INTRODUCTION — Melanoma is the third most common cancer causing brain metastases in the United States, after cancers of the lung and breast, which appears to reflect the relative propensity of melanoma to metastasize to the central nervous system (CNS). Brain metastases are responsible for 20 to 54 percent of deaths in patients with melanoma, and among those with documented brain metastases, these lesions contribute to death in up to 95 percent of cases.

As with other primary tumors, patients with melanoma metastatic to brain typically present with symptoms of increased intracranial pressure (eg, headache), focal neurologic deficits, and/or seizures. Brain metastases from melanoma may have a particularly high propensity for spontaneous hemorrhage.

RISK FACTORS — Although all patients with melanoma are at risk for metastasis to the brain, certain characteristics that are associated with increased risk of systemic metastases also correlate with the subsequent development of brain metastasis. These factors include:

Male gender
Melanomas arising on mucosal surfaces or the skin of the trunk, head, or neck
Wide, thick, deeply invasive, or ulcerated primary lesions
Acral lentiginous or nodular lesions on histologic examination
Involvement of more than three regional lymph nodes, either at diagnosis or relapse
Visceral metastasis at the time of diagnosis, especially if the visceral metastasis is disseminated to more than one organ

PROGNOSIS — Melanoma patients with brain metastases have a poor prognosis. In two series totaling almost 1400 patients, the median survival was four months, and one year survival rates were 9 and 19 percent, respectively

Despite the general grim prognosis, occasional patients do well. Favorable prognostic signs include the presence of a single brain metastasis without other visceral metastatic disease, and an initial presentation with a metastasis to the brain. In contrast, multiple brain lesions, extensive visceral metastases, or a primary lesion of the head and neck region carry an unfavorable prognosis.

In patients with brain metastases from other primary tumors, good performance status and limited intra- and extracranial disease are associated with a favorable prognosis. The Radiation Therapy Oncology Group (RTOG) used recursive partitioning analysis (RPA) to analyze over 1200 patients with brain metastases and develop a prognostic index to predict the outcome following palliative whole brain radiation therapy (WBRT). When RPA was applied to 74 patients with brain metastases from melanoma, the median survival durations for patients in RPA classes I, II, and III were 10.5, 5.9, and 1.8 months, respectively.

FAVORABLE PROGNOSIS PATIENTS — In favorable prognosis patients (ie, limited or no extracranial disease, a good performance status, and a single or limited number of brain metastases), aggressive treatment to eradicate metastases in the brain is associated with an improved outcome. Surgery traditionally was used to treat patients with a single or limited number of lesions, often supplemented with WBRT. Subsequent advances have made stereotactic radiosurgery (SRS) an alternative, particularly when lesions are not surgically accessible or when multiple lesions are present.

Surgery — Surgical resection appears to be more effective than WBRT in favorable prognosis patients with melanoma metastatic to brain. Most data supporting the use of surgery in this setting were generated from patients with a single lesion. However, patients with multiple surgically-accessible lesions and little or no extracranial disease may also have an improved prognosis when treated with surgical resection, rather than WBRT.

Randomized trials comparing surgery plus WBRT to WBRT alone in patients with brain metastases from diverse primary tumors support a survival benefit associated with the use of surgery. The results of these trials are discussed elsewhere.

In a retrospective analysis of 686 patients treated for brain metastases between 1985 and 2000 at the Sydney Melanoma Unit, the median survival for those undergoing surgery with or without RT (8.9 and 8.7 months, respectively) was significantly greater than that of patients treated with RT alone or supportive care (3.4 and 2.1 months, respectively). These results suggest that there is a benefit for surgery as a component of the initial treatment, although much of the observed difference may reflect a patient selection bias.

In patients who are not candidates for surgical resection because of the size or location of the lesion, biopsy is indicated when the diagnosis is in doubt.

Stereotactic radiosurgery — SRS utilizes multiple convergent radiation beams to deliver a high single dose to a radiographically discrete treatment volume.

There is substantial experience with SRS in patients with melanoma brain metastases suggesting that outcomes are similar to those with other malignancies metastatic to the brain. Freedom from progression is achieved in approximately 90 to 95 percent of lesions, with disappearance or a reduction in size occurring in about 55 percent. Local control is primarily influenced by the dose and lesion size, while an increasing number of lesions predict a higher likelihood of relapse at other sites in the brain

The effectiveness of SRS compared to conventional surgery remains uncertain. In general, our first choice is surgery if a single metastatic lesion is causing neurologic symptoms and is easily accessible. SRS is preferred for lesions in eloquent or relatively inaccessible areas of the brain and in patients who have previously received WBRT and have new, progressive, or symptomatic lesions.

Adjuvant WBRT — Both surgical resection and SRS are generally followed by postoperative radiation]. The goal of WBRT is to treat any residual cancer cells at the site of resection and elsewhere within the brain.

When WBRT is used as an adjuvant, there is the potential for delayed complications. This risk must be weighed against the potential benefits and long-term prognosis. The late effects of WBRT are discussed elsewhere.

After surgery — No randomized trials have been conducted in melanoma patients to assess the effectiveness of postoperative WBRT, but at least one such trial in patients with diverse tumor types supports the use of this approach. In patients with solitary brain metastases from all primaries, a randomized controlled trial indicated that patients receiving WBRT were significantly less likely to fail in the brain, although there was no difference in overall survival or the duration of functional independence.

Although extrapolation of such results to radioresistant tumor types may not be appropriate, observational studies in melanoma patients who had a resection of a solitary brain metastasis suggest that WBRT improves the outcome In one study of 35 patients, the likelihood of dying from neurologic causes was much lower in the 19 patients who received radiation group (24 versus 85 percent, compared to those who were treated with surgery alone), although overall survival was similar and depended upon control of systemic disease

After SRS — We generally recommend the administration of WBRT after SRS in patients with brain metastases from a variety of tumor types. The data from randomized trials supporting this approach are derived from studies in patients with multiple tumor types, and are discussed elsewhere.

In patients with metastatic melanoma, data supporting WBRT after SRS is limited. Randomized trials have not been performed in patients with melanoma, but an observational study suggested the possible efficacy of adjuvant WBRT after SRS in this patient population. In this study, SRS was performed on 118 lesions in 60 patients with melanoma, 51 of whom also received WBRT. Although WBRT combined with SRS did not improve either survival or local tumor control, there was a nonsignificant reduction in new brain metastases (23 versus 44 percent) in patients given WBRT. Consequently, decisions about adjuvant WBRT after SRS in patients with melanoma need to be individualized, taking into account the size and number of CNS lesions and the status of the systemic disease.

POOR PROGNOSIS PATIENTS — Poor prognosis patients are generally treated with WBRT rather than surgery or SRS. However, surgery is occasionally performed to resect a large symptomatic or life-threatening lesion.

Radiation therapy — WBRT is widely used as the treatment for poor prognosis patients with melanoma metastatic to the brain. Even with treatment, the prognosis for these patients is poor. In the Sydney Melanoma Unit study, 234 patients treated with WBRT alone had a median survival of 3.4 months, compared to 2.1 months in 210 who received supportive care only.

Chemotherapy — Melanoma is relatively resistant to chemotherapy, and systemic treatment for parenchymal brain metastases has generally been ineffective due to drug resistance, inadequate drug delivery to the brain, and/or treatment-related toxicity. The use of chemotherapy for the treatment of brain metastases is discussed elsewhere.

Radiation therapy has been combined with various systemic agents (fotemustine, temozolomide, thalidomide). These combinations of RT plus systemic therapy have produced CNS response rates of <10 percent and often increased toxicity. These data call into question not only the value of these chemoradiation combinations, but also the value of WBRT itself in the treatment of CNS metastases from melanoma

Patients with leptomeningeal metastases from melanoma have a poor prognosis, with a median survival less than two months. Intrathecal chemotherapy may delay the progression of neurologic signs and symptoms, but benefits are limited.

SUMMARY AND RECOMMENDATIONS — Brain metastases are a frequent complication in patients with metastatic melanoma and when present are an important cause of both morbidity and mortality. In general, the approach to patients with melanoma metastatic to brain is similar to that for patients with other types of tumors.

Favorable prognosis patients — Those patients with a relatively favorable prognosis (RPA class I or II,  have a single or limited number of brain metastases, good performance status, and limited or no extracranial disease. In this setting, aggressive treatment directed against the brain metastases is warranted:

We recommend surgical resection for patients with a single or limited number of surgically accessible metastases

We recommend stereotactic radiosurgery (SRS) for patients with lesions in surgically inaccessible locations and those who are not surgical candidates or refuse surgery

We suggest whole brain radiation therapy (WBRT) following surgery or SRS to reduce the incidence

Poor prognosis patients — Patients with extensive extracranial disease and/or a poor performance status have a very poor prognosis (RPA class II or III,  and are less likely to benefit from aggressive treatment directed against their brain metastases. In this setting we suggest WBRT or best supportive care