Multidisciplinary Management of Brain Metastases

Learning Objectives

After completing this course, the reader will be able to:

1. Identify the clinical factors that predict survival after a diagnosis of brain metastasis.

2. Select appropriate multidisciplinary treatments for patients with new and recurrent brain metastases.

3. Describe the circumstances in which focal therapy, such as surgery or stereotactic radiosurgery, is likely to be beneficial for patients with brain metastases.

CME Access and take the CME test online and receive 1 AMA PRA Category 1 Credit™ at CME.TheOncologist.com

Abstract

Metastatic brain tumors are the most common intracranial neoplasms in adults. The incidence of brain metastases appears to be rising as a result of superior imaging modalities, earlier detection, and more effective treatment of systemic disease. Therapeutic approaches to brain metastases include surgery, whole brain radiotherapy (WBRT), stereotactic radiosurgery (SRS), and chemotherapy. Treatment decisions must take into account clinical prognostic factors in order to maximize survival and neurologic function whilst avoiding unnecessary treatments. The goal of this article is to review important prognostic factors that may guide treatment selection, discuss the roles of surgery, radiation, and chemotherapy in the treatment of patients with brain metastases, and present new directions in brain metastasis therapy under active investigation. In the future, patients will benefit from a multidisciplinary approach focused on the integration of surgical, radiation, and chemotherapeutic options with the goal of prolonging survival, preserving neurologic and neurocognitive function, and maximizing quality of life.

Neoplasm metastasis, Brain neoplasms, Therapeutics, Ionizing radiation, Drug therapy

Introduction

Metastatic brain tumors are the most common intracranial neoplasm in adults, and although the exact incidence is unknown, it has been estimated to be as high as 200,000 cases per year in the U.S. alone [1]. Recent population-based data suggest that 8%–10% of adults with cancer will develop symptomatic brain metastases during their lives [2, 3]. The majority of brain metastases originate from one of three primary malignancies: lung cancer (40%–50%), breast cancer (15%–25%), and melanoma (5%–20%). Among these, melanoma has the highest propensity to metastasize to the brain, with a 50% rate of brain involvement reported in patients dying of melanoma [4]. The frequency of metastatic brain tumors appears to be rising as a result of superior imaging modalities and earlier detection as well as longer survival after a primary cancer diagnosis because of more effective treatment of systemic disease.

The distribution of brain metastases generally parallels blood flow, with 80% occurring in the cerebral hemispheres, 15% in the cerebellum, and 5% in the brainstem [5, 6]. In the era of magnetic resonance imaging (MRI), the majority of patients have multiple brain metastases at diagnosis. Common clinical features include headache, neurological deficit, and seizures. Detailed neuropsychological testing demonstrates cognitive impairment in 65% of patients with brain metastases, usually across multiple domains [7, 8]. Neurologic deficits may be a result of destruction or displacement of brain tissue by expanding tumor, peritumoral edema leading to further disruption of surrounding white matter tracts, increased intracranial pressure, and/or vascular compromise.

The management of brain metastases can be divided into symptomatic and therapeutic strategies. Symptomatic therapy often includes corticosteroids to reduce peritumoral edema and anticonvulsants to prevent recurrent seizures. In addition, there is increasing data to suggest that medications such as methylphenidate and donepezil can improve cognition, mood, and quality of life in patients with brain tumors [9, 10]. Supportive care for patients with brain metastases has been comprehensively reviewed elsewhere [11] and is not discussed further in this review.

Therapeutic approaches to brain metastases include surgery, whole brain radiotherapy (WBRT), stereotactic radiosurgery (SRS), and chemotherapy. Many patients are treated with a combination of these, and treatment decisions must take into account factors such as patient age, functional status, primary tumor type, extent of extracranial disease, prior therapies, and number of intracranial lesions. The goal of this article is to review important prognostic factors that may guide treatment selection, discuss the roles of surgery, WBRT, SRS, and chemotherapy in the treatment of patients with brain metastases, and present new directions in brain metastasis therapy under active investigation.

Prognostic Factors

Appropriate management of patients with brain metastases requires an assessment of independent prognostic factors in order to maximize survival and neurologic function whilst avoiding unnecessary treatments. Demographic and clinical variables predictive of survival in patients with brain metastases have been well studied [12, 13]. Important variables include: age, performance status (most commonly designated by the Karnofsky performance status [KPS] score [14], Table 1), number of brain metastases (single or multiple), primary tumor type (lymphoma, germ cell, and breast versus other), and systemic tumor activity (controlled versus uncontrolled). Of these, the KPS score has consistently been shown to be the major determinant of survival, secondary only to treatment regimen in most studies. Time from primary tumor diagnosis to development of brain metastases holds prognostic value as well, particularly for breast and melanoma primaries, with long intervals being favorable [13, 15].

Table 1

Karnofsky performance status (KPS) scale

A three-tiered prognostic categorization was derived from 1,200 patients in the Radiation Therapy Oncology Group (RTOG) database using recursive partitioning analysis (RPA) [12] and validated in several subsequent studies (Table 2) [16–18]. Patients were pooled from three consecutive RTOG trials spanning 1979–1993 and all received 30–38.4 Gy of WBRT in various fractionation schedules. The overall survival duration for patients in RPA class 1, defined as those with a KPS score ≥70, age <65 years, controlled primary tumor, and no extracranial sites of disease, was 7.1 months. The median survival duration was only 2.3 months for patients in RPA class 3, defined as all patients with a KPS score <70. The median survival duration for the remainder, RPA class 2, was 4.2 months. Although the distinction of single versus multiple brain metastases did not retain significance in the original RPA model, it may hold additional prognostic value within classes 1 and 2 [19]. In 97 patients with a single brain metastasis randomized to surgical resection with or without postoperative WBRT, survival did not differ significantly for RPA class 1 (10.9 months) versus class 2 (9.8 months) [20]. Similarly, in a single-center retrospective analysis of 916 patients treated with WBRT, survival for 2 years or more was observed in both RPA class 1 and class 2 patients [21]. Within both classes, survival duration was significantly longer for patients with a single brain metastasis (13.5 months for RPA class 1, 6.0 months for RPA class 2) compared with those having multiple metastases (8.1 months for RPA class 1, 4.1 months for RPA class 2).

Table 2

Median survival duration according to RPA class for patients treated with WBRT [12]

Abbreviations: KPS, Karnofsky performance status; RPA, recursive partitioning analysis; WBRT, whole brain radiation therapy.

An alternative prognostic scoring system has been derived from patients undergoing SRS, called the score index for radiosurgery (SIR), which takes into account not only age, KPS score, and systemic disease status but also lesion number and largest lesion volume [22, 23]. The 0–10 point SIR scale appears to be comparable to RPA class in predicting outcome but needs to be validated in larger data sets with more varied treatment regimens. What each scale emphasizes is the close association between clinical variables and outcome, and the need to carefully adjust for these factors when comparing competing treatment modalities in nonrandomized and all but large randomized studies.

Treatment Options

WBRT

The mainstay of treatment for brain metastases over the past five decades has been corticosteroids and WBRT. Nonrandomized studies suggest that WBRT increases the median survival time by 3–4 months over approximately 1 month without treatment and 2 months with corticosteroids alone. Although reports of the response rate after WBRT alone vary, complete responses (CRs) or partial responses (PRs) have been documented in approximately 60% of patients in randomized controlled studies conducted by the RTOG [24]. Symptom stabilization or improvement occurs in roughly the same proportion, though symptom response has been vaguely defined in many studies, is rarely independent of corticosteroid use, and may be overestimated [25].

Although several fractionation schedules have been studied, meta-analyses suggest that differences in dose, timing, and fractionation do not significantly alter the median survival times of patients receiving WBRT for brain metastases [26]. The most common regimen employed is 35 Gy delivered in 2.5-Gy fractions over 14 treatment days. Daily fractions >3 Gy likely increase the risk for neurotoxicity. In clinical practice, WBRT is commonly delivered to patients with multiple brain metastases not amenable to surgery or SRS, poor functional status, or active or disseminated systemic disease with effective palliation of neurological symptoms.

Radiation Sensitizers

Multiple attempts have been made to improve upon the results of WBRT alone, particularly by adding agents suspected of having radiosensitizing effects in preclinical models. The following radiosensitizers have been studied in randomized controlled trials, all failing to show benefit in either local brain tumor control or overall survival: lonidamine [27], metronidazole [28], misonidazole [29], motexafin gadolinium [8], bromodeoxyuridine [30], and RSR13 (efiproxiral) [31]. All six trials reported serious adverse effects in the radiosensitizer arm. Based on a subgroup analysis in patients with lung cancer showing a longer median time to neurological progression (5.5 months for WBRT plus motexafin versus 3.7 months for WBRT alone) and improved neurologic function as assessed by a blinded events review committee, a randomized trial of motexafin in this subgroup has been launched. Similarly, the small subset of 42 breast cancer patients in the efiproxiral study showed a doubling of the median survival time with the addition of efiproxiral (hazard ratio, 0.51; p = .003; not adjusted for multiple comparisons), and a randomized trial for breast cancer patients is currently enrolling.

WBRT Plus Chemotherapy

Over the years, several chemotherapeutic agents have been studied in combination with WBRT for patients with brain metastases, including chloroethylnitrosoureas, tegafur, fotemustine, and teniposide [32–34]. Although most have shown higher response rates in the experimental arm, all have been at the expense of greater toxicity with no benefit in overall survival. More recently, the combination of WBRT and low-dose (75 mg/m²) daily temozolomide (TMZ) has shown promising response rates with acceptable toxicity in patients with newly diagnosed brain metastases from a variety of solid tumors [35]. Antonadou et al. [35] found a 96% objective response rate (38% CR rate, 58% PR rate) in 24 assessable patients randomized to TMZ plus WBRT, versus 66% (33% CR rate, 33% PR rate) in 21 assessable patients treated with WBRT alone, although the median survival time did not differ (8.6 months versus 7.0 months; p = .447). A more recent study by Verger et al. [36], enrolling 82 patients with the same study design, failed to replicate the high objective response rate found by the previous authors, although the authors did find a significantly higher neurological progression-free survival rate in the TMZ arm (72% versus 54%; p = .03). A randomized phase III trial (134 patients) was completed by Antonadou et al. [37] with similar results to their first, but final results have not yet been published.

In summary, although several studies have shown promising tolerability and response rates for concurrent TMZ and WBRT, particularly for non-small cell lung cancer (NSCLC) and melanoma, current data do not yet support the widespread use of the combination in patients with new brain metastases. A future treatment strategy may be to assess tumor O⁶-methylguanine–DNA methyltransferase methylation status as a way of preselecting a group of patients with a greater likelihood of responding to TMZ.

Prophylactic Cranial Irradiation

Administering WBRT to cancer patients at high risk for brain relapse but without overt brain metastases at diagnosis, or prophylactic cranial irradiation (PCI), has been the subject of inquiry for over 30 years. In patients with small cell lung cancer (SCLC), who have a ≥50% estimated 2-year risk for central nervous system (CNS) relapse [38], PCI reduces the risk for brain metastases by 50% and increases overall survival by 16%–18% in patients with a CR to chemotherapy based on a meta-analysis of 12 randomized trials [39, 40]. PCI has also been investigated in high-risk NSCLC patients, though not systematically, and there is currently insufficient evidence to support its use in this population [41]. A multicenter randomized phase III trial sponsored by the Eastern Cooperative Oncology Group and RTOG is under way in stage IIIA and IIIB NSCLC patients comparing PCI with observation, with an accrual goal of 1,000 patients [42]. Endpoints include overall survival (primary), CNS relapse rate, quality of life, and neurocognitive function.

The risk for late cognitive decline in SCLC patients after PCI has been addressed in six prospective studies, including two randomized trials. Arriagada et al. [43] studied 300 patients with SCLC in CR who were randomized to PCI or observation and found no significant difference in the 2-year incidence of neuropsychological changes (as assessed by a neurologist) in the two groups. Similarly, Gregor et al. [44] carried out detailed psychometric and quality of life testing on 136 patients enrolled in a multicenter randomized trial of PCI versus observation in SCLC patients achieving a CR after induction chemotherapy and found no difference in neurocognitive function at 12 months. Of note, a significant proportion of patients in both arms were impaired both at baseline and at 12 months, indicating the importance of serial measurements and a control arm when assessing neurocognitive outcomes in cancer patients receiving multimodal therapy. Based on available data, it is probable that PCI delivered either as 20 Gy in 10 fractions or 30 Gy in 10 fractions causes little significant toxicity up to 2 years after PCI, but longer-term follow-up data are needed to fully assess late toxicity.

SRS

SRS employs multiple convergent beams to deliver a single, large dose of radiation to a discrete target volume. The three most common delivery systems are the linear accelerator, gamma knife, and cyclotron, which make use of high-energy photons, gamma rays, and protons, respectively. Most brain metastases have distinct radiographic and pathologic margins, making them attractive targets for SRS. Maximum-tolerated doses of 15, 18, and 24 Gy have been established by the RTOG 90–05 protocol for tumors 31–40 mm, 21–30 mm, and ≤20 mm in maximum diameter, respectively [45]. In addition, a recent retrospective analysis suggests that doses >20 Gy do not improve local control and increase toxicity for lesions ≤20 mm [46]. SRS has emerged as a common treatment modality for newly diagnosed patients, alone or in combination with WBRT, and as salvage therapy for progressive intracranial disease after WBRT.

One-year actuarial local control rates in the range of 71%–79% have been reported with the use of SRS alone for single and multiple brain metastases [46–49]. Response rates are mixed for tumors that have traditionally been considered “radioresistant,” for example, renal cell carcinoma, melanoma, and sarcoma, with some studies showing comparable response rates [50, 51] and others demonstrating local control rates <50% [52, 53]. Importantly, SRS does not address distant failure in the brain. Complications from SRS include early treatment-induced edema, reported in 4%–6% of patients within 1–2 weeks of treatment [54, 55]; seizures within the first 24–48 hours, reported in 2%–6% [54, 56]; and delayed radiation necrosis, reported in 2%–17% [45, 49, 54, 56, 57]. The risk for radiation necrosis increases with larger tumor volume, higher radiation dose, and prior radiotherapy.

WBRT With or Without SRS

Three randomized [58–60] and two retrospective [61, 62] studies have directly examined whether the addition of SRS to WBRT provides therapeutic benefit over WBRT alone (Table 3). In the RTOG 95–08 trial, 333 RPA class 1 and class 2 patients with one to three brain metastases not exceeding 4 cm in largest diameter were randomized to WBRT (37.5 Gy in 15 fractions) with or without SRS [58]. Patients were stratified by number of metastases and status of extracranial disease. No significant difference was seen in the primary outcome measure of median overall survival (5.7 months for WBRT plus SRS versus 6.5 months for WBRT alone; p = .14). However, in the prespecified subset of patients with single lesions, the median survival time was longer in the WBRT plus SRS group (6.9 months versus 4.9 months; p = .04) despite a 19% failure to receive SRS rate. These results are analogous (though not directly comparable) to the surgical randomized trials, which have shown survival benefit in select patients with single lesions treated with WBRT plus resection versus WBRT alone [63, 64].

Table 3

Randomized studies comparing surgery or SRS plus WBRT versus WBRT alone

In patients with multiple [2–3] metastases, the addition of SRS did not improve survival or local control in the RTOG 95–08 trial, contrary to results of a previous small randomized study [59] and a large retrospective series by Sanghavi et al. [61] comparing survival in 500 patients from 10 institutions who had received WBRT plus SRS to all lesions against the historical RTOG RPA cohort [12] (treated with WBRT alone). In that study, the median overall survival time in the WBRT plus SRS group was superior to that of the RPA cohort across all RPA classes (class 1, 16.1 months versus 7.1 months; class 2, 10.3 months versus 4.3 months; class 3, 8.7 months versus 2.1 months). A cautionary note is that the Sanghavi et al. [61] cohort, by being eligible and selected for SRS after WBRT, may have had an inherently better prognosis than the patients in the historical WBRT cohort, above and beyond what can be adjusted for by RPA class.

In summary, firm data indicate that radiosurgical boost after WBRT improves survival in select patients with single brain metastases. Although the addition of SRS to WBRT improves local control in patients with up to four metastases, it does not affect overall survival in patients with multiple metastases, and it remains speculative whether select patients with multiple metastases and indolent extracranial disease may benefit from SRS boost.

SRS With or Without WBRT

Significant controversy exists over whether a subset of patients, particularly those with a limited number of brain metastases (with the definition of limited varying substantially by institution), can be treated effectively with SRS alone. The putative rationale for withholding WBRT is to spare patients the risk for late neurotoxicity from WBRT. Patients who forgo upfront WBRT are typically monitored closely with serial MRI scans and treated with WBRT or additional SRS at recurrence. To date, one randomized trial has been published examining this question [47], and randomized phase III trials are nearing completion in Europe (European Organization for Research and Treatment of Cancer protocol 22952–26001, surgery or SRS with or without WBRT in patients with one to three brain metastases) [19] and the U.S. (American College of Surgery Oncology Group protocol Z-0300, SRS with or without WBRT in patients with one to three metastases). In addition, six retrospective comparative studies with >100 patients each have been published since 1998 [17, 46, 49, 65–67].

In the first randomized trial published on the subject, Aoyama et al. [47] recently reported results of 132 patients with one to four brain metastases, each <3 cm, randomized to WBRT plus SRS or SRS alone. They found no significant difference in the median overall survival time (7.5 months for WBRT plus SRS versus 8.0 months for SRS alone; p = .42), in accordance with previous retrospective studies. It is of note that the trial was significantly underpowered to address the primary outcome of overall survival [68]. As with virtually every previous study, they confirmed superior combined local and distant intracranial control rates with the addition of WBRT (1-yr actuarial overall control rate, 53.2% versus 23.6%; p < .001). Salvage therapy for CNS progression was required significantly more often in the SRS group (43% versus 15%; p < .001). No significant difference in neurological outcome was observed, although only limited neurocognitive data (Mini-Mental State Examination only) were presented and were not available for the majority of patients in the trial.

The impact of WBRT on neurocognitive function has not been well studied, and the most frequently cited article detailing the detrimental effects of WBRT is a case series published >15 yrs ago in patients who received higher fractions of radiation (300–600 cGy) than are used today [69]. The only randomized data available for SRS with or without WBRT in patients with brain metastases is from the recent Aoyama et al. [47] trial, which failed to show worse neurological outcomes in the WBRT arm. In addition, two randomized trials in SCLC patients receiving PCI have shown no difference in neurocognitive outcome up to 2 years after WBRT [43, 44]. Many have argued that withholding WBRT actually worsens neurologic outcomes because of higher rates of symptomatic intracranial relapse [70].

It is clear that brain metastases themselves result in some degree of neurocognitive dysfunction in many patients, as shown by Chang et al. [7] in a recent pilot study of patients with one to three brain metastases treated with SRS. At baseline, 67% of patients (10 of 15) had impairment on at least one test of neurocognitive function. Similarly, in the phase III trial of WBRT with or without the radiosensitizer motexafin, the most significant predictor of neurologic and neurocognitive decline, as well as deterioration in quality of life, was disease progression in the brain [71].

In summary, there are now randomized controlled data from both surgical [72] and SRS [47] trials that the omission of WBRT in patients with new brain metastases results in significantly worse local control and distant intracranial disease control, though it does not appear to affect overall survival. No studies to date have adequately addressed whether the negative impact of poorer intracranial disease control on quality of life is higher than the risk for delayed neurotoxicity in long-term survivors. The relationship between neurologic and neurocognitive function and quality of life is being increasingly recognized [73], and therefore prospective studies specifically designed to measure these outcomes are still needed.

SRS at Recurrence

The use of SRS for recurrent brain metastases after WBRT has been investigated in several small series and appears to be an effective treatment in patients with good functional status and controlled or indolent extracranial disease. Reported local and overall brain control rates are in the range of 57%–100% and 65%–78%, respectively [45, 74–77]. Acute (seizure, nausea/vomiting, alopecia, headache) and subacute (edema, radionecrosis, hemorrhage) complications are reported in 7%–33% of patients, with large tumor diameter or volume being the most important risk factor for necrosis.

Surgery

Resection of brain metastases was uncommon before the widespread availability of computed tomography imaging [78]. Since the 1980s, resection of most single brain metastases has become a standard, though potentially underused, treatment option in patients with good functional status and controlled or indolent extracranial disease [79]. Benefits of surgical resection include the provision of a definitive diagnosis, rapid relief of neurological symptoms caused by mass effect, and establishment of local control. Advances in surgical technique over the past two decades have led to lower rates of morbidity and in-hospital mortality, with current estimates of in-hospital mortality for resection of brain metastases as low as 1.8% in high-volume centers [79].

Although many retrospective case series in the 1980s suggested a survival benefit from resection of single brain metastases, concern for selection bias remained until a series of three randomized trials in the 1990s was published (Table 3), all addressing whether patients with a single operable brain metastasis benefit from surgery followed by WBRT compared with WBRT alone [64, 80, 81]. Patchell et al. [64] found that patients in the surgery plus WBRT arm had a longer median survival time (9.2 months versus 3.4 months; p < .01), higher local control rate (80% versus 48%; p < .02), longer duration of functional independence, defined as a KPS score ≥70 (38 weeks versus 8 weeks; p < .005), and longer freedom from death resulting from neurological causes compared with patients treated with WBRT and biopsy alone [64]. Three years later, Vecht et al. [81] reported an overall median survival time of 10.0 months in patients undergoing resection and WBRT compared with 6.0 months in patients treated with WBRT alone (p = .04). Similar results could not be replicated by Mintz et al. [80]; however, this study included a patient population with more active extracranial disease and a lower KPS score than in the previous studies. Taken together, the Patchell et al. [64] and Vecht et al. [81] studies indicate that patients with a single brain metastasis and good prognostic features benefit from aggressive local control followed by WBRT.

In patients with multiple metastases, surgery is usually limited to patients with a dominant, symptomatic, or life-threatening lesion and/or those who require a tissue diagnosis before proceeding with therapy. However, two recent single-center retrospective studies suggest that patients with good prognostic features and two to three metastases may gain similar survival benefit from surgery when the dominant lesion is resected [82, 83]. Bindal et al. [84] reported similar results for patients with multiple metastases undergoing resection of all lesions, whose median survival duration was 14 months, again suggesting that a highly selected subset of patients with a limited number of metastases may benefit from aggressive surgical management.

Surgery With or Without WBRT

As with SRS, there is ongoing debate regarding the use of WBRT after surgical resection of single (or multiple) metastases, and the key elements driving the discussion are summarized above. Analogous to the Aoyama et al. [47] study, Patchell et al. [72] found that patients treated with postoperative WBRT had superior local (90% versus 54%; p < .001), distant (86% versus 63%; p < .01), and overall (82% versus 30%; p < .001) intracranial control rates compared with those who underwent surgical resection alone. No overall survival benefit was seen, although patients in the WBRT group were less likely to die from neurologic causes than patients in the observation group (14% versus 44%; p = .003).

Surgery Versus SRS

It is clear that aggressive local management of solitary or oligometastatic brain disease, either with SRS or surgery, combined with WBRT, leads to better outcomes in patients with good prognostic features. No randomized study has directly compared the two, and many patients are eligible for either treatment. Several retrospective comparisons have reported longer survival times after surgical resection [85]; however, most have found essentially no difference [86–88]. The choice must therefore be made on a case-by-case basis, with consideration given to the following factors: for large tumors with extensive edema and mass effect, surgery is probably superior to SRS for quick and reliable relief of symptoms, provided the lesion can be safely resected; because even minor swelling in the posterior fossa can cause hydrocephalus, surgery is often recommended for cerebellar metastases; radiosurgery has the advantage of being noninvasive and can be used to treat surgically inaccessible lesions in the brain stem, basal ganglia, and eloquent cortex.

Chemotherapy

Chemotherapy has traditionally played a limited role in the treatment of brain metastases and has been reserved for patients who have failed other treatment modalities or for diseases known to be “chemosensitive,” such as lymphoma, SCLC, germ-cell tumors, and, to a lesser degree, breast cancer. Skepticism regarding the usefulness of chemotherapy for brain metastases stems primarily from concerns that most agents are either too large or hydrophilic to cross the blood–brain barrier (BBB). The degree to which a given agent is believed to penetrate the BBB is usually based on pharmacokinetic animal and/or human studies comparing plasma with cerebrospinal fluid drug concentrations after i.v. or oral administration. This method may underestimate the concentration of drug delivered to the tumor, however, because brain metastases are known to have local BBB breakdown (demonstrated on MRI by contrast enhancement and peritumoral edema). This is corroborated by studies showing roughly equivalent intracranial and extracranial response rates to chemotherapeutic agents assumed to have little BBB penetration, particularly when first-line agents for the systemic cancer are chosen [89–91]. The success of an agent may therefore rest more heavily upon its inherent activity against the systemic tumor than its putative ability to cross the BBB. Clinical data supporting the utility of chemotherapy for brain metastases in various solid tumors are limited primarily to small phase II studies, often in heavily pretreated patient populations. The text that follows summarizes existing data for chemotherapy in the three most common primary tumors: NSCLC, breast cancer, and melanoma.

NSCLC

Although several chemotherapeutic regimens have modest activity against NSCLC, particularly those that are platinum based, patients with brain metastases from NSCLC tend to be heavily pretreated and therefore have less chance of responding to prescribed second- or third-line agents. When used up-front, cisplatin (CDDP) has activity both as a single agent and in combination in patients with brain metastases from NSCLC. Response rates of 30% for single-agent CDDP [92] and 28%–45% for CDDP in combination with paclitaxel and either vinorelbine or gemcitabine [93], vinorelbine and gemcitabine [89], etoposide [94], fotemustine [95], and teniposide [96] have been reported in chemotherapy-naïve patients treated with up-front chemotherapy. In the majority of these trials, reported response rates and overall survival times were comparable with those in patients with metastatic NSCLC outside the CNS, again suggesting that the responsiveness of brain metastases relies heavily on the chemosensitivity of the primary tumor. More recently, TMZ has demonstrated modest activity in recurrent brain metastases from NSCLC, with response rates in the range of 0%–20% reported in several studies [97–100]. Higher response rates have been reported for TMZ combined with vinorelbine in patients with NSCLC, though at the expense of greater toxicity [101].

Limited data also exist for the responsiveness of brain metastases to the epidermal growth factor receptor (EGFR) inhibitors gefitinib (Iressa®; AstraZeneca Pharmaceuticals, Wilmington, DE) [102–104] and erlotinib (Tarceva®; Genentech, Inc., South San Francisco, CA) [105, 106]. In the largest prospective series reported, Ceresoli et al. [103] found a 27% disease control rate (4 of 41 patients [10%] with PRs, 7 of 41 patients [17%] with stable disease) in 41 patients with measurable brain metastases, 18 of whom had received WBRT >3 months prior to study entry. As with extracranial disease, the response of brain metastases to EGFR inhibitors appears to depend upon the presence of EGFR mutations [107]. In addition to EGFR inhibitors, vascular endothelial growth factor pathway inhibitors such as bevacizumab and small-molecule receptor tyrosine kinase inhibitors such as sorafenib and sunitinib have shown activity in NSCLC and may potentially be useful for brain metastases either alone or in combination; all are entering clinical trials presently.

Breast Cancer

As with NSCLC, combination chemotherapy known to be active in systemic breast cancer has shown comparable response rates with those in extracranial disease when administered in the up-front setting (prior to WBRT) to patients with new brain metastases (Table 4). Objective response rates in the range of 43%–59% have been reported in patients with new brain metastases from breast cancer treated with cyclophosphamide in combination with 5-fluorouracil (5-FU) and prednisone [108], 5-FU, prednisone, methotrexate (MTX), and vincristine [108], and 5-FU and MTX [91]. Rosner et al. [108] also reported a 43% objective response rate in seven patients treated with MTX, vincristine, and prednisone. These regimens may be considered either before or after WBRT in patients with newly metastatic breast cancer to brain who have not yet received cyclophosphamide-based chemotherapy, particularly in the presence of active systemic disease. The combination of cisplatin and etoposide has shown activity in two phase II studies totaling 78 patients with new brain metastases from breast cancer, with 12 CRs, 21 PRs, and a median overall survival time in the range of 31–58 weeks [94, 109]. Also in the up-front setting, Oberhoff et al. [110] reported a 37% objective response rate (one CR, five PRs) in 24 women treated with topotecan (1.5 mg/m² daily for 5 days every 3 weeks) for new brain metastases, with hematologic toxicity being the major side effect.

Table 4

Chemotherapeutic regimens with activity in brain metastases from breast cancer

In the recurrent setting, TMZ is the best studied because of its favorable BBB penetration, ease of administration, and low toxicity profile. Unfortunately, TMZ has shown essentially no activity as a single agent or in combination with vinorelbine for brain metastases from breast cancer [97, 98, 101, 111], demonstrating that BBB penetration alone is not sufficient for response. Modest response rates have been reported for TMZ combined with CDDP (0 of 15 patients with CRs, 6 of 15 patients with PRs) or the oral 5-FU prodrug capecitabine (1 of 24 patients with CRs, 3 of 24 patients with PRs) in small phase II studies, likely reflecting the activity of CDDP or capecitabine rather than TMZ [112, 113]. There are several case reports suggesting that capecitabine has activity in new and recurrent brain metastases from breast cancer [114–117], and larger studies are called for to establish its potential role as a single agent or in combination with other treatment modalities. Likewise, larger studies are needed to better define the role of high-dose MTX for recurrent brain metastases [118]. MTX is active in breast cancer and has good BBB penetration at high doses; however, concern for toxic leukoencephalopathy, particularly when administered after WBRT, limits its use for many patients.

Finally, there is interest in the role of oral targeted agents, particularly in women with human epidermal growth factor receptor (HER)-2–positive tumors, who are known to be at high risk for brain metastases. This risk stems from several factors, including the inherent propensity of HER-2–positive tumors to metastasize viscerally, better systemic disease control and survival in patients treated with trastuzumab (Herceptin®; F. Hoffmann-La Roche, Basel, Switzerland), and poor BBB penetration of trastuzumab, which may provide a “sanctuary site” for brain tumor growth while systemic disease remains controlled [119]. The dual EGFR and HER-2 tyrosine kinase inhibitor lapatinib showed modest activity in a recent phase II study of lapatinib for recurrent brain metastases in women treated with trastuzumab for HER-2–positive breast cancer [120]. Lin et al. [120] reported two PRs, six minor responses, and five patients with stable disease for >16 weeks in 34 patients treated, and a larger multinational phase II trial has recently completed enrollment.

Melanoma

Effective therapy for brain metastases in melanoma remains elusive for all but a select subset of patients who achieve disease stabilization with aggressive local control and WBRT. Internationally, the best-studied agent is fotemustine, a nitrosourea with good BBB penetration that is not available in the U.S., which has shown response rates in the range of 5%–25% as a single agent and in combination with WBRT for newly diagnosed brain metastases [32, 121–123]. In addition, a phase III randomized trial comparing fotemustine with dacarbazine in patients with disseminated malignant melanoma showed a trend toward a longer median time to occurrence of brain metastasis (22.7 months versus 7.2 months; p = .059) in the 182 patients without brain metastases at trial entry [123].

Because of its favorable CNS penetration profile and comparable activity with dacarbazine in metastatic melanoma, TMZ has been the focus of several phase II studies in patients with brain metastases, with modest results. Agarwala et al. [124] conducted a phase II study in 151 radiotherapy-naïve patients with new brain metastases treated with monthly TMZ (150–200 mg/m² × 5 days) and found an objective response rate of 6%. Hoffmann et al. [125] reported a similar response rate of 9% (one CR, two PRs) in 34 patients treated with TMZ alone or in combination with radiotherapy. TMZ in combination with WBRT appears to be safe and well tolerated, but modest results in a small phase II study (objective response rate, 9.7%; progression-free survival duration, 2 months) [126] indicate that phase III randomized data are needed before routine use can be recommended. In a randomized phase II study, the combination of TMZ and thalidomide, chosen for its antiangiogenic properties, resulted in a 25% disease stabilization rate (CR plus PR plus stable disease) [127]; however, the toxicity profile reported in a separate phase II study (intracranial hemorrhage in 29%, thrombosis in 12.5%) raises concerns about the safety of this combination [128].

Future Directions

In the next 5 years, the results of several ongoing multicenter randomized trials will become available to further define the role of various radiation sensitizers and chemotherapeutic agents in combination with SRS, WBRT, or both (Table 5). In addition, a growing number of phase II trials is being initiated to examine the utility of targeted agents in specific disease groups. The safety and efficacy of angiogenesis inhibitors in the treatment of stable and active brain metastases, a topic to this point understudied because of concerns regarding intracranial hemorrhage, will likely be established, as growing evidence in patients with glioblastoma suggests that these agents are relatively safe and carry a low risk for bleeding. An additional benefit of angiogenesis agents may be their ability to control peritumoral edema and reduce steroid dependence, two factors that contribute greatly to morbidity and mortality in patients with brain metastases. As suggested in glioblastoma patients treated with concurrent TMZ and WBRT, the combination of traditional cytotoxic agents or targeted molecular drugs with radiation may hold promise for improving upon control rates of WBRT alone. Finally, given our growing experience with the feasibility and safety of a variety of agents in the treatment of brain metastases, as well as the fact that intracranial responses often track extracranial responses in chemotherapy-naïve patients, more efforts should be made to include patients with brain metastases in disease-specific phase I and II trials for metastatic cancer.

Table 5

Ongoing or recently completed multicenter clinical trials for new or recurrent brain metastases from NSCLC, breast cancer, or melanoma

Abbreviations: ACO-SOG, American College of Surgeons Oncology Group; ECOG, Eastern Cooperative Oncology Group; EORTC, European Organization for Research and Treatment of Cancer; HER-2, human epidermal growth factor receptor 2; JCOG, Japan Clinical Oncology Group; NCCTG, North Central Cancer Treatment Group; NSCLC, non-small cell lung cancer; RTOG, Radiation Therapy Oncology Group; SCLC, small cell lung cancer; SRS, stereotactic radiosurgery; TMZ, temozolomide; WBRT, whole brain radiation therapy.

Conclusion

In summary, significant progress has been made over the past two decades for a subset of patients with single or oligometastatic brain metastases and well-controlled systemic disease. Clear survival benefit has been demonstrated for patients with single metastases and favorable prognostic features who undergo surgery or SRS and WBRT versus WBRT alone. Unfortunately, these patients represent a small minority of the estimated 200,000 patients with brain metastases yearly in the U.S. For these patients, treatment strategies that prolong survival remain lacking; however, for this population intracranial disease control, time to neurologic progression, neurologic function, and quality of life may be more relevant endpoints because of the competing risk for death from systemic disease. An algorithm for approaching treatment decisions in patients with brain metastases is presented in Figure 1. For the majority of patients, most of whom have multiple metastases, WBRT remains the standard of care. Interest in the area of brain metastases is burgeoning, however, and in the future radiosensitizers, chemotherapy, and targeted agents will likely play increasingly important roles in the treatment of recurrent brain metastases. Also of interest are chemoprevention strategies for patients at high risk for CNS failure and methods for enhancing CNS drug delivery through BBB modulation or manipulation of multidrug resistance pathways in the brain. In the future, patients will benefit from a multidisciplinary approach focused on the integration of surgical, radiation, and chemotherapeutic options with the goal of preserving neurologic and neurocognitive function and quality of life.

Figure 1

Algorithm for the initial treatment of brain metastases. ^aOmission of up-front MBRT is an alternative in patients who are followed closely for progression after surgery or SRS. ^bConsider for patients with one dominant lesion causing mass effect, or pathologic diagnosis required.

Abbreviations: KPS, Karnofsky performance status; SRS, stereotactic radiosurgery; WBRT, whole brain radiation therapy.

Acknowledgments

Dr. Eichler is the Richard B. Simches Scholar in Neuro-Oncology at the Massachusetts General Hospital.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Author notes

Disclosure of potential conflicts of interest is found at the end of this article.