The brain microenvironment, including brain vascular endothelial cells and stromal cells (microglia and astrocytes), provides growth and invasion advantages to the disseminated tumour cells. Astrocytes are intimately involved in maintaining normal homeostasis of the brain microenvironment, accomplished through transport of nutrients to the neurons and facilitation of neural signal transduction. These mechanisms could be co-opted to protect tumour cells in brain metastases from the cytotoxic effect of chemotherapeutic agents through a decreased apoptosis or survival genes upregulation [4]. The growth of metastatic brain tumours is also critically dependent on angiogenesis and on the expression of vascular endothelial growth factor (VEGF). Disordered angiogenesis results in structural and functional abnormality of tumour-associated blood vessels, characterized by defective endothelial cells, pericyte covering and basement membranes. These abnormalities can directly restrict the delivery of oxygen, leading to intratumoural hypoxia that increases resistance to both systemic agents and ionizing radiation. Another protection for the tumor cells is the BBB. The BBB is a selective barrier between the systemic circulation and cerebrospinal fluid that is formed by specialized endothelial cells lining the cerebral microvasculature, together with pericytes and astrocytes. Large hydrophilic molecules, including chemotherapeutic and molecular-targeted drugs, are excluded from the central nervous system unless they can be actively transported by receptor-mediated transcytosis. In addition, the BBB expresses high levels of drug efflux pumps such as the P-glycoprotein (PgP)/multi-drug resistance proteins, which actively remove some chemotherapeutic drugs from the brain [5, 6]. Selective permeabilisation of the BBB that may facilitate tumor-specific access of chemotherapeutic agents is under experimental investigation [7]. Finally, phenotypic alterations in tumours cells may promote the progression of cancer. Activation signalling pathways such as of epidermal growth factor receptor (EGFR) (via gene amplification, overexpression and/or mutation) or through the EML4–ALK translocation can be observed in both primary lung cancer cases and brain metastases.

Considered as the preferred imaging for brain metastasis, contrast-enhanced magnetic resonance imaging (MRI) has been enriched by new experimental techniques, mainly dynamic contrast-enhanced imaging (perfusion-weighted MRI: PWI) and magnetic resonance spectroscopy (MRS). PWI enable to estimate the microvascular environment by measuring the cerebral blood volume (CBV). MRS analyses levels of choline and of N-acetylaspartate that are modified (respectively increased and decreased) in brain tumours. Such markers are then mainly useful to differentiate brain metastasis (recurrence) from radiation necrosis [8, 9]. Preliminary data also suggest 11C-methionine positron emission tomography (MET-PET) to be a promising tool for similar indication [10]. Other MRI advances such as magnetization transfer and triple dose gadolinium imaging are being studied to further improved lesion detection [11].

The patient’s DS-GPA score is important in determining whether or not surgery is appropriate. Surgical resection combined with adjuvant radiation therapy is a standard treatment since the 90s in patients with a good prognosis and solitary accessible brain metastasis. Two randomized clinical trials demonstrated that surgery plus whole brain radiotherapy (WBRT) increased survival benefit in patients with favourable baseline criteria over WBRT alone [12, 13]. Patchell et al. showed in the 25 patients treated with the combined modality fewer local recurrences (20 vs. 52 %), improved survival (40 vs. 15 weeks), and a better quality of life. Nowadays, radiosurgery (SRS) to the tumor bed is proposed instead of WBRT after surgery. In retrospective studies, reported local control rates were promising and comparable with postoperative WBRT. Still, wile deferring WBRT, there is a risk of recurrence elsewhere in the brain and radionecrosis rates may be higher [14–16]. A randomized phase III trial is currently comparing WBRT vs. SRS in patients with brain metastases that have been removed by surgery [http://​clinicaltrials.​gov: NCT01372774].

Others technical improvements may confer to neurosurgeons the chances to perform complete surgery while minimize injury to normal tissue. Intraoperative MRI is specifically assessed in primary malignant brain tumours and permitted, in a randomized trial, to detect residual tumor and additional tumor resection [17]. Such procedures were also assessed in non-randomized studies for brain metastases with encouraging results [18]. Awake craniotomy and neurophysiological monitoring for functional assessment during resection have also been addressed in small series. There, complete microsurgical resection of metastatic tumours in the primary motor cortex was feasible and efficacious [19, 20]. Other promising technologies include optical and molecular visualization such as 5-aminolevulinic acid (5-ALA) fluorescence or fluorescein staining of malignant tissues within the operative bed [21].

With the multiplication of high energy x-rays produced by linear accelerators, SRS (in unique or multiple fractions) have become a common therapeutic modality for patients with good prognosis and a limited number of brain metastases. The principle of such treatment implies concentrating high-dose ablative radiation within a tumor while avoiding radiation of the surrounding healthy tissue that could lead to neurocognitive decline, as observed after WBRT. Metastases are usually small (<3 cm) and well limited lesions making them ideal targets for SRS. Moreover, focal high dose irradiation, as compared with neurosurgery, has the ability to treat surgically inaccessible areas, several lesions, and to be non-invasive and cost-effective. There is no randomized trial comparing surgery and SRS but an increasing number of uncontrolled studies suggest that focal irradiation may be as effective, with local control rates of 81–98 % for lung cancer [22, 23]. Only a small randomized study showed equivalent outcomes after Gamma Knife SRS vs. surgical resection plus WBRT but only 64 patients were included [24].

There are, in opposite evidences on the benefit of adding an SRS boost to WBRT as compared with WBRT alone [25, 26]. The much larger of the two randomized trials addressing this question (RTOG: Radiation Therapy Oncology Group 9508) showed that the 167/333 patients receiving SRS had an 6-month improved performance status. Survival was similar in the whole population but was significantly longer with SRS in patients with a single brain metastasis and in patients with favourable prognosis [25]. The question of the addition of WBRT after SRS has also been addressed in prospective randomized trials. The larger one was conducted by the European Organisation for Research and Treatment of Cancer (EORTC 22952–26001). In this trial, 359 patients with one to three brain metastases were randomly assigned to WBRT or observation following definitive treatment (either SRS or surgery). Progression to a performance status of more than 2 (primary endpoint) was similar in the two arms but WBRT significantly decreased the local relapse (at initial site and within the brain). There was no significant difference in overall survival (median 10.9 and 10.7 months with WBRT and observation, respectively) [27]. Consistent with these results, two others prospective studies reported that the use of WBRT plus SRS (no surgery in those trials) did not improve survival for patients with a limited number of brain metastases, but intracranial relapse occurred considerably more frequently in those who did not receive WBRT [28, 29]. On the other hand, effects of WBRT after SRS (or surgery) may negatively impact on cognitive function [30, 31]. Another large randomized trial conducted by the North Central Cancer Treatment Group is currently addressing this issue (NCT00377156).

Close future will possibly include SRS for selected patients with up to ten lesions. Yamamoto et al. have recently showed that SRS without WBRT was feasible as the initial treatment for patients with five to ten brain metastases. In this prospective study, patients with five to ten brain metastases did not seem to fare worse than those with two to four brain metastases in terms of overall survival, intracranial tumour control, neurological deterioration and death, leukoencephalopathy, or salvage treatment [32]. Results on the ongoing randomised trial comparing SRS and WBRT for patients with five or more brain metastases (NCT01731704) will hopefully better define the role of SRS in this setting.

WBRT alone remains the preferred treatment for patients with a poor prognosis according to DS-GPA, patient who are not candidates for surgery or SRS, or for prophylactic cerebral irradiation (PCI). Moreover, histological subtypes such has small-cell lung carcinoma (SLC) may carry a much higher risk of distant brain failure with SRS alone. WBRT is generally thought to improve survival by several months in comparison with best supportive care (BSC) [33–35]. In fact there is no available prospective data excepted an interim report of the Medical Research Council trial that included patients with brain metastases from non-small cell lung cancer (NSCLC). In the 151/500 initially planned randomized patients, no difference in survival or quality-adjusted life was observed between patients treated with BSC vs. BSC plus WBRT [36].

Technical or pharmacological strategies to reduce the risk of neurocognitive decline after WBRT are under active investigation. Hippocampal-sparing radiation is presently investigated in the RTOG 0933 phase II study (NCT01227954). In the preliminary results, evaluable patients (42 patients with brain metastases from histologically proven malignancy other than SLC) had significantly less decline in measures of delayed recall at 4 months as compared with historical controls. Furthermore, only three patients relapsed within the hippocampal avoidance region [37]. Concurrent memantine, an oral N-methyl-D-aspartate (NMDA) receptor antagonist, has been evaluated in patients receiving WBRT in a randomized trial. Memantine non-significantly delayed time to cognitive decline but only 29 % patients included (n = 149) were evaluable for of memory decline at 24 weeks, defined as the primary endpoint [38]. Radiation sensitizers given with WBRT, including motexafin gadolinium and efaproxiral have failed to demonstrate their superiority as compared with placebo in phase III trials [33–35]. WBRT has also been delivered with concurrent systemic chemotherapy or targeted therapies.

The primary approaches to the treatment of patients with brain metastases include surgery or radiotherapy. Lung cancers are however generally chemosensitive and systemic chemotherapy may induce objective response in patients with brain metastases [39–42]. Despite such responsiveness to chemotherapy, WBRT may not be omitted in patients receiving chemotherapy [43]. Systemic treatments such as temozolomide are of interest in salvage management [44]. Adding WBRT to chemotherapy has been evaluated in a phase III trial conducted by the EORTC in 120 SCLC patients pre-treated or not with chemotherapy. Patients were randomized to receive teniposide alone or with WBRT. The combination of chemotherapy and WBRT had a significantly higher response rate but overall survival was not improved [45]. Several chemotherapy regimens have as well been assessed concurrently with WBRT in NSCLC patients with disappointing results [46–48]. The role of temozolomide associated with WBRT in NSCLC is controversial. While some studies have reported good response rates and a limited toxicity [49–51] others (including the prematurely stopped RTOG 0302 phase III trial) had demonstrated deleterious effects and no effect on survival [52, 53].

Molecular-targetable alteration can be observed in a large proportion of primary NSCLC. Prospective evidence for targeted therapies delivered in patients with brain metastases from NSCLC is summarized in Table 16.2. As the primary tumour, brain metastases also express increased EGFR levels [54, 55]. Anti-EGFR tyrosine kinase inhibitors, erlotinib and gefitinib, have antitumor activity in NSCL patients with brain metastases [56–60]. Responses rate and outcomes appear to be increased in patients with known EGFR mutations as compared with wild-type tumours [61]. Concurrent erlotinib and WBRT has been prospectively studied in NSCLC patients with good overall response rate (86 %) and a median overall survival was 12 months [62]. However, in the randomized RTOG 0302 trial addressing the use of erlotinib (or temozolomide or no systemic treatment) with WBRT and SRS, overall survival and toxicity were worse with systemic therapies as compared with WBRT/SRS alone [3, 52]. The RTOG 0302 has however been criticized as the allocation of treatment was not biomarker-driven, statistical power was low, and better outcomes were observed in the control arm as compared within historical studies. As long as no other randomized prospective data is available, the use of anti anti-EGFR may be dedicated to salvage treatment after irradiation or when radiotherapy may be delayed (symptomatic extracerebral disease with absence of neurologic symptoms) [63].