Clinical aspects

Description

A team approach encompassing all treatment modalities is beneficial for optimal diagnosis and treatment of brain metastases (69). Specifically, the options of surgical resection, WBRT, and stereotactic radiosurgery must be readily available, with well-established referral channels established if necessary. At a minimum, the team should include members from neurosurgery, radiation oncology, and medical oncology, as well as the patient’s primary physician. In addition, tertiary care centers can involve members from the neurooncology, neurology, neuroradiology, neuropsychology, rehabilitation, and internal medicine teams.

The evaluation of a patient with suspected brain metastases includes a complete history, physical and neurologic examination, Karnofsky Performance Score determination, and a review of the patient's primary tumor history, if applicable. Imaging should include a gadolinium-enhanced MRI scan of the brain. Even in the absence of neurologic deficits, routine brain imaging in patients with newly diagnosed lung cancer will identify brain metastases in 3% to 15% of patients (43). With the exception of lung cancer and melanoma, routine brain imaging in patients with cancer and no neurologic deficits may not be warranted.

If a metastatic lesion is suspected based on the clinical evaluation and MRI scan, and if the primary site is unknown, a limited workup for the primary tumor is typically appropriate. The lung is the primary site in 70% of cases in which brain metastases are the initial presentation of an extracranial primary tumor; therefore, a chest radiograph and a chest CT scan are the most productive diagnostic tests (108). Although a CT scan of the abdomen may reveal the occasional renal or gastrointestinal primary tumor, further radiologic studies are rarely diagnostic.

Initial treatment in patients with symptomatic cerebral edema includes corticosteroids (often dexamethasone 4 mg by mouth every 6 or 8 hours, plus appropriate medication to prevent steroid gastrointestinal side effects). Although approximately 20% of patients with brain metastases present with seizures (Nguyen and De Angelis 2004), the use of prophylactic anticonvulsants may not significantly reduce the risk of a first seizure (33; 72; 59).

It is essential for appropriate surgical decision-making to have the medical oncologist's best estimate of the patient's functional survival based on the primary tumor and the extent of noncentral nervous system disease. With rare exceptions, surgical resection or stereotactic radiosurgery is not indicated if the "systemic" survival is less than 4 to 6 months. If age or general medical status indicate a risk for general anesthesia, evaluation by a neuroanesthesiologist is helpful to clarify that risk. In some cases, surgical resection can be accomplished under local anesthesia with intravenous sedation, reducing the medical risk of the surgery.

The choice of stereotactic radiosurgery versus surgical resection is rarely difficult when lesion characteristics such as size, location, and number, and patient characteristics such as neurologic status, medical status, and age are all considered. In some situations, a combination of surgical resection (eg, in the case of a large, symptomatic lesion) or stereotactic cyst aspiration (to reduce the size of cystic metastases) plus stereotactic radiosurgery is optimal. If, after considering all factors, it would be equally reasonable to proceed with either surgical resection or stereotactic radiosurgery, a thorough discussion with the patient of the risks and benefits of each option will ensure a truly informed decision.

A tissue diagnosis is usually indicated prior to treatment. One reason for this is that up to 10% of patients with a strongly suspected metastatic disease in fact have nonmetastatic pathology (86; 01). Molecular subtypes of brain metastases may also predict response to radiotherapy (73). A clearer understanding of the differences in molecular mutational phenotypes between brain metastases and their paired primary tumors is evolving (17). Over the past decade, significant advances have been made with regard to the understanding of the molecular biology underlying many brain metastases, and this has led to an increase in the availability of specific molecular targeted therapies. More specifically, erlotinib, gefitinib, afatinib, and osimertinib have demonstrated efficacy in EGFR-mutated nonsmall cell lung cancer (75; 117; 91; 101), trastuzumab for HER2-positive breast cancer (42), and dabrafenib and vemurafenib in BRAF-mutated melanoma (03). In addition, crizotinib, alectinib, ceritinib, and bagratinib have demonstrated benefit in EML4-ALK or ROS1 gene fusion (27). Ceritinib, which is a more potent ALK inhibitor than crizotinib, has been the subject of much investigation as ALK-rearranged non-small cell lung cancers invariably develop resistance to crizotinib. A phase I trial demonstrated a clinical response to ceritinib in patients with crizotinib-resistant disease (55). The extent of PDL1 expression can also influence decisions regarding the utilization of immune checkpoint inhibitors. The availability of these targeted therapies reinforces the importance of tissue diagnosis in the setting of newly diagnosed brain metastases. It is important to note that a stereotactic CT- or MRI-guided biopsy has a 1% to 7% chance of major morbidity (eg, permanent neurologic deficit) or mortality (58; 62).

For surgical resection, the operating room team must be proficient in the surgical approaches to brain lesions, image-guided neuronavigation, intraoperative neuromonitoring, and neuroanesthesia for cerebroprotection. Techniques and equipment important for surgical resection include volumetric CT and MRI image guided neuronavigation, intraoperative ultrasound, intraoperative electrophysiological monitoring and mapping (evoked potentials and direct brain stimulation), and an ultrasonic aspirator. The role of intraoperative MRI in the surgical resection of brain metastases is probably much less than that for optimizing the resection of invasive primary brain tumors, such as malignant gliomas (93; 81; 106).

In the setting of both surgical resection and stereotactic radiosurgery, there may be a transient worsening of neurologic status (usually an exacerbation of a preexisting focal neurologic deficit) due to treatment-related cerebral edema at the site of resection. A short course of corticosteroids (1 week or less in most cases of surgical resection, but often longer with stereotactic radiosurgery) may be helpful. If surgical resection is halted before significant changes occur in the monitoring parameters (ie, before evoked potential amplitude drops by greater than 50%), then a permanent neurologic deficit is unlikely to result, despite a transient worsening for up to a week that parallels the time course of postoperative cerebral edema.

In addition to prolong survival, the goal of surgery for brain metastases is to maintain or improve functional status for as long as possible. In a series of 218 brain metastases in 172 patients treated with Gamma Knife stereotactic radiosurgery (median minimum tumor dose 18.5 Gy), the local tumor control rate was 87%; tumor hemorrhage or edema requiring corticosteroid treatment occurred in 3%, and the overall persistent complication rate was 6% (Hasegawa et al 2003). Radiosurgery in patients with multiple brain metastases who are not candidates for surgical resection results in a mean survival of approximately 9 months, dependent more on the extent of extracranial disease rather than the brain metastases (05; 89; Hasegawa et al 2003; 53; 36).

With surgical resection, local recurrence is common if complete resection of the metastatic lesions is not achieved. To select patients most likely to benefit from surgical resection of recurrent metastatic lesions, a grading system was developed (I-IV) that accounts for the presence of systemic disease, preoperative KPS, time to recurrence, age, and histology of the primary tumor (13). The overall median survival of patients was 11.5 months after re-operation and was inversely related to grade: grade IV (n = 6) 3.4 months, grade III (n = 14) 6.8 months, and grade II (n = 19) 13.4 months. The median survival of grade I patients (n = 9) had not been reached by the conclusion of the study.

Given the current state of healthcare economics, it is important to consider the cost differential for surgical resection versus stereotactic radiosurgery. Not surprisingly, the total cost of surgical resection has been found to be 1.3 to 2.6 times the cost of stereotactic radiosurgery (92; 99; 20). However, as minimally invasive surgical techniques reduce the hospital stay for many patients undergoing surgical resection, the cost discrepancy between these treatment strategies continues to decrease. Importantly, an analysis of an administrative claims database shed light on the economic burden associated with the development of brain metastases (39). After comparing expenses before and after brain metastasis diagnosis, the authors reported an increased cost of $25,579 per-patient-per-6-months (PPP6M) in the post-diagnosis period.

The criteria for selecting between surgical resection and stereotactic radiosurgery in patients with brain metastases are continually evolving. Nonetheless, regardless of which treatment modality is selected, there is strong evidence that early and aggressive treatment is critical in order to achieve the primary goal of maximum functional neurologic status for the longest duration of time (22; 29; 53; 34). A retrospective study of 720 patients who underwent Gamma Knife radiosurgery for non-small cell lung cancer demonstrated the importance of early stereotactic radiosurgery treatment and repeat stereotactic radiosurgery in the setting of tumor recurrence (15).

With regard to follow up imaging after initial treatment of brain metastases, there is evidence to suggest that routine MRI surveillance is superior to symptom-based reimaging in patients treated initially with stereotactic radiosurgery (65). Nevertheless, the specific frequency and duration of follow up have not been well studied, nor has the optimal follow-up regimen in patients treated initially with surgical resection.

Indications

Quality of life and functional status must be taken into account when considering offering surgery to a patient with brain metastases. Although no clearcut guidelines exist, one must consult with the patient and family, as well as with the patient's oncologist and internist. Factors important to the decision of whether to offer open excision or stereotactic radiosurgery include: (1) the patient's life expectancy based on the systemic disease, (2) the patient's medical condition (risk of general anesthesia and postoperative recovery), (3) likelihood of success (disease control and resolution of neurologic deficits), and (4) the patient's wishes after given all available information and options. A Karnofsky Performance Score (KPS) greater than 70 is often considered favorable for surgery (54), although some patients with a KPS greater than 70 may be better suited for nonsurgical management, and conversely, some patients with a lower KPS score may benefit from surgery (05; 63). At least 1 study has demonstrated the feasibility of surgery in carefully selected elderly patients (up to 80 years of age) (38). Risk factors for early death (less than 6 months) after resection of brain metastases include a decrease in KPS after surgery, lack of postoperative systemic therapy, and the presence of uncontrolled extracranial malignancies (09). Although surgery has been recommended with a life expectancy as short as 2 months (86), in most cases, an estimated life expectancy of 4 to 6 months is preferable before proceeding with surgical resection or stereotactic radiosurgery.

Although randomized prospective studies have demonstrated superior outcomes in patients who undergo surgical resection in combination with radiotherapy for single brain metastases (86; 80), considerable evidence can be found to support either surgical resection (14; 05) or stereotactic radiosurgery (16; 89; Hasegawa et al 2003; 53; 74) for patients with brain metastases. Multiple studies have demonstrated comparable survival in patients with single brain metastases as compared with patients with multiple lesions (16; 53; 95). In 1 retrospective study, mean survival time was longer in patients who underwent complete resection of brain metastases, irrespective of the number of lesions, as compared with survival time in those patients in whom 1 or more lesions were left unresected (14 vs. 6 months) (14). Metastasis to the brainstem frequently precludes surgical resection; however, local control may still be achieved in such cases with stereotactic radiosurgery (60).

Surgical resection is indicated for lesions with symptomatic mass effect. In addition, several studies have documented the benefit of surgical resection of recurrent tumor after stereotactic radiosurgery or WBRT (13; 52; 02). Out of 2884 patients with metastatic brain tumors treated with stereotactic radiosurgery at 1 center over 14 years, 58 patients underwent resection of a recurrent or progressive lesion after stereotactic radiosurgery (52). Median survival time (in months) after resection was: 9.4, 7.6, and 3.1 in RPA class I, II, and III, respectively. At least 1 study has looked at the specific technique employed for surgical resection as a predictor of outcome (87) but so far no meaningful differences have been identified between surgical techniques.

Fractionated stereotactic radiosurgery, which seeks to maximize local control and minimize radiation-induced toxicity, is an alternative to surgical resection and single-fraction SRS. A study of 135 patients with 171 brain metastases demonstrated overall local control rates of 88% and 72% at 1 and 2 years, respectively, with 7% of patients developing radiographic evidence of radiation necrosis. Local control rates varied depending on the histology of the lesion, with breast carcinoma having the best response at 92%, followed by non-small cell lung cancer at 88%, and melanoma at 68% (74).

Given the frequency of brain metastases from lung primary and the poor prognosis of patients with lung cancer overall (little more than 10% with 5-year survival), it is important to identify subgroups of patients with a more favorable prognosis. The value of surgical treatment of both the lung primary and the brain metastasis (where feasible), with 5-year survival rates approaching 3 times the overall lung cancer survival rate, was first noted in the late 1980s (06) and confirmed in a 2010 study (67). For lung cancer patients with brain metastases, evaluation of the lung primary by an experienced thoracic surgeon is crucial to optimize functional survival given that progression of extracranial disease is more often the cause of death than progression of intracranial disease (06; 05; 67; 74).

Contraindications

The major contraindications to surgery are a life expectancy of less than 4 to 6 months and a KPS of less than 70 (without expectation of significant improvement after surgery). For surgical resection, the patient's medical condition (ie, risk of undergoing general anesthesia) must be considered. That being said, the increasing use of minimally invasive stereotactic and image-guided techniques have allowed for shorter anesthesia times and lower complication rates, thereby making surgical resection a viable option even for patients with significant comorbidities.

The size and location of the metastatic tumors usually dictate whether surgical resection or stereotactic radiosurgery is the more appropriate treatment strategy. Lesions larger than 3 cm in diameter are rarely candidates for stereotactic radiosurgery. If there is a significant cystic component to a brain metastasis, stereotactic aspiration of the cyst to reduce the size to less than 3 cm, followed by stereotactic radiosurgery, can be considered (85). Alternatively, stereotactic radiosurgery alone may be indicated for smaller cystic metastases (28). Lesions producing disabling neurologic deficits may be more appropriate for surgical resection than for stereotactic radiosurgery, depending upon the location. The choice depends on the likelihood of stereotactic radiosurgery resulting in lesion resolution, anticipated post-treatment edema, and the need for rapid neurologic recovery. Deep lesions (ie, those located in the thalamus or midbrain), complicating medical conditions, and multiple metastases would be indications for stereotactic radiosurgery. Combined treatments including surgical resection of 1 or more of the multiple lesions followed by radiosurgery for the remaining lesions is also a reasonable treatment strategy in many patients. In short, both surgical resection and radiotherapy are essential in the management of brain metastases, and any care team managing this patient population must be able to offer both treatment strategies (05; 63; 34).

Outcomes

The goal of surgery for brain metastases should be the preservation or improvement of functional neurologic status, particularly when long-term control of systemic disease is feasible. Overall survival appears to be most dependent on RPA class. As described above, studies have demonstrated a median survival of 21.4 months in RPA class I patients with a single brain metastasis who underwent surgical resection and adjuvant therapy (107) and 27.6 months for those patients who underwent stereotactic radiosurgery and adjuvant therapy (68). One report of 122 patients undergoing resection of brain metastases found that 10% of patients had postoperative deterioration of neurologic status, and this percentage was even higher in the subset of patients who had previously undergone radiation therapy (32).

Surgical resection of a recurrent metastatic tumor with or without prior stereotactic radiosurgery was associated with a median survival of around 10 months, comparable to some series for survival after initial surgical resection (13; 110; 52; 02).

Several randomized trials have demonstrated that stereotactic radiosurgery improves overall survival and functional performance over WBRT alone in properly selected patients. The largest phase III trial undertaken was the RTOG 9508, in which 333 patients with up to 3 brain metastases were randomly assigned to either stereotactic radiosurgery plus WBRT or WBRT alone. This study demonstrated a statistically significant improvement in survival for patients with a single lesion. Subset analysis also revealed improvements in survival for patients with 2 to 3 metastases having the following characteristics: RPA class 1, age < 50, and patients with non-small cell lung cancer (98; 04). One randomized trial found local control was 70% with stereotactic radiosurgery alone and 82% to 86% with stereotactic radiosurgery and adjuvant WBRT, but failed to demonstrate a survival benefit (survival generally dictated by the burden of extracranial disease) (98; 07; 08). A third phase III randomized trial was terminated early as patients undergoing WBRT were found to be at a greater risk of developing significant decline in learning and memory function at 4 months post-treatment, compared to stereotactic radiosurgery alone (23). In patients with 1 to 3 brain metastases that were randomized to receive stereotactic radiosurgery alone or stereotactic radiosurgery with WBRT, a survival benefit was not found (median OS, 10.4 months SRS vs. 7.4 months for SRS plus WBRT; hazard ratio, 1.02; P = .92) and cognitive deterioration at 3 months was seen more in patients treated with stereotactic radiosurgery plus WBRT (18). Finally, in a large European effort, the European Organization for Research and Treatment of Cancer (EORTC) conducted a phase III trial of patients with up to 3 brain metastases, stable systemic disease, and prior local therapy (either SRS or resection) who were randomized to observation or whole-brain radiation therapy (56). WBRT reduced the 2-year relapse rate both locally and elsewhere in the brain for patients after SRS and resection. Death due to intracranial progression was significantly reduced by the use of WBRT (44% vs. 28%); however, there was no difference in overall survival. Another study comparing WBRT and stereotactic radiosurgery alone found higher rates of leptomeningeal disease (LMD) and inferior distant brain control, but found no difference in local control or in overall survival between the 2 groups (88).

Following resection of brain metastases, should WBRT or stereotactic radiosurgery be given? A study of 700 patients with brain metastases who received stereotactic radiosurgery had 47 patients who underwent postexcision stereotactic radiosurgery (48). A 94% local control rate was achieved with stereotactic radiosurgery – all 3 failures occurring within 16 weeks of stereotactic radiosurgery and in patients whose tumor size was greater than 15 cm³. Other retrospective studies found similar results (97; 90; Luther et al 2013). These data are superior to WBRT following surgical resection, and stereotactic radiosurgery avoids the risk of widespread (but not focal) late radiation necrosis. Recurrence after stereotactic radiosurgery to the resection bed often occurs outside of the planned treatment volume, and thus inclusion of a 2 to 3 mm margin around the area of postoperative enhancement seen on imaging has been recommended (Luther et al 2013). Recurrence in the form of leptomeningeal disease (LMD) is seen in about 13% of patients following stereotactic radiosurgery to the resection bed, with the risk being higher in patients with breast cancer histology. Whether WBRT would have decreased the risk of leptomeningeal disease in these patients is unclear (10).

A survey of radiation oncologists demonstrated the variability in treatment strategies among radiation oncologists for patients with brain metastases and suggests that the lack of prospective randomized studies is only part of the explanation for this variability (61). Access to the various treatment modalities (notably stereotactic radiosurgery and state-of-the-art open neurosurgical resources) and reimbursement are likely to be major factors affecting treatment decisions for patients with metastatic brain tumors.

Adverse effects

Complications of surgery for brain metastases include death (operative mortality being defined as death due to any cause within 30 days of surgery), neurologic worsening, and other complications such as infection, pulmonary embolism, and myocardial infarction. The operative mortality with surgical resection should be less than 5% and in stereotactic radiosurgery less than 2% (16; 89; Hasegawa et al 2003). Complications such as postsurgical hemorrhage or cerebral edema (requiring surgical intervention) following either surgical resection or stereotactic radiosurgery should also occur in less than 2% to 3% of cases (89; Hasegawa et al 2003). In a large retrospective study (458 patients), seizures developed in 2.8% of patients, raising the question of whether or not to use prophylactic anticonvulsants (89). However, the American Academy of Neurology (AAN) performed a meta-analysis that showed that the use of prophylactic anticonvulsants did not significantly reduce the risk of a first seizure; thus, the AAN does not recommend prophylaxis (33; 72; 59).

Radiation necrosis is a late complication of WBRT, stereotactic radiosurgery, or any type of radiotherapy, and this complication can lead to steroid dependence or the need for open surgical intervention. Tumor diameter of greater than 1 cm and the use of immunotherapy are among the many factors that have been associated with an increased risk of radiation necrosis (57; 26). Radiation necrosis can be difficult to distinguish from recurrent tumor. Several studies have addressed radiographic features that can differentiate radiation necrosis from tumor recurrence (44; 116; 112; 25). Narloch and colleagues suggest that radiation necrosis should be strongly suspected in patients with enlarging lesions presenting greater than 9 months after stereotactic radiosurgery (77). In their study, they reported an incidence of radiation necrosis of 93.8% in patients that meet these criteria. In the stereotactic radiosurgery treatment of skull-base tumors, the optic nerve and chiasm are more sensitive than the cranial nerves in the cavernous sinus (particularly with doses of 15 Gy or more) (64). These late complications of stereotactic radiosurgery usually have not been considered in the comparison between surgical resection and stereotactic radiosurgery for brain metastases and would offer some advantage to surgical resection that is not commonly recognized (12). For further reading, the complications of stereotactic radiosurgery for brain metastases have been extensively reviewed (113; 26).

WBRT was once the standard treatment for brain metastases and has generally been the standard treatment following either surgical resection or stereotactic radiosurgery. WBRT improves both local and distant intracranial control following either surgical resection or stereotactic radiosurgery (111; 109). In the past, patients with metastatic brain tumors routinely died within a year of diagnosis, the complications of radiation-induced dementia, brain atrophy, and death were exceedingly rare, especially when doses of 3 Gy per fraction or less were used (105). However, as patients live longer from their primary tumor and systemic metastases, late complications of WBRT are increasing in prevalence, and therefore, the use of WBRT as a treatment option is being reconsidered (Hasegawa et al 2003; 105; 66). A retrospective study suggests that local radiation therapy may be superior to WBRT as an adjuvant to surgical resection (47), suggesting that WBRT may not be the best initial therapy in select patients (Hasegawa et al 2003). A noninferiority, phase 3 randomized trial in non-small cell lung cancer patients with brain metastases not eligible for resection or stereotactic radiosurgery found that WBRT conferred no benefits in overall survival or quality-adjusted life years as compared with optimal supportive care (76). However, WBRT likely still has a vital role in the management of brain metastases in carefully selected patients (40).

Maximizing local control of brain metastases and preventing radiation complications are the most important factors in preserving cognitive function. New efforts to minimize radiation-induced neurocognitive decline that are being investigated include chemoprevention and additional conformal radiation techniques such as hippocampal avoidance with WBRT to spare the neural stem cells used in memory. For further information, the literature on short-term and long-term neurocognitive effects of WBRT has been reviewed in detail (105; 114; 35).

The overall prognoses in brain metastases with various treatment options are as follows: corticosteroids alone (1 to 2 months); corticosteroids plus WBRT (3 to 5 months); stereotactic radiosurgery (6 to 14 months); or open excision plus adjuvant therapy (8 to 21 months) (86; 80; 05; 12; 107; 53). Outcomes range over wide intervals due to a number of significant factors that are histology-specific (100). The prognosis for surgical treatment of recurrent brain metastases (ie, those that recur locally following open excision or stereotactic radiosurgery) in terms of both survival and functional independence is nearly as good as for the initial surgical treatment (13; 110; 52). An analysis of patients with brain metastases who survived more than 5 years showed benefit from aggressive upfront surgical resection or stereotactic radiosurgery in comparison with WBRT alone (24).

As patients survive longer from their primary disease, questions arise regarding re-recurrence of metastases and long-term effects of WBRT. The only fully described series of open excision for re-recurrence compared 17 patients who underwent a third open excision with 9 who did not (13). The former group survived a median of 8.6 additional months, the latter 2.8 months. Randomization was not performed, but these data mirror the survival times closely with and without open excision in the treatment of initial metastatic lesions. In light of the increasing problem of late radiation-induced dementia as patients with brain metastases live longer, the routine use of WBRT is being reconsidered (111; Hasegawa et al 2003; 66). The controversy over routine use of WBRT, as well as guidelines for the management of brain metastases, is the subject of a European Federation of Neurological Societies Task Force report (96) and has been discussed in reviews (34). Another issue post-intervention for brain metastasis is the clinical and economic utility of follow-up imaging – whether it should be performed routinely or only in the presence of new neurologic symptoms (65).

Special considerations

Fortunately, brain metastases are relatively rare in females of childbearing age. If a pregnant woman should be diagnosed with brain metastasis and surgery (open excision or stereotactic radiosurgery) is indicated, the risk to the fetus of either a general anesthetic (for open excision) or radiation exposure (for stereotactic radiosurgery) must be considered. Open excision under local anesthesia may be an alternative that presents less risk to the fetus. Unless the metastasis is diagnosed late in the pregnancy (or is small and asymptomatic), it is unlikely that surgery could be postponed until after delivery. If possible, an alternative option may be to delay the surgery until a more favorable time for the fetus (eg, second trimester).

Clinical vignette

Case 1. A 53-year-old man had been diagnosed with lung cancer 3 years prior to intracranial symptoms. Cranial CT revealed enhancing masses in the left temporal lobe and left cerebellar hemisphere, and he received 50 Gy whole brain irradiation.

Within 2 weeks of completing radiation, he complained of headache and suffered progressive obtundation. The 2 lesions (metastatic adenocarcinoma) were excised in the same operation. The left temporal lesion recurred 1 year later and was re-excised.

Seven months following the re-excision of the left temporal metastasis, no tumor is present.

The patient died from progression of the pulmonary disease 23 months after the first craniotomy.

Case 2. A 45-year-old male presented with severe obtundation and right hemiparesis. Chest x-ray and CT revealed a mass, and contrast CT and MRI of the brain showed 3 large cystic lesions: 2 left frontal and 1 left cerebellar.

All 3 lesions were excised in the same operation without repositioning (right lateral decubitus). Postoperative WBRT (40 Gy) resulted in the disappearance of the residual left frontal metastatic adenocarcinoma.

The patient survived 2 years following the craniotomy before succumbing to the spread of the unresectable lung lesion.

Case 3. A 52-year-old male presented with a generalized seizure but no abnormalities on neurologic examination. Brain CT showed a 1 cm enhancing right occipital mass; chest CT showed a 2 cm right middle lobe mass without mediastinal lymphadenopathy.

He was referred for resection of the brain lesion, to be followed by lobectomy for the lung lesion. The size and location of the brain lesion were well-suited for radiosurgery. Thus, he underwent stereotactic brain biopsy and stereotactic radiosurgery (18 Gy to the tumor periphery) on the same afternoon. The histological diagnosis was large cell carcinoma. Following 30 Gy whole brain irradiation in 10 fractions, he underwent right middle lobectomy. He remained neurologically intact and had no radiographic evidence of tumor recurrence.

Case 4. An 82-year-old female presented with right facial weakness and slightly unsteady gait. Her past medical history is relevant for excision of a melanoma from her left upper back 32 months previously. At the time she had no evidence of other metastases and underwent fractionated stereotactic radiosurgery (25 Gy in 5 fractions) to the right brainstem lesion. As expected, she required oral steroids for nearly 6 months before the right facial droop fully resolved, but she remained fully functional and independent. Unfortunately pulmonary and hepatic metastases appeared, which were not controlled by chemotherapy. She remained neurologically intact until her death from systemic disease 9 months following diagnosis of the brainstem metastasis.

Scientific basis

Brain metastases occur in up to 40% of the 1.4 million patients diagnosed with cancer per year in the United States (31), making metastatic brain tumors 5 to 10 times more common than primary brain tumors. The incidence of brain metastases is increasing with more sensitive diagnostic imaging techniques and increased overall survival due to more effective treatments of the primary tumor and systemic metastatic disease. The most common tumors to metastasize to the brain include lung (30% to 40% of all cerebral metastases), breast (20% to 30%), melanoma, gastrointestinal tract, and urinary tract (5% to 10% each). The percentage of individuals with a primary tumor type who will develop intracranial metastases varies depending on the primary tumor. Melanoma is the primary tumor with the highest tendency to metastasize to the brain; however, the most common sites of origin for brain metastases are the lung and breast due to the higher overall incidences of these cancers. Up to 70% of melanoma, 20% to 35% of lung, breast, urinary tract, prostate, and lymphoma, and 5% to 10% of gastrointestinal tract and female genital tract primary tumors will progress to intracranial metastases (49). For non-small cell lung cancer (70), early detection and treatment of the primary disease may decrease the risk of developing brain metastases to 10% (45). In 10% to 20% of cases, the metastatic tumor is detected before the primary, and in 10% to 20% of these cases the primary site goes undetected even at autopsy (49; 108). The spread of tumor to the brain is hematogenous in most cases (104).

The location of the metastatic tumor in the intracranial compartment is an important consideration in surgical decision-making. Apart from a predilection for prostate cancer to metastasize to the skull and dura, most metastases occur at the gray-white junction—where there is a rapid reduction in vascular diameter—and/or in the distal vasculature (watershed areas) (46). Other important considerations for surgical decision-making are proximity to eloquent cortex, histology, size (> 3 cm in diameter being a relative contraindication for stereotactic radiosurgery), number of lesions, character of the lesions (ie, solid, cystic, or ring-enhancing), and clinical condition of the patient (05; 85). Infratentorial metastases, if symptomatic, are likely to require surgical resection in order to decrease the mass effect rapidly. Nonetheless, the prognosis for infratentorial metastases is comparable to that for supratentorial metastases with aggressive management (05; 51).

Patients may develop brain metastases late in the course of the disease, or alternatively, signs and symptoms due to a brain lesion may be the first presentation of systemic malignancy. In cases without a primary tumor diagnosis, open surgical excision or stereotactic biopsy for tissue confirmation is necessary to guide further treatment (104). It is important to note that even when a systemic tumor diagnosis is already established, brain lesions may not be metastatic; up to 10% of lesions suspected to be metastatic tumor prove to be otherwise (eg, primary brain tumor, abscess) (86; 01). Gadolinium-enhanced MRI is the primary and most sensitive imaging technique for diagnosis of suspected brain metastases. Advanced MRI techniques such as MR spectroscopy may help to distinguish a metastatic brain tumor from a primary brain tumor (83); delayed contrast extravasation MRI, with imaging 1 hour or more after contrast infusion, may differentiate between tumor recurrence and necrosis (116). Additionally, other imaging techniques, such as fiber tract imaging, functional positron emission tomography (fPET), and magnetoencephalography can be used to elucidate functional anatomy and its relationship to the tumor. Intraoperative MR imaging is also being investigated as a tool for improving brain tumor surgery (106).

5-Aminolevulinic acid (5-ALA) is a novel imaging modality with a potential role in the diagnosis of brain metastases, although it has not yet been approved in the United States. 5-ALA, a precursor molecule in the heme pathway, is metabolized into protoporphyrin IX, which is capable of fluorescence after excitation with a violet-blue light (370–440 nm) (102). The utility of 5-ALA stems from its preferential uptake and breakdown by CNS tumors, thereby allowing surgeons to better visualize tumor margins (102; 50). There is level 1 evidence to support the role of 5-ALA in the resection of high-grade gliomas (30). Further investigation is necessary to determine the utility of this novel imaging modality in the diagnosis and management of other CNS tumors, including brain metastases.

The Radiation Therapy Oncology Group (RTOG) studied factors that may play a role in survival in patients with brain metastases and developed recursive partitioning analysis (RPA) classes. Patients in RPA class I have a Karnofsky Performance Score (KPS) of 70 or higher, are less than 65 years of age, have controlled primary tumor, and have no extracranial metastases; these patients are the best candidates for surgery and/or radiosurgery. RPA class II entails a KPS greater than 70, without the other criteria for class I. RPA class III patients have a KPS of less than 70. Patients in RPA classes II and III may also benefit from surgery, but surgical resection should only be offered to carefully selected individuals. The RPA classification system was validated in a large retrospective series which concluded that RPA classification has prognostic significance for patients with a single brain metastasis who undergo surgical resection and adjuvant therapy (107). This series also found that a median survival of 21.4 months can be achieved in RPA class I patients with a single brain metastasis who undergo both surgery and adjuvant therapy (107). A separate study found that RPA class I patients have a median survival of 27.6 months if treated with stereotactic radiosurgery and adjuvant therapy (68). The intermediate RPA class II classification has less prognostic utility given its broad definition; however, efforts to subclassify this group may help to guide decision-making for or against surgery (21). Drawbacks of the RTOG RPA system (eg, not addressing the number of metastasis) have fostered a new system: the graded prognostic assessment (GPA) (100). This system allots points to a number of factors that are specific to the underlying histology. It provides a more nuanced and contemporary assessment of the prognosis of patients with brain metastases. Importantly, perhaps the strongest predictor of survival in patients with brain metastases is the degree of control of extracranial disease, including both the primary tumor, as well as noncentral nervous system metastases (80; 36). A newly proposed grading system attempts to quantify the extent of extracranial disease in patients with brain metastases by measuring various surrogates of disease extent, including albumin, lactate dehydrogenase, and number of extracranial organs affected by metastasis (78). This study concludes that patients with elevated LDH, low albumin, and involvement of at least 2 extracranial organs might be considered for supportive care rather than surgical resection and/or stereotactic radiosurgery.

The radiosensitivity of the metastatic lesion is usually not a major factor when considering open excision versus stereotactic radiosurgery, apart from highly radiosensitive metastases such as germ cell, lymphoma, and small-cell lung primary tumors. Local control with single fraction stereotactic radiosurgery is usually excellent even in the context of the more radioresistant metastases such as melanoma, sarcoma, and renal cell carcinoma (82). These radioresistant metastases tend to decrease in size slowly and then stabilize with a central hypodensity (representing necrotic material) on CT or MRI scan (16).