Management of Brain Metastases
Management of Brain Metastases
Brain metastases are the most common type of brain tumor in adults and are an increasingly important cause of morbidity and mortality in cancer patients. In recent years, important advances have been made in the diagnosis and management of brain metastases. These advances include the widespread use of magnetic resonance imaging (MRI), enabling small metastases to be detected; the introduction of stereotactic radiosurgery; and the performance of studies that have clarified the role of surgery and postoperative radiation therapy for single brain metastases. As a result, most patients receive effective palliation, and the majority do not die from their brain metastases. However, further studies are needed to define the optimal role of conventional treatments and to develop more effective novel therapies. [ONCOLOGY 13(7):941-961, 1999]
Brain metastases are a common complication in cancer patients and an important cause of morbidity and mortality. They develop in approximately 10% to 30% of adults and 6% to 10% of children with cancer.[1-6] Each year in the United States, an estimated 97,800 to 170,000 new cases of brain metastasis are diagnosed.[1,2,6] This number may be increasing as a result of the increased ability of magnetic resonance imaging (MRI) to detect small metastases and improvements in systemic therapy, leading to longer patient survival.[1,6-9]
In adults, the primary tumors most often responsible for brain metastases are lung cancer (50%), breast cancer (15% to 20%), unknown primary tumor (10% to 15%), melanoma (10%), and colon cancer (5%).[1-3,10] In children, the most common sources of brain metastases are sarcomas, neuroblastoma, and germ cell tumors.[1,4,11]
Studies using MRI suggest that the proportion of single metastases is lower than was previously believed, accounting for only one-third to one-fourth of patients with cerebral metastases.[7,12] Metastases from breast, colon, and renal cell carcinomas are often single, while melanoma and lung cancer have a greater tendency to produce multiple metastases.[1,13]
The most common mechanism of metastasis to the brain is by hematogenous spread. These metastases are usually located directly beneath the junction of the gray and white matter. Brain metastases tend to occur at this site because the blood vessels decrease in size at this point and act as a trap for clumps of tumor cells. Brain metastases also tend to be more common at the terminal “watershed areas” of arterial circulation.[1,13]
The distribution of brain metastases roughly follows the relative weight of (and blood flow to) each area. Approximately 80% of brain metastases are located in the cerebral hemispheres, 15% in the cerebellum, and 5% in the brainstem. For unclear reasons, pelvic (prostate and uterus) and gastrointestinal tumors have a predilection to metastasize to the posterior fossa.
It is estimated that more than two-thirds of patients with cerebral metastases experience neurologic symptoms during the course of their illness. The clinical features of brain metastases are extremely variable, and the presence of brain metastases should be suspected in any cancer patient who develops new neurologic symptoms.
The majority of patients with brain metastases present with progressive neurologic dysfunction resulting from a gradually expanding tumor mass and the associated edema, or, rarely, from the development of obstructive hydrocephalus. Approximately 10% to 20% of patients present acutely with seizures, while another 5% to 10% present acutely as a result of strokes caused by embolization of tumor cells, invasion or compression of an artery by tumor, or hemorrhage into a metastasis.[8,14,15] Melanoma, choriocarcinoma, and thyroid and renal carcinomas have a particular propensity to bleed.
The clinical presentation of brain metastases is similar to that of other brain tumors and includes headaches, focal neurologic dysfunction, cognitive dysfunction, and seizures. Headaches occur in approximately 40% to 50% of patients with brain metastases. These are usually dull, nonthrobbing, and often indistinguishable from tension headaches. The headaches are usually on the same side as the tumor, although they can be diffuse. Headaches characteristic of increased intracranial pressure, such as early morning headaches, or headaches exacerbated by coughing, bending, and straining, are present in less than half of patients with brain metastases. The headaches may be associated with nausea, vomiting, and transient visual obscurations. Patients with multiple metastases and posterior fossa metastases have a higher frequency of headaches. (Papilledema is observed in fewer than 10% of patients at the time of presentation.)
Focal neurologic dysfunction is the presenting symptom in 20% to 40% of patients. Hemiparesis is the most common complaint, but the precise symptom varies depending on the location of the metastases. Cognitive dysfunction, including memory problems and mood or personality changes, are the presenting symptoms in one-third of patients, while seizures are the presenting symptom in another 10% to 20%.[17-20]
Brain metastases must be distinguished from primary brain tumors, abscesses, demyelination, cerebral infarctions or hemorrhages, progressive multifocal leukoencephalopathy, and the effects of treatment, including radiation necrosis. In a study by Patchell et al, 11% of patients who were initially felt to have a single brain metastasis eventually were found to have a different diagnosis after the lesion was biopsied. Half of the nonmetastatic lesions were primary brain tumors, while the other half were infections. The false-positive rate for diagnosis of multiple metastases undoubtedly is significantly lower than the 11% rate for single metastases. Nonetheless, in any patient in whom the diagnosis of brain metastases is in doubt, a biopsy should be performed since this is the only reliable method of establishing the diagnosis.
Breast cancer patients with a single dural-based lesion pose a particular diagnostic dilemma. Since the incidence of meningiomas is increased in patients with breast cancer, it is important to differentiate a dural-based metastasis from a meningioma.[22,23] Frequently, imaging studies are inconclusive, and a biopsy or surgical resection of the lesion is needed.
In addition to diagnosing brain metastases, it is also important to differentiate patients with a single or solitary metastasis from those with multiple brain metastases since their subsequent treatment differs. The term “single brain metastasis” refers to a single cerebral lesion, with no implication made regarding the extent of extracranial disease. “Solitary brain metastasis” describes the relatively rare occurrence of a single brain metastasis that is the only known site of metastatic cancer in the body.
Although computed tomographic (CT) scans detect the majority of brain metastases, the best diagnostic test for brain metastases is contrast-enhanced MRI. [12,24,25] This test is more sensitive than enhanced CT scanning or nonenhanced MRI in detecting lesions in patients suspected of having cerebral metastases, and in differentiating these metastases from other central nervous system (CNS) lesions.[ 24,25] Radiographic features that help differentiate brain metastases from other CNS lesions include the presence of multiple lesions (which helps distinguish metastases from gliomas or other primary tumors), localization of the lesion at the gray-white matter junction, more circumscribed margins, and relatively large amounts of vasogenic edema compared to the size of the lesion.
In the majority (80%) of patients, brain metastases develop after the diagnosis of systemic cancer (metachronous presentation).[1,2] However, in some patients, brain metastases may be diagnosed before the primary tumor is found (precocious presentation) or at the same time as the primary is detected (synchronous presentation).
For patients who present with brain metastases without a known primary tumor, the lung should be the focus of the evaluation. Over 60% of these patients will have a lung primary or pulmonary metastases from a primary tumor located elsewhere.[1,26,27] If the chest radiograph is nondiagnostic, a chest CT scan should be performed, as this significantly increases the likelihood of detecting a lung tumor. These patients also should have a CT scan of the abdomen and pelvis and a bone scan to determine the extent of metastatic disease. Breast cancer is an uncommon cause of brain metastases without a known primary tumor, possibly due to its earlier detection on physical examination, and its tendency to produce brain metastases in the setting of widely disseminated disease. 
The management of patients with brain metastases can be divided into symptomatic and definitive therapy. Symptomatic therapy includes the use of corticosteroids for the treatment of peritumoral edema, anticonvulsants for control of seizures, and anticoagulants or inferior vena cava filters for the management of venous thromboembolic disease. Definitive therapy includes treatments directed at eradicating the tumor itself, such as surgery, radiotherapy, and chemotherapy.
Corticosteroids were first used for treating peritumoral edema by Kofman et al in 1957 in patients with breast cancer. Galicich et al introduced the use of dexamethasone in 1961, and this has remained the standard treatment for peritumoral edema ever since. Corticosteroids produce their antiedema effect by reducing the permeability of tumor capillaries,  and are indicated in any patient with symptomatic edema.
Most patients are started on dexamethasone, which, compared with other corticosteroids, has relatively little mineralocorticoid activity, thus reducing the potential for fluid retention. In addition, dexamethasone may be associated with a lower risk of infection and cognitive impairment.
Dexamethasone therapy is usually started as a 10-mg loading dose, followed by 4 mg four times a day; however, there is some evidence that lower doses may be as effective. Although most patients improve symptomatically within 24 to 72 hours, neuroimaging studies may not show a decrease in the amount of edema for up to 1 week. In general, headaches tend to respond better than do focal deficits. If 16 mg of dexamethasone is insufficient, the dose may be increased up to 100 mg/d. Steroid dose is usually tapered following irradiation, although the tapering process may begin earlier in patients with minimal peritumoral edema.
Adverse Effects—Despite their usefulness, corticosteroids are associated with a large number of well-known side effects, including myopathy, weight gain, fluid retention, hyperglycemia, insomnia, gastritis, acne, and immunosuppression. The frequency of these complications can be reduced by using the lowest possible dose.
There is increasing evidence that brain tumor patients who receive corticosteroids are at increased risk of developing Pneumocystis carinii pneumonia. This complication can be prevented by treating patients who are on prolonged courses of a corticosteroid, especially those over the age of 50 years, with trimethoprim/sulfamethoxazole prophylaxis.
As mentioned previously, seizures are the presenting symptom in approximately 10% to 20% of patients with brain metastases, and occur at some stage of the illness in another 10% to 20% of patients.[17-20] Patients with brain metastases who present with seizures should be treated with standard anticonvulsants. In order to minimize toxicity, the lowest effective anticonvulsant dose should be used and polytherapy should be avoided whenever possible. Electroencephalography may be useful if the diagnosis of seizures is in doubt but is not routinely needed for patients who give a clear history of seizures or, conversely, do not have symptoms suggestive of seizures.
Adverse Effects and Drug Interactions—In addition to the usual complications of anticonvulsants, brain tumor patients experience an increased incidence of particular side effects, especially drug rashes. Approximately 20% of brain tumor patients treated with phenytoin and undergoing cranial irradiation develop a morbilliform rash and a small percentage develop Stevens-Johnson syndrome.[35,36] Stevens-Johnson syndrome also has been described in brain tumor patients receiving carbamazepine, while patients receiving phenobarbital have an increased incidence of shoulder-hand syndrome.
In addition to producing adverse effects, anticonvulsants also have clinically significant interactions with other drugs commonly used in patients with brain metastases. Phenytoin induces the hepatic metabolism of dexamethasone and significantly reduces its half-life and bioavailability. Conversely, dexamethasone may also reduce phenytoin levels.
A number of chemotherapeutic agents commonly used in cancer patients interact with phenytoin, causing serum drug levels to fall and potentially leading to breakthrough seizures. Also, hepatic enzyme–inducing anticonvulsants, such as phenobarbital and phenytoin, may interfere with chemotherapeutic agents, such as paclitaxel (Taxol).
Role in Patients With Supratentorial Metastases—Because the risk of seizures in patients with infratentorial metastases is very low, anticonvulsant therapy usually is not indicated. The role of anticonconvulsant therapy in patients with supratentorial brain metastases who have not had a seizure is controversial.
Cohen et al retrospectively reviewed 160 patients with brain metastases who had not suffered a seizure. They found that patients receiving prophylactic phenytoin had the same frequency of late seizures (10%) as did patients receiving no antiseizure prophylaxis.
Glantz et al conducted a prospective, placebo-controlled, randomized study evaluating the efficacy of valproic acid in protecting 74 patients with newly diagnosed brain metastases from seizures. There was no significant difference in the incidence of seizures between patients receiving valproic acid (35%) or placebo (24%), suggesting that prophylactic anticonvulsants were not effective in these patients.
Weaver et al conducted a prospective, randomized study of prophylactic anticonvulsants in 100 brain tumor patients who had not had seizures, including 60 with metastases. Overall, 26% of patients had seizures during the study. There was no difference in the seizure rate between patients who did and did not receive anticonvulsants.
Recently, Glantz et al performed a meta-analysis of the randomized clinical trials addressing this issue. They concluded that there is no statistical evidence showing a significant benefit of prophylactic anticonvulsants.
Recommendations—Because of the increased incidence of allergic reactions in patients with brain metastases receiving anticonvulsant therapy, and the lack of clear evidence that anticonvulsant therapy reduces the incidence of seizures, routine anticonvulsant therapy is probably unnecessary in patients with brain metastases who have not experienced a seizure. Possible exceptions to this are patients with brain metastases in areas of high epileptogenicity (eg, the motor cortex), patients with multiple metastases from melanoma,  and patients with both brain metastases and leptomeningeal metastases. These patients have a higher incidence of seizures and may benefit from prophylactic anticonvulsant therapy.
Treatment of Venous Thromboembolic Disease
Venous thromboembolic disease is common in patients with brain metastases, occurring in approximately 20% of patients. The optimal therapy is unknown. These patients are often perceived to be at increased risk of intracranial hemorrhage when treated with anticoagulants because of the vascularity of the tumors and anecdotal case reports of hemorrhage. As a result, the majority of brain metastases patients with venous thromboembolic disease are managed with inferior vena cava filtration devices rather than anticoagulation. However, Levin et al found that complications occur in up to 60% of brain tumor patients with venous thromboembolic disease who are treated with inferior vena cava filters.
Moreover, several retrospective studies have suggested that the risk of intracranial hemorrhage may not be significantly increased in patients with primary brain tumors who are anticoagulated after the immediate postoperative period. More recently, Schiff and DeAngelis reviewed the Memorial Sloan-Kettering experience with anticoagulation in patients with brain metastases who developed venous thromboembolic disease. Of the 42 patients who received anticoagulation at some stage of their treatment, only 3 (7%) experienced cerebral hemorrhage, 2 in the setting of overanticoagulation.
These studies suggest that anticoagulation may be more effective than inferior vena cava filter placement, and is acceptably safe when the prothrombin time is maintained within the normal range, especially in patients with brain metastases that generally do not hemorrhage, such as breast cancer.