Fax: (819) 820-6452
Enhanced chemotherapy delivery by intraarterial infusion and blood-brain barrier disruption in the treatment of cerebral metastasis
Article first published online: 8 JAN 2007
Copyright © 2007 American Cancer Society
Volume 109, Issue 4, pages 751–760, 15 February 2007
How to Cite
Fortin, D., Gendron, C., Boudrias, M. and Garant, M.-P. (2007), Enhanced chemotherapy delivery by intraarterial infusion and blood-brain barrier disruption in the treatment of cerebral metastasis. Cancer, 109: 751–760. doi: 10.1002/cncr.22450
- Issue published online: 2 FEB 2007
- Article first published online: 8 JAN 2007
- Manuscript Accepted: 21 NOV 2006
- Manuscript Revised: 26 OCT 2006
- Manuscript Received: 14 AUG 2006
- blood-brain barrier disruption;
- brain metastasis;
Cerebral metastases are clinically significant in 10% to 30% of patients with neoplasia. Multiple cerebral metastases are typically treated with palliative radiotherapy. There is no consensus on the role of enhanced chemotherapy delivery as an adjuvant treatment modality in this disease. In this report, the authors detailed their experience with intraarterial (IA) chemotherapy infusion with and without blood-brain barrier disruption (BBBD) in patients with multiple cerebral metastases.
From November 1999 to May 2005, 38 patients with multiple cerebral metastases were enrolled in a prospective study. Patients were treated with IA carboplatin, except for those with cerebral metastases of systemic lymphoma, who were administered IA methotrexate. Osmotic BBBD was offered to patients without the presence of a significant mass effect. These regimens were coupled with intravenous etoposide and cyclophosphamide. Cycles were repeated every 4 weeks.
Survival was calculated from study entry and radiologic response was based on MacDonald criteria. Kaplan-Meier estimates were generated for all subgroups. Mean and median survival obtained was as follows: 34 and 29.6 months for the whole group; 33.6 and 42.3 months for ovarian carcinoma; 15.3 and 13.5 months for lung adenocarcinomas; 8.3 and 8.8 months for small cell lung carcinoma; 8.9 and 8.1 months for breast carcinoma; and 24.8 and 16.3 months, respectively, for cerebral metastasis from systemic lymphoma.
Even with a small number of patients in each subgroup, the results obtained seem promising for multiple brain metastasis of ovarian carcinoma, adenocarcinoma of lung, small cell lung carcinoma, and systemic lymphoma. Cancer 2007;109:751–60. © 2007 American Cancer Society.
Brain metastases are the most common intracranial tumors among adults. They occur in 20% to 40% of cancer patients, which represents over 170,000 new cases per year in the US.1, 2 The metastatic process to the central nervous system (CNS) usually implies an uncontrollable disease, and typically represents an exclusion criterion in most Phase II and III clinical studies. The emergence of cerebral metastasis significantly impacts the quality of life of the patient, producing neurological morbidity.3 More so, it adversely impacts survival, as their presence entails per se a poor prognosis. According to the recursive-partitioning analysis (RPA) of the Radiation Therapy Oncology Group (RTOG), median survival of brain metastases patients ranges between 2 and 7.1 months, thus illustrating the poor prognosis observed in these patients.4
Most patients will present with multiple lesions.5, 6 Magnetic resonance imaging-based studies reveal that approximately 80% of patients have more than 1 metastasis at presentation.6 In these patients the approach is usually considered palliative, and whole-brain radiation therapy (WBRT) is offered in an attempt to briefly stabilize disease progression. However, the role of radiotherapy remains controversial.7 Even though this treatment modality has been shown effective for local tumor control, survival benefits have been modest, ranging from 3 to 5 months.4–6 More so, WBRT has been associated with significant side effects.8, 9
Chemotherapy has been of limited use in patients with brain metastasis because of the blood-brain barrier (BBB). The delivery impediment of therapeutic molecules related to the physicoelectric characteristics of the BBB and the activity of efflux pump systems is such that a significant area under the curve is hard to maintain.10–12 Based on the rationale of bypassing the BBB, different approaches have been proposed.13 One such approach consists of increasing the effective concentration of a chemotherapy agent by intraarterial (IA) administration. By doing so, this regional delivery paradigm increases local plasma peak concentration and local area under the curve, ultimately translating into a 3 to 5.5 × factor increase in intratumoral concentration.13, 14 This strategy can also be augmented by manipulating the BBB permeability through different means.15, 16 To date, there is no consensus on the role of these approaches in the treatment of brain metastases. We hereby relate our experience with the IA administration of chemotherapy in cerebral metastatic disease, with and without osmotic blood-brain barrier disruption (BBBD).
MATERIALS AND METHODS
This prospective Phase II study was conducted at the Centre Hospitalier Universitaire de Sherbrooke (CHUS) from November 1999 to May 2005. The protocol was approved by the Institutional Review Board and informed consent was obtained in accordance with institutional regulations. Patients with histologically confirmed brain metastases were eligible. All histological subtypes were allowed, except notoriously chemoresistant histologies, such as hypernephromas and melanomas.
Eligibility criteria also included a Karnofsky Performance Status (KPS) >50 and measurable disease on initial contrast enhanced computed tomography (CT)/ magnetic resonance imaging (MRI). Prior radiotherapy and chemotherapy was allowed. The complete list of inclusion criteria is detailed in Table 1. The primary endpoint of this study was survival.
|Histological confirmation of cerebral lesion|
|Previous radiotherapy and chemotherapy permitted|
|Normal renal and hepatic functions|
|Normal cardiac and pulmonary functions|
|White cells ≥2.5 × 103/mm3|
|Granulocytes ≥1.5 × 103/mm3|
|Platelet ≥100 × 103/mm3|
After enrollment and initial evaluation, patients were treated every 4 weeks (1 cycle) for up to 12 cycles. The procedure was accomplished in a standardized way. After general anesthesia, a catheterization via transfemoral puncture of either the right or left carotid or vertebral artery was performed, the parent vessel(s) used depending on the tumor(s) location. Before appropriate vessel selection a diagnostic angiogram was obtained in the vessel(s) of interest. However, a 4-vessel angiogram was not performed routinely. The catheter was placed at the level of C1-C2 for carotid and at the level of C5-C6 for vertebral circulation. After determination of the adequate infusion rate for mannitol infusion to open the barrier, mannitol was infused at high flow over 30 seconds in the selected artery. During this infusion the whole vascular tree in the selected distribution was filled with mannitol, whereas reflux in the common and external carotid artery was kept to a minimum. The chemotherapy was then infused at a rate calculated to prevent streaming.16
Two different chemotherapy regimens were used based on tumor histology. Methotrexate (5 gm), for systemic lymphomas, or carboplatin (400 mg/m2) for all other histologies, was infused via the catheter. Both regimens are detailed in Table 2. Leucovorin rescue was used with the methotrexate protocol. All patients received Neupogen (5 μg/kg subcutaneous daily) for 7 days.
|Chemotherapy agents||Doses, mg/m2|
Because the procedure induces a transitory increase in intracranial pressure, only 1 vascular territory could be disrupted per session. Thus, patients with multiple metastases covering more than 1 vascular distribution were approached as follows: different vascular distributions were treated alternately from cycle to cycle. The presence of a significant mass effect, as exemplified by a closed quadrigiminal cisterna, dilatation of the contralateral ventricular system, or uncal herniation, represented an absolute contraindication to BBBD. Some patients presenting an excessive mass effect but otherwise meeting all the inclusion criteria were offered IA chemotherapy without BBBD (Table 3). In this setting, general anesthesia is not required, and as the BBB is not breached the chemotherapy dose was equally split between each vascular distribution treated, thus covering all lesions for each cycle.
|Intraarterial (IA)||IA + BBBD|
All patients were monitored with complete blood and platelet counts every week. A biochemical, kidney and liver function profile, and electrolyte assessment was requested every 4 weeks. Neurological and general examination was performed before each cycle. Assessment of overall response was based on tumor evaluation from CT/MRI scans obtained before each treatment and interpreted according to the MacDonald criteria (Table 4). Data were prospectively collected from study entry and at every visit during the study.
|Complete response (CR)||Complete disappearance of all enhancing tumor and no corticosteroid use other than physiologic doses with stable or improved neurological condition|
|Major partial response (MPR)||>75% reduction in enhancement|
|Partial response (PR)||>50% reduction in enhancement and stable corticosteroid use with stable or improved neurological condition|
|Progressive disease (PD)||>25% increase in contrast enhancement or any new tumor nodule, with or without neurological progression|
|Stable disease (SD)||Other circumstances|
Mean and median survivals from study entry were calculated from the first cycle of treatment to death or May 1, 2005. Based on these, Kaplan-Meier estimates were also extrapolated.
From November 1999 to May 1, 2005, 38 patients were accrued in this study and grouped in 5 distinct histological classes (Table 5).
|Histologic subtype||No. of cases||Average no.of metastases|
|Small cell lung carcinoma (SCLC)||9|
|Anal epidermoid carcinoma||1|
The median age was 55, 9 years (20–74 years) for the whole group at enrollment, and there were 25 women (65.8%) and 13 men (34.2%). These 38 patients were submitted to 199 procedures (mean of 5.2 cycles per patient). Nineteen were exposed to BBBD, whereas the remaining 19 patients were treated with IA chemotherapy. On accrual, 63.2% of patients had received radiotherapy and 73.7% had already been exposed to chemotherapy for their primary disease (Table 6). The agents used were: carboplatin, cisplatin, etoposide, taxol, topotecan, 5FU, navelbin, vincristine, vinblastin, and adriamycin. In an attempt to characterize our study population prognostically, we stratified our patients according to the RTOG partitioning analysis classes. Three patients were stratified as class 1, 18 as class 2, and 17 as class 3. Overall, 92.1% of patients were in classes 2 or 3 (Table 7). The mean number of metastatic lesion was 3.5 lesions per patient.
|RPA classes||Criteria||No. of patients||Median survival time expected, mo|
|I||All of the following:|
|1. Karnofsky ≥70|
|2. Age ≤65||3||7.1|
|3. Controlled primary tumor with no extracranial metastases|
|II||1. Karnofsky ≥70 and at least 1 of the following||18||4.2|
|2. Age ≤65|
|3. Uncontrolled or synchronous primary tumor presence of extracranial metastases|
Tumor Response, Time to Tumor Progression (TTP), and Overall Survival
The mean overall survival from diagnosis of brain metastases was 34 months, whereas median overall survival was 29.6 months. Mean and median overall survivals from study entry were, respectively, 19.9, and 13.5 months, for the entire group of patients (Fig. 1). Figure 2 summarizes the assessment of best radiological response to the treatment for each group based on MacDonald criteria. Twenty-three patients (60%) progressed during treatment, for a mean TTP of 127 days. Interestingly, of 10 patients already exposed to systemic administration of drugs used in our regimen, 7 showed a response, 1 patient remained in stable disease (SD), and 2 progressed.
Lung Carcinoma (n = 18)
Lung metastases patients presented a median survival of 11.2 months. Two histological subgroups were identified and analyzed separately. Adenocarcinomas (n = 9) depicted a mean and median survival time of 15.3 and 13.5 months, respectively. The best response to treatment was 1 major partial response (MPR) and 4 partial response (PR). For small cell lung carcinoma (SCLC) patients (n = 9), we observed 1 complete response (CR), 2 MPR, and 3 PR, with 8.3 months of mean and 8.8 months of median survival (Fig. 3). As SCLC patients typically presented with multiple lesions and a significant mass effect related to peritumoral edema, all were treated with IA chemotherapy, without BBBD.
Lymphoma (n = 8)
Systemic lymphoma patients with brain metastasis presented an average and median survival of 24.8 and 16.3 months, respectively. Two patients completed the full treatment regimen (12 cycles), and are still in CR after 57.2 and 51.9 months (Fig. 4). Interestingly, 7 of 8 patients presented a response.
Ovarian Metastasis (n = 5)
This group showed an average and median survival of 33.6 and 42.3 months, respectively. These patients presented and all or non-behavior when studying radiological responses, in that 4 of them had a CR, whereas 1 progressed and died 8 months after treatment initiation. Two of these 5 patients are still in CR with a follow-up of 34.1 and 27.8 months (Fig. 5). None of the responders had been exposed to radiation therapy.
Breast Carcinoma (n = 4)
A mean and median survival of 8.9 and 8.1 months, respectively, were observed. Best radiological response observed was 1 PR. The 3 other patients progressed.
Complications and Adverse Events
For the present study, adverse events are classified in 2 distinct categories: hematologic toxicity and adverse events associated with the procedure itself. In terms of hematologic toxicity, we observed 1 grade 3 occurrence of anemia and 1 grade 4 thrombocytopenia according to the National Cancer Institute common toxicity criteria. Two occurrences of grade 3 and 1 of grade 4 neutropenia were also observed. Three adverse events related to the procedure were recorded. One occurrence of severe neck pain after perfusion of chemotherapy in the vertebral artery was observed in a patient treated with IA chemotherapy without BBBD. This syndrome was controlled with steroids. Two patients presented with a postinfusion orbital pseudotumor syndrome.17 Their condition completely recovered after high-dose steroids.
Brain metastases occur in 20% to 40% of cancer patients and are associated with a dismal prognosis. Typically, patients with multiple brain metastases are exposed to WBRT as a palliative measure.1, 2, 6 The rationale supporting the use of this approach relies on the demonstration that it has a high potency of controlling neurological symptoms.18 However, survival improvement is modest, ranging from 3 to 5 months.3, 5 This survival has not been altered despite 3 decades of clinical research aimed at improving outcome for these patients.6 The short median survival benefit has to be weighed against the possibility of inducing side effects such as fatigue, alopecia, and transient worsening of neurological symptoms.5, 6, 8 More worrisome are the risks of inducing delayed toxicity such as diffuse cerebral injury, producing progressive dementia, gait ataxia, and sphincter dysfunction in the few patients presenting a long-term survival.16 DeAngelis et al.9 described an incidence of radiation-induced dementia of 1.9% to 5.1% in patients with brain metastases. This was probably an underestimate, as only the most severely affected patients were identified and no formal neurocognitive testing was accomplished. These side effects need to be considered in light of the palliative context.
The role of standard chemotherapy in the treatment of brain metastases has always been marginal, as penetration of the chemotherapy agent beyond the BBB is limited.11 By way of its anatomic and physiologic properties, the normal BBB prevents passage of ionized water-soluble compounds with a molecular weight greater than 180 Da.19 Although the integrity of the barrier is often compromised within or around tumors, this alteration in permeability is variable and heterogeneous, nonetheless greatly limiting drug penetration.19 More so, drug distribution is uneven, with a preferential accumulation in the necrotic areas of the tumor.5 Drug penetration at the edge of the tumor is nonexistent, as has been shown in a study using fluorescein as a marker of permeability.20 The presence of the P-glycoprotein (P-gp) efflux pump at the luminal surface of the brain capillaries and the constant flow of cerebral spinal fluid (CSF) emanating from perivascular spaces further contribute to a decrease in the actual concentration and time of exposure of tumor cells to chemotherapy. Interestingly, despite the fact that this limitation in delivery imposed by the barrier is more and more acknowledged, it remains underdiscussed in the field of neurosciences.11, 12
Intraarterial Chemotherapy and BBBD
Different approaches have been advocated to improve delivery across the BBB. One such approach has been extensively tested in the clinic, and is commonly used by our team, both in the preclinical and clinical setting. It involves the IA infusion of chemotherapy with or without prior administration of a hypertonic solution to produce a transient increase in permeabilization of the barrier in a given cerebrovascular distribution.15, 16 The so-called IA chemotherapy infusion refers to the regional delivery of chemotherapy to the CNS, which produces a significant increase in plasma peak concentration and in the area under the curve related to the first pass effect.13 This translates in a significant increase in intra-tumoral chemotherapy concentration by a 3–5.5-fold factor, as most of these lesions are highly vascularized.19 Newton et al.13 reported their experience with IA chemotherapy in a group of 27 patients exposed to IA carboplatin and IV etoposide. All the patients in this series had been exposed to radiation before enrollment, and 67% had been exposed to chemotherapy. The authors reported a 54% response rate and a 20-week median survival from study initiation. This lead to an impressive TTP of 16 weeks in the cohort, thus validating this approach as promising in the treatment of this patient's population. Doolittle et al.21 reported their results on brain metastases treatment as part of a broader series. Although not specified in the article, most of those patients had been treated by IA chemotherapy without BBBD. Thirteen patients with brain metastasis of unspecified histology were so treated with the following results: 3 CR, 2 PR, and 5 SD. However, no details as to the specific histology in relation to response and duration of response are given for this subset of patients, thus greatly limiting the reach of the results.
One way to further increase delivery is to add an osmotic manipulation of the BBB to the IA chemotherapy infusion. The permeation of the BBB is obviously transitory (30 minutes to 2 hours according to experimental data), but lasts for a sufficient time to allow the IA infusion of a therapeutic molecule.19 The procedure would also allow a transient decrease in the function of the P-gp pump, thereby allowing for improved delivery of chemotherapy agents that are also known substrate of this pump.22 Even if the barrier is breached in the tumoral tissue, its integrity remains virtually intact at the periphery of the tumor.19 Sato et al.20 elegantly demonstrated in vivo that the mannitol infusion increases the permeability of the barrier at the edge of the tumor, in the peritumoral area. Recently, we reported encouraging results obtained with malignant glioma patients using the BBBD procedure.16 Fragmentary data were also available on a group of 10 patients with brain metastasis. The median survival time from study entry was 9 months. Interestingly, the extent of the increase in permeability has been shown to correlate with long-term survival in a population of lymphoma patients treated with the BBBD procedure combined with IA methotrexate infusion.23 In that study the authors clearly established a relation between the clinical outcome of the patient and the intensity of delivery obtained by the procedure, emphasizing the relevance of improving delivery strategies.
As stated earlier, the selection of patients directed toward BBBD was partially based on the presence of a significant mass effect, and some patients were denied BBBDs but were offered IA chemotherapy based on that criterion. Because BBBD produces a significant increase in interstitial brain water content, patients with a significant mass effect could be at risk for herniation, and thus strict criteria are applied to exclude these patients from BBBD.16 Unquestionably, this constitutes a huge bias factor in detriment of the IA cohort.
Recently, Newton et al.14 reported on the importance of performing a 4-vessel angiogram before selecting the appropriate vessel(s) for treatment. In a series of 78 patients (39 with metastasis), those authors showed a 6.4% alteration in treatment plan after having performed a 4-vessel angiogram, and not just a diagnostic angiogram of the presumed vessel(s) of interest, as we do. The fact that the relevant vessel might not be selected is worrisome, even if the risk is fairly low (6.4%). As this instance can be easily avoided by a 4-vessel angiography before the final planning of the treatment, this appropriate measure should be implanted in every study on IA with or without BBBD. We intend to do so in future trials.
The analysis of a heterogeneous group of patients as a whole, such as our series of brain metastasis, is always difficult and data can be influenced by many confounding factors. One way to limit these confounding factors is to stratify the patients according to validated prognostic factors. The recursive-partitioning analysis (RPA) of the RTOG databases allows this stratification along 3 different prognostic classes.4 According to these classes, median survival of patients with brain metastases ranges between 2 and 7.1 months (Table 7). As most of our patients (35 of 38, or 92.1%) fitted criteria of class 2 or 3, expected median survival for this group of patients would lean between 2.3 and 4.2 months. Thus, a mean and median survival of 19.9 and 13.5 months, respectively, compares favorably with this projection. It also allows us to preclude favorable survival biases as an explanation for these results.
Lung is the primary source of metastasis to the brain. SCLCs and adenocarcinomas have similar frequencies of seeding to the brain, averaging a 30% incidence.24 Overall, more than 50% of patients with brain metastases from lung cancer present with multiple lesions.25
A median survival of 4–8 months was reported for adenocarcinoma patients treated with either systemic chemotherapy regimens with and without prior WBRT.24 The 9 patients with adenocarcinoma enrolled in the present study presented 13.5 months of median survival from study entry. All these patients were treated at recurrence, after having failed radiation therapy. Only 1 patient is still alive at 20.7 months.
The European Organisation for Research and Treatment of Cancer (EORTC) Lung Cancer Cooperative Group evaluated the efficacy of WBRT as the sole treatment for the SCLC subgroup and found a median disease-free survival of only 5.4 months, even though the reported stabilization or improvement in neurological status was 60%.25 Systemic chemotherapeutic regimens have been used without prior WBRT with a median survivals ranging from 3 to 4 months.26 Some prospective trials for the treatment of SCLC brain metastases have found median survival of 3–6 months when these regimens were used after WBRT.24 In our series, a median survival of 8.8 months was obtained for this group, which demonstrates a limited advantage with respect to results obtained from the literature. Only 3 patients were treated with radiotherapy after having failed chemotherapy. From chemotherapy failure, these patients showed a median survival of 2.7 months.
Brain represents a rare site of metastasis for patients suffering from an ovarian carcinoma. Autopsy series reported an incidence of CNS metastasis of 2% to 12%. This site thus represents only 0.7% to 1.8% in a large series of patients with brain metastasis.27–29 Melichar et al.28 found that a combination of surgery, radiotherapy, and conventional chemotherapy was associated with a median survival of 20 months and that surgery with chemotherapy resulted in a median survival of 15 months. However, because of the small cohort size (2–12 patients), extrapolation of these intervals is limited.30 Our results are also marred by a small sample size of 5 patients, thus limiting the reach of our analysis. A median and mean survivals of 42.3 and 33.6 months, respectively, from study entry were obtained in these patients. More important, 2 of these 5 patients are still alive and in CR after having completed the entire course of treatment. All of the responders (n = 4) were exposed to chemotherapy alone, and thus were spared the side effects and discomfort related to radiation therapy. Even with a limited number of patients, we suggest that this treatment modality is an adequate substitute to radiation therapy in this particular histology, and should thus be considered up front.
Ten to 15% of patients with advanced breast cancer will eventually present with CNS involvement. CNS metastases would be a major contributing factor in the death of 68% of these patients.31 Patients are typically in an advanced stage of their disease and have already been exposed to multiple chemotherapy regimens.32 The same can be said of our patients, who had all been exposed to radiation therapy and to multiple chemotherapy regimens (average of 4).
Fenner and Possinger33 reviewed 8 trials on standard chemotherapy as treatment for brain metastasis of breast cancer. They concluded that median survival was 6 months and that this figure was similar to WBRT. Different investigators have tested temozolomide in pretreated and preirradiated patients and have obtained a median survival time of 4 to 7 months.5, 34 Similar results were produced in our series of 4 patients, with a median survival of 8.1 months. The best response obtained was 2 SD. One factor that could explain the poor results obtained in this group is the fact that all patients were heavily pretreated for their systemic disease before enrollment in our study. These results are in sharp contrast to the results published by Newton et al.,13 who reported a CR/PR in 4 of 9 patients with breast cancer. Thus, a larger and more diverse cohort will be needed to study the impact of this treatment in this class of patient.
Systemic Lymphoma With Brain Metastases
Parenchymal brain metastases from systemic lymphoma have been reported infrequently. In a series of 498 patients with malignant non-Hodgkin lymphoma, only 30 presented a secondary CNS involvement. The prognosis of this condition is typically poor, with a mean survival with treatment of 3.5 months.35 The patients in our series presented a median survival of 16.3 months, with 2 patients still in CR at 57.2 and 51.9 months. These results represent a significant improvement over previously published series, and this approach should be considered in this rare occurrence.
Doolittle et al.21 reported on the safety of this procedure, when performed in a standardized fashion, in a multicenter setup. Our data further support this contention.
Poor Response Factors
When reviewing the pretreatment radiological characteristics of each patient, we found 2 surrogates adversely impacting response: heavy tumor burden independent of the histological nature of the brain metastasis, and important bihemispheric edema. The nature of the edema may be vasogenic or postradiotherapy. All patients showing these features either alone or in combination achieved a poor or no response at all.
Even if the number of patients in each subgroup is small, the results obtained in this study seem promising for the following histologies: ovarian carcinoma, lung adenocarcinoma, SCLC, and lymphoma.
- 6Current strategies in whole-brain radiation therapy for brain metastases. Neurosurgery. 2005; 57: s4–33–44., .
- 21Safety and efficacy of a multicenter study using intra-arterial chemotherapy in conjunction with osmotic opening of the blood-brain barrier for the treatment of patients with malignant brain tumors. Cancer. 2000; 88: 637–647., , , et al.