS to the surgical cavity,70—86 even though there has not been a survival advantage associated with postoperative radiotherapy. Similarly, patients with 1 to 3 brain metastases ≤ 4 cm may be considered for SRS ± WBRT; those undergoing WBRT may be considered for memantine during WBRT and for 6 months following WBRT. Patients with metastases ≥ 4 cm may be considered for multifraction SRT to an intact metastasis and/or to a resection cavity in the postoperative setting. Patients with > 4 metastases may be considered for SRS, though this is currently not considered standard of care.
Technologic advances over the past several decades have resulted in improved imaging techniques, allowing for improved sensitivity in detection of brain metastases and earlier treatment of smaller, asymptomatic lesions. This progress in radiologic imaging has been paralleled by improvements in neurosurgical techniques, novel systemic agents, immunomodulators, developments in radiation delivery capabilities, further understanding of tumor biology, and deeper appreciation of potential treatment toxicities. In the coming decades, anticipated increases in cancer diagnoses paired with these therapeutic advances will likely result in an increased number of cancer survivors. It stands to reason that further understanding of patient, tumor, and treatment factors will allow for greater individualization of therapeutic recommendations with goals of optimizing local control of existing brain metastases, preventing development of new brain metastases, and minimizing treatment-related toxicity.
ABOUT THE AUTHORS
Affiliations: Abigail L. Stockham, MD, is a neuro-radiation oncology fellow at the Brigham and Women’s Hospital/Dana-Farber Cancer Institute in Boston, MA. Nils D. Arvold, MD, is attending physician, Neuro-Radiation Oncology, Brigham and Women’s Hospital/Dana- Farber Cancer Institute.
Disclosures: Drs. Stockham and Arvold report no conflicts of interest to disclose.
Address correspondence to:
Abigail L. Stockham, MD, Brigham and Women’s Hospital/Dana- Farber Cancer Institute, 75 Francis Street, ASB1-L2, Boston, MA 02115. Phone: (617) 732-6313, #3; fax: (617) 975-0932; Email: email@example.com.