Systemic Therapy for Lung Cancer Brain Metastases: A Rationale for Clinical Trials
Systemic Therapy for Lung Cancer Brain Metastases: A Rationale for Clinical Trials
ABSTRACT: Despite the high prevalence of brain metastases in patients with metastatic lung cancer, these patients have been excluded from enrollment in clinical trials of new therapeutic drugs. The reasons for exclusion have centered on concerns that the blood-brain barrier may impede drug delivery into brain metastases, that brain metastases confer a dismal survival for metastatic lung cancer patients, and that brain metastases carry risk for cerebrovascular hemorrhage. A focused, updated review of these issues, however, clearly shows that these particular concerns are unwarranted. An extensive review of clinical trials on the efficacy of chemotheraputic agents against lung cancer brain metastases is also provided. This collective information describes an area in need of therapeutic development and supports an initiative to evaluate novel targeted therapies for lung cancer brain metastases.
The brain is a frequent site of metastases in patients with advanced lung cancer and can be associated with substantial morbidity. Historically, poor prognosis associated with brain metastases has led to therapeutic nihilism as a self-fulfilling prophecy. Over the past 10 years, however, with earlier detection of brain metastases and improved local treatment options, survival and prognosis have improved. These improvements are two of several factors that provide justification for including patients with brain metastases in clinical trials of new systemic therapies. We address several commonly cited beliefs that have hindered the development of therapeutic agents for central nervous system malignancies, and we advance reasons that we feel justify including patients with brain metastases in clinical trials of new therapeutic approaches for lung cancer.
The Blood-Brain Barrier May Not Substantially Impede Drug Delivery Into Brain Metastases
The microvasculature of the brain parenchyma is lined by a continuous, nonfenestrated endothelium with tight junctions and has little pinocytic vesicle activity.[1,2] This blood-brain barrier (BBB) limits the entrance of circulating macromolecules into the brain parenchyma. The BBB and the lack of a lymphatic system are responsible for maintaining the brain as an immunologically privileged site and for protecting the brain against the entry of most drugs and invasion by microorganisms. This barrier has been hypothesized to be a reason why brain neoplasms and metastases are often resistant to chemotherapeutic drugs.[4,5]
The observation of brain metastases in 55% of patients who achieved a complete response (CR) with neoadjuvant chemotherapy (often developing very early after completion of therapy) has been interpreted as evidence for the brain being a pharmacologic sanctuary. Very small metastases that have not yet developed neovascularizition may be protected from chemotherapy. However, there is little if any evidence that this has any relevance in the delivery of chemotherapy to established brain metastases with highly permeable angiogenic vasculature sufficiently abnormal to permit enhancement on magnetic resonance imaging (MRI) scans.[1-4]
Abundant evidence exists that the BBB is not fully operational in brain tumors.[7-9] First, the barrier does not prevent the entry of circulating metastatic cancer cells into the brain parenchyma. In addition, many malignancies in the brain appear to degrade the integrity of the BBB, permitting tumor drug delivery.[10,11]
Second, animal studies demonstrate that molecular tracers as large as 1.5 kD can be extravasated from tumor vasculature of experimental brain metastases.[2,12] The progressive growth of brain metastases is associated with increased expression of vascular endothelial growth factor and leads to tumor vascular permeability, uncharacteristic of surrounding normal brain parenchymal vasculature.[13,14]
Third, the BBB can be permeable in ischemic regions of the brain where increased endothelial pinocytosis, opening of the interendothelial tight junctions, and damage to endothelial cells can occur.[15,16] Degeneration and central necrosis often occur in large (0.2-mm2) brain metastases, and the BBB in these lesions is not intact, possibly due in part to endothelial cell damage or a direct effect of vascular endothelial growth factor (VEGF).
Even more direct data exist on the accumulation to therapeutic levels of chemotherapeutic agents in brain tumors.[17-34] These data have been obtained from analyses conducted after surgical resection, biopsy, or autopsy removal of brain tumors from chemotherapy-treated patients. Direct tissue measurement of drug levels has demonstrated that chemotherapy agents accumulate heterogeneously but often to a much greater extent than in the surrounding normal brain[12,35-37] or cerebrospinal fluid.[26,29,38-40] Pharmacokinetic and anatomic studies of the BBB have suggested that lipophilic drugs penetrate the normal central nervous system (CNS) much more readily than do hydrophilic drugs. However, there is little evidence that this is true for brain tumors, and hydrophilic drugs have shown activity against brain tumors.
The hydrophilic agent cisplatin, like most other chemotherapeutic drugs that have been studied, accumulates in human brain tumors much more than penetration into the normal CNS would predict.[40,41] After therapeutic dosing, cisplatin could be measured in brain tumors from autopsied patients at potentially cytotoxic concentrations. In patients undergoing biopsies after receiving small doses of cisplatin, the concentration of cisplatin in brain tumors, after correction for pharmacokinetics and tissue distribution, is equivalent to that in most normal tissues, except the liver and normal brain, where concentrations are higher and lower, respectively.[31,43]
Brain Metastases Do Not Necessarily Mandate a Dismal Survival for Metastatic NSCLC
Retrospective data from M. D. Anderson Cancer Center demonstrate that for patients with non–small-cell lung cancer (NSCLC) and metastases at one or two organ sites, survival is slightly worse if the brain is involved than if it is not (median survival of 7–8 vs 9–10 months, respectively; unpublished data). However, for patients with involvement of more than two organ sites, the presence of brain metastases does not appear to influence survival at all. Thus, it appears that the total burden of cancer is a more important prognostic factor than the presence of brain metastases per se.
More routine use of MRI screening for brain metastases can detect these lesions earlier, long before they are an imminent cause of death or disability. Earlier detection of oligometastatic brain lesions allows the opportunity to treat with effective stereotactic radiosurgery (SRS) techniques, which are much faster and less toxic than is traditional whole-brain radiotherapy (WBRT), so that systemic therapy is not delayed or compromised. The detection and treatment of brain metastases at a low volume of disease is associated with better survival.
Historically, WBRT has been the mainstay of treating brain metastases, and the median survival time after WBRT has been 2.5 to 7 months.[46-48] The prognosis for brain metastasis patients before WBRT has been categorized according to a recursive partitioning analysis (RPA) schema from the Radiation Therapy Oncology Group (RPA class 1 = Karnofsky performance scale [KPS] score > 70%, age < 65 years, and no extracranial disease; RPA class 2 = KPS score > 70% and age ≥ 65 years or active extracranial disease; RPA class 3 = KPS score ≤ 70%). The median survival time has been 5 to 7 months for RPA class 1, 3 to 4 months for RPA class 2, and 2.5 to 3 months for RPA class 3.[45,47,49-51]
The prognosis and patterns of treatment failure for brain metastasis patients after WBRT have been categorized according to an RPA schema based on decision nodes that include KPS score ≤ 80%, age > 60 years, radiation dose < 66 Gy, weight loss > 5%, and malignant pleural effusion. This RPA schema has demonstrated median survival differences ranging from 3.3 to 12.6 months. Surgical resection of brain metastases has been generally reserved for patients in RPA class 1 or for patients with large symptomatic metastases refractory to radiotherapy.
Since the late 1990s, SRS has been used with increasing frequency for the treatment of brain metastases.[54-61] For up to four tumors smaller than 3 cm and as small as 0.5 cm, SRS, using GammaKnife techniques (from a fixed cobalt source) or a linear accelerator, has demonstrated effective local control of brain metastases as a single-day outpatient procedure without the alopecia, fatigue, or potential neurocognitive toxicity encountered with WBRT.[58,59,61] WBRT is still essential in treating multiple or disseminated brain metastases, and neurosurgical resection is still the best option for management of large brain metastases that are resistant to radiotherapy or markedly symptomatic. However, SRS can now be used to treat most other cases of brain metastases effectively with fewer side effects, and systemic therapy can be resumed immediately.[54,58,59,61] In addition, SRS as consolidation therapy has been shown to improve survival after WBRT in patients with solitary brain metastases.
Although the efficacy of SRS has not been directly compared with neurosurgical resection for NSCLC brain metastases, local control rates have invariably been high; the median survival time after SRS has been approximately 5 to 11 months and as high as 21 months for patients without extracranial disease, which is significantly better than the historical median survival time after WBRT.[54,57-60,63-65] This improvement is likely due to the practice of SRS being used selectively to treat lower-volume, early oligometastatic disease in the brain with a better prognosis, whereas WBRT is still the gold standard therapy for bulky and extensive metastatic disease in the brain with a poorer prognosis.[55,66,67]
SRS and WBRT appear to have equivalent efficacy, however, in patients with a similar disease burden. The overall survival for patients with up to four brain metastases treated with SRS alone or with concomitant WBRT has been the same in both retrospective and prospective randomized multi-institutional trials.[58,68,69] New sites of metastases are more likely to develop after SRS than after WBRT, but SRS can be repeated at new sites if the metastases develop gradually as oligometastatic disease[70,71] and can be used to effectively treat tumors that progress after WBRT.
Other Factors Influencing Prognosis
In addition to the availability of SRS, which allows effective early intervention for NSCLC brain metastases with minimal morbidity, other factors have somewhat reduced the effect of brain metastases on prognosis. Specifically, routine MRI screening is now often detecting asymptomatic small brain metastases early, while patients have a relatively long life expectancy and a preserved performance status. Overall, the presence of brain metastases is no longer necessarily associated with the very poor prognosis that has historically been associated with them.
Because of this shift, it would be very reasonable to routinely include patients with brain metastases in clinical trials of new systemic therapies—particularly patients with small minimally symptomatic brain metastases (whether or not they have been treated with radiotherapy) and patients with oligometastatic brain disease that has been treated with SRS or surgical resection. If a new systemic therapy is effective against the small and minimally symptomatic brain metastases, the morbidity associated with WBRT might be deferred.