Fax: (212) 717-3342
Phase II study of temozolomide and thalidomide in patients with metastatic melanoma in the brain†
High rate of thromboembolic events (CALGB 500102)
Version of Record online: 19 SEP 2006
Copyright © 2006 American Cancer Society
Volume 107, Issue 8, pages 1883–1890, 15 October 2006
How to Cite
Krown, S. E., Niedzwiecki, D., Hwu, W.-J., Hodgson, L., Houghton, A. N., Haluska, F. G. and for the Cancer and Leukemia Group B (2006), Phase II study of temozolomide and thalidomide in patients with metastatic melanoma in the brain. Cancer, 107: 1883–1890. doi: 10.1002/cncr.22239
The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute.
- Issue online: 3 OCT 2006
- Version of Record online: 19 SEP 2006
- Manuscript Accepted: 9 AUG 2006
- Manuscript Revised: 20 JUL 2006
- Manuscript Received: 9 MAY 2006
- National Cancer Institute. Grant Numbers: (CA31946), (CA77651), (CA33601), CA12449)
- malignant melanoma;
- brain metastases;
Preliminary studies suggesting that extended-dose temozolomide with thalidomide is safe and active in patients with metastatic melanoma have led to frequent use of this oral regimen. To confirm these observations the combination was tested in a multicenter Phase II trial in patients with melanoma brain metastases.
Eligible patients had melanoma brain metastases, with or without systemic metastases. The primary endpoint was response rate in brain metastases. Patients received temozolomide at a dose of 75 mg/m2/day for 6 weeks with a 2-week rest between cycles, and thalidomide (escalated to 400 mg/day for patients age <70 years or to 200 mg/day for patients age ≥70 years). A 2-stage design required ≥3 responses in the first 21 patients before enrolling 29 additional patients in the second stage.
Sixteen eligible patients were enrolled. No objective responses were observed. The median survival was 23.9 weeks. Seven patients withdrew because of tumor progression; 7 were removed during Cycle 1 because of adverse events, including allergic reaction (1 patient), severe fatigue (1 patient), sudden death (1 patient), and thromboembolic events (pulmonary embolism in 3 patients and deep vein thrombosis in 1 patient); 2 patients withdrew when the study was suspended and subsequently closed. No associations could be established between baseline characteristics and toxicity.
The proportion of patients with lethal or potentially life-threatening adverse events was high (0.31, 95% confidence interval, 0.11–0.59), and the absence of objective responses made it unlikely that further accrual would demonstrate the efficacy of the regimen. These observations provide little support for the use of this combination for melanoma brain metastases unless safe and effective methods to prevent thrombosis are developed. Cancer 2006. © 2006 American Cancer Society.
Malignant melanoma frequently metastasizes to the brain. When it does, median survival is short. Various factors, which include age; gender; serum lactate dehydrogenase (LDH) level; performance status; the sites, numbers, and volume of brain metastases; the presence or absence of extracranial metastases and neurological signs and symptoms; and the type of treatment have each been associated in ≥1 studies with outcome.1–3 Most treatments have been associated with only modest prolongation of survival compared with corticosteroid treatment alone, and more effective treatments are needed.
Temozolomide, an orally bioavailable alkylating agent with excellent central nervous system (CNS) penetration,4, 5 has shown activity in certain CNS tumors6 and was approved by the U.S. Food and Drug Administration for the treatment of glioblastoma multiforme and anaplastic astrocytoma. It has also shown activity in some patients with melanoma brain metastases. In a study of 151 patients who did not require immediate radiotherapy, 8 of 117 patients (7%) who had received no prior systemic chemotherapy and 1 of 34 patients who had received prior systemic chemotherapy for melanoma showed objective responses.7 The median overall survival was 3.5 months in those patients previously untreated with chemotherapy and 2.2 months in those previously treated. In another study, 6 of 25 patients (24%) showed partial response of brain metastases after treatment with temozolomide alone or combined with docetaxel or cisplatin.8 In that study, the median survival was 4.7 months. It has also been suggested that temozolomide may inhibit subsequent development of brain metastases in patients with extracranial melanoma.9, 10
In prior clinical trials, temozolomide was administered on 5 consecutive days, every 28 days, at a dose of 150 to 200 mg/m2/day. It was hypothesized, however, that lower daily doses administered over an extended period might result in greater efficacy by depleting the DNA repair protein, O6-alkyl guanine DNA-alkyltransferase, and providing greater cumulative drug exposure.11–13 In addition, Hwu et al.14 reasoned that additional benefit might be gained by combining temozolomide with thalidomide, an agent with angiogenesis-inhibitory and immunomodulatory activities.15 In Phase I and II trials in patients with metastatic melanoma without brain metastases, objective response rates of 42% and 32%, respectively, were documented when temozolomide was administered in 8-week cycles at a dose of 75 mg/m2/day for 6 weeks, followed by a 2-week break, together with thalidomide given daily for 8 weeks.14, 16 These high response rates, together with the good safety profile and ease of administration reported for this regimen, has led many practicing oncologists to use this combination to treat advanced melanoma. This same regimen was also used to treat 26 patients with melanoma brain metastases.1 Of 15 patients who completed 1 full treatment cycle, 3 showed an objective response. However, many patients did not complete a full treatment cycle because of adverse events that included CNS hemorrhage (7 patients), deep vein thrombosis (DVT) or pulmonary embolism (PE) (3 patients), and severe rash (1 patient).1 The median survival of all 26 patients in this study was 5 months.
Given the poor prognosis and limited treatment options available, together with these preliminary data, the Cancer and Leukemia Group B (CALGB) initiated a multicenter, Phase II trial of the combination of extended-dose temozolomide and thalidomide in patients with malignant melanoma metastatic to the brain.
MATERIALS AND METHODS
All patients gave written informed consent in accordance with human experimental guidelines of the Department of Health and Human Services and the Human Investigations Committees at each of the participating sites. Eligible patients were at least 18 years old, had a National Cancer Institute Common Toxicity Criteria (CTC) performance status (PS) of 0 or 1, and a histologic or cytologic diagnosis of metastatic melanoma with at least 1 measurable brain metastasis on computed tomography (CT) or magnetic resonance imaging (MRI) scan according to RECIST criteria.17 Patients with or without extracranial disease were eligible to participate. Patients were permitted to have received no more than 1 prior chemotherapy regimen, but were not permitted to have received systemic chemotherapy for brain metastases or a regimen that contained temozolomide on a continuous (>5 consecutive days) dosing schedule. Prior immunotherapy, surgery for brain metastases, or whole-brain or stereotactic radiation for brain metastases was allowed. Patients who had received brain radiation were required to have progressive disease in at least 1 measurable brain metastasis, or to have developed a new measurable brain metastasis. Before entry, at least 3 weeks must have elapsed since stereotactic radiotherapy or surgery, and at least 4 weeks since whole-brain radiation, cytokine therapy, or chemotherapy (6 weeks for nitrosoureas or mitomycin C).
Patients were excluded for conditions that would interfere with oral intake; history of active angina or myocardial infarction within 6 months; significant ventricular arrhythmia or uncontrolled cardiac arrhythmia; history of DVT or PE; known human immunodeficiency virus (HIV) infection; neuropathy ≥Grade 2; uncontrolled seizures; use of warfarin, heparin, or their derivatives; requirement for daily aspirin, ibuprofen, clopidogrel, or other antiplatelet therapy; or requirement for bisphosphonates. Pregnant and nursing women were excluded. Women of childbearing potential were required to agree to either abstain from sexual intercourse or use 2 methods of birth control for 28 days before treatment, during the entire course of thalidomide treatment, and for 28 days after the last thalidomide dose. Men were required to agree to abstain from unprotected sexual intercourse and to request that their female partners use a second method of birth control in addition to the male barrier method.
Patients were also required to have granulocyte count ≥1500/μL; a platelet count ≥100,000/μL; serum creatinine ≤2 mg/dL; serum alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, and LDH ≤2.5 times the upper limit of normal; thyroid-stimulating hormone within normal limits; and electrocardiogram showing no acute abnormality or uncontrolled arrhythmia. Women who could become pregnant were required to have a negative serum β-human chorionic gonadotropin (β-HCG). Patients receiving anticonvulsants were required to have serum drug levels within the therapeutic range.
Each treatment cycle was 8 weeks long. All patients received temozolomide at a dose of 75 mg/m2/day for 6 weeks, followed by 2 weeks without temozolomide. Patients age <70 years had therapy with thalidomide initiated at a dose of 200 mg/day; the dose was escalated in 100-mg increments every 2 weeks to a maximum of 400 mg/day according to tolerance. Patients age ≥70 years began thalidomide at a dose of 100 mg/day, which was escalated in 50-mg increments every 2 weeks to a maximum of 200 mg/day according to tolerance. All drugs were given in the evening. Thalidomide was given 30 to 60 minutes before temozolomide. Additional oral antiemetics were given as needed. Patients were instructed to use laxatives and stool softeners to prevent constipation. Patients were to receive continued cycles of therapy until tumor progression (intracranial or extracranial) was documented or dose-limiting toxicity developed. Toxicity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE), version 3.0.
Dose modification guidelines for temozolomide included withholding the drug for any Grade 4 hematologic toxicity or Grade 3 or 4 nonhematologic toxicity. If the toxicity resolved to ≤Grade 1 within 4 weeks, temozolomide was to be restarted at a dose of 50 mg/m2/day; if the toxicity did not resolve within 4 weeks, study treatment was to be permanently discontinued.
Dose modification guidelines for thalidomide included gradual escalation of the dose at 2-week intervals during the first treatment cycle unless Grade 2 rash or Grade 3 or 4 nonhematologic toxicity developed. Subsequent thalidomide dosing was to be withheld for graded neurologic, dermatologic, and gastrointestinal toxicities and resumed at the next lower dose if the toxicity resolved to ≤Grade 1 within 4 weeks. Patients who developed toxic epidermal necrolysis and patients who developed DVT or PE or required anticoagulation were required to have all study treatments permanently discontinued.
Tumor responses at all metastatic sites were to be assessed at the end of each 8-week treatment cycle using previously described criteria.17 Physical examinations and blood tests to monitor safety and toxicity were scheduled at Weeks 2 and 4 and then every 4 weeks.
Statistical Analysis and Data Management
The primary endpoint of the study was response (complete or partial) in brain metastases. The trial was designed to test the null hypothesis that the response rate was ≤10% versus the alternative that it was ≥25% in 2 stages with 21 patients and 29 patients entered at each stage, respectively. Early stopping for low response would occur if ≤2 responses were observed in the first 21 patients. The regimen would be considered a success after Stage 2 if ≥8 responses were observed among the 50 patients studied. The power and significance level under this design were .9 and .1, respectively. The protocol included an early stopping rule for excess neurotoxicity. Descriptive statistics and time-to-event estimates are based on all patients studied unless otherwise indicated. Progression-free survival was measured from study entry until documented progression of disease or death from any cause; overall survival was measured from study entry until death from any cause.
Patient registration, data collection, and data analysis were performed by the CALGB Statistical Center. Data quality was ensured by careful review of data by CALGB Statistical Center staff and the study chairperson.
As part of the CALGB quality assurance program, members of the Data Audit Committee visit all participating institutions at least once every 3 years to review source documents. The auditors verify compliance with federal regulations and protocol requirements, including those pertaining to eligibility, treatment, adverse events, tumor response, and outcome in a sample of protocols at each institution. Such onsite review was performed for a subgroup of 3 patients (18.5%) of the 16 patients treated under this study.
The study was activated on October 15, 2003; accrual was temporarily suspended on February 9, 2005 because of unexpected thromboembolic adverse events. After review of the adverse events, the trial was subsequently closed to accrual on March 15, 2005. A total of 16 eligible patients (7 women and 9 men) were enrolled. The targeted accrual of 21 patients to Stage 1 was not reached.
The median age of the patients was 54.5 years (range, 23–83 years). Thirteen patients were age <70 years. A PS of 0 was reported in 9 patients and a PS of 1 was reported in 7 patients.
Four patients had metastases confined to the brain and 12 also had extracranial metastases. The median number of brain metastases was 3 (range, 1–9 metastases). The median number of extracranial disease sites was 3 (range, 0–13 sites). Eight patients were receiving corticosteroids at the time of study entry, and 1 woman was taking oral contraceptives. Three patients had received chemotherapy or interferon previously for metastatic melanoma and in 8 patients brain metastases had been treated with radiation (whole brain or stereotactic) and/or surgery. The median time from diagnosis of brain metastases to registration on study was 2.2 months (range, 0.4–8.4 months).
Response, Survival, and Toxicity
The median treatment duration was 4.7 weeks (range, 0.3–14.6 weeks). Only 4 patients received at least 7 weeks of treatment and 2 others received at least 6 weeks. No treatment responses were observed. Using an intent-to-treat analysis (n = 16), the exact 95% confidence interval (95% CI) estimates for the response rate was given by (0.0–0.21). The probability of at least 3 responses being observed in the next 5 patients studied at Stage 1 was .1 under the alternative hypothesis (P = .25). Two patients voluntarily withdrew during the first cycle, 1 at the time of study suspension and the other at the time of permanent closure to accrual, before their responses could be assessed. Neither had experienced dose-limiting toxicity after 3.9 weeks and 7.3 weeks, respectively. Seven patients were removed from study after a median of 7.9 weeks (range, 1.4–14.9 weeks) because of tumor progression in the brain (5 patients), extracranial sites (1 patient), or both (1 patient). The remaining 7 patients were removed because of graded adverse events after a median of 4.7 weeks (range, 2.7–6.7 weeks) (in some patients, the time on the study was slightly longer than time on treatment, because the decision to permanently discontinue treatment was made after temporarily withdrawing study drugs). Adverse events leading to treatment discontinuation included a Grade 3 allergic reaction in 1 patient (facial edema and itching after 3 days of therapy), severe fatigue in 1 patient (Day 40), and sudden death (Day 27). In the latter patient, a head MRI performed on Day 16 to evaluate headaches demonstrated no evidence of hemorrhage, and a lower extremity Doppler ultrasound performed on Day 15 to evaluate mild leg swelling did not reveal DVT. Four other patients were removed during the first cycle because of PE (3 patients) or DVT (1 patient), which occurred after a median of 25 days (range, 11–40 days). One patient with PE had concurrent MRI evidence of acute intracranial bleeding. Additional Grade 3 toxicities unrelated to the thrombotic events occurred in 1 patient each, and included lymphopenia, thrombocytopenia, elevated prothrombin time, and infection.
As of November 30, 2005, 11 patients had died after a median of 11.3 weeks (range, 1.7–36.3 weeks) and 5 patients were alive with disease after a median of 31.7 weeks (range, 3.9–40.3 weeks) of follow-up. The median overall survival for all 16 patients was 23.9 weeks (range, 1.7–40.3 weeks) (Fig. 1). The median progression-free survival (progression or death) was 7.3 weeks (range, 1.43–40.3 weeks) (Fig. 2).
Correlates of Toxicity
The high rate of thromboembolic events (4 patients) and the single sudden death (a postmortem examination was not performed, but possible causes of death included PE, arrhythmia, myocardial infarction, intracranial hemorrhage, and brain herniation) prompted an analysis of potential contributing factors (Table 1). Because of the small numbers of patients in the event subgroup, we descriptively compared baseline characteristics of these 5 patients with those of the 11 patients without these events. There were no obvious differences in known risk factors between the patients with and without these events. Patients with thromboembolism or sudden death had a somewhat shorter median time from the diagnosis of brain metastases to registration on study than those without these events, which is consistent with their lower rate of prior surgery or radiation therapy for brain metastases. A higher percentage of patients with thromboembolism/sudden death had received prior systemic therapy for melanoma than those without these events.
|Thromboembolic event or sudden death|
|Yes (n = 5)||No (n = 11)|
|Median age, y||44||55|
|No. of patients age ≥70 y||1 (20%)||2 (18%)|
|Time from diagnosis of brain metastases to registration, mo||Median: 0.7||Median: 2.7|
|Range: 0.4–3.9||Range: 0.4–8.4|
|Mean (SD): 1.3 (1.5)||Mean (SD): 3.1 (2.3)|
|No. receiving dexamethasone||3/5 (60%)||5/11 (45%)|
|No. receiving oral contraceptives (women only)||1||0|
|No. receiving erythropoietin*||0||0|
|Prior radiation or surgery for brain metastases||1 (20%)||7 (64%)|
|Prior systemic therapy||2 (40%)||1 (9%)|
|Performance status 1||2 (40%)||5 (45%)|
|Neurologic symptoms present||2 (40%)||4 (36%)|
|Extracranial metastases present||4 (80%)||8 (73%)|
|LDH > ULN||1 (20%)||4 (36%)|
|No. of brain metastases||Median: 2||Median: 5|
|Range: 2–7||Range: 1–9|
|Mean (SD): 3.2 (2.2)||Mean (SD): 4.6 (3.0)|
|Time on treatment, wk||Median: 3.9||Median: 5.6|
|Range: 2.9–5.9||Range: 0.3–14.6|
|Progression-free survival, wk||Median: 6.6||Median: 8.0|
|Range: 4.4–40.3||Range: 1.4–14.1|
|Overall survival, wk||Median; 8.7||Median: 24.3|
|Range: 4.4–40.3||Range: 1.7–39.9|
Of the 4 patients with documented thromboembolic events, 3 later died after a median of 8.7 weeks on study, but none of the deaths was attributed to the adverse event. Time on treatment, progression-free survival, and overall survival appear similar between the groups with and without thromboembolic events. Again, the small numbers in each subgroup provide little statistical power with which to compare these endpoints.
Although previous reports suggested that temozolomide, alone or in combination with thalidomide, was well tolerated and could induce regression of melanoma brain metastases,1, 7, 8 our study was stopped prematurely because of concerns regarding safety and efficacy. After only 16 of the planned initial group of 21 patients had been accrued, 5 had experienced a DVT, PE, or sudden death during the first treatment cycle. Not only was the proportion of patients with these life-threatening or lethal adverse events unacceptably high (0.31; 95% CI, 0.11–0.59), but the absence of objective tumor responses also made it unlikely that further accrual would demonstrate efficacy of the regimen according to the predefined criteria. Moreover, we were not able to identify baseline characteristics that would allow us to exclude patients at increased risk for these adverse events. Nonetheless, the overall median survival was 23.9 weeks.
Several factors could have contributed to the thrombotic events we observed. Cancer itself has long been associated with an increased risk of thromboembolic events (TEE).18–20 The mechanisms responsible include endothelial damage mediated by cytokines, activation of platelets and clotting factors, and production of procoagulants by tumor cells.18, 19 Although it has been possible to identify cancer patients demonstrating laboratory evidence reflecting activation of coagulation, to our knowledge these have not been predictive of subsequent TEE.21 Brain tumors are associated with a particularly high risk of TEE. As recently reviewed,22 brain tissue is a rich source of tissue factor, which may be released into the systemic circulation with resultant activation of the coagulation cascade. Patients with brain tumors also demonstrate abnormalities of the fibrinolytic system.
In addition, cancer treatments may increase the risk of thrombosis. High rates of TEE have been reported in patients with nonsmall cell lung cancer,23 germ cell tumors,24 and urothelial cancers25 who are treated with cisplatin-based chemotherapy. Among women with breast cancer, adjuvant chemotherapy has been associated with a 3-fold to 4-fold increase in the risk of TEE, and tamoxifen increased this risk.26–34 Although the mechanisms responsible for the increased risk of TEE associated with cancer therapies are poorly understood, they most likely are similar to those proposed for cancer itself.18, 19 To our knowledge, there have been no reports of an increased risk of TEE associated with administration of temozolomide alone, but detection of an increased risk in patients with primary brain tumors, in whom there is the greatest experience with temozolomide, may have been obscured by their already high risk of thrombotic events.22 Likewise, there is no evidence that thalidomide by itself is associated with an unusually high risk of TEE.35, 36 It has been suggested, however, that TEE are underrecognized and underreported as a complication of cancer chemotherapy.23
Although TEE have been observed previously in melanoma patients who received the combination of temozolomide and thalidomide, the reported incidence was lower than we observed. In a Phase I study in 12 patients without brain metastases, 1 patient developed DVT and PE after prolonged car travel.14 In a Phase II study in similar patients, only 1 case of DVT was reported among 38 patients.16 A 12% incidence of TEE was reported among 26 patients with melanoma brain metastases who received temozolomide with thalidomide at the same dose and schedule used in this trial.1 Despite these findings and the absence of evidence that the components of the temozolomide and thalidomide regimen are individually associated with an increased risk of TEE, there is ample evidence, primarily from studies in multiple myeloma37–40 (but also other tumor types39, 41–43), that the combination of thalidomide with dexamethasone or chemotherapy markedly increases this risk. In myeloma patients receiving thalidomide, the use of an anthracycline-containing regimen38 and acquired activated protein C resistance44 were associated with a particularly high risk of TEE.
Given its widespread use for melanoma treatment, oncologists need to be aware that the temozolomide and thalidomide combination may be associated with a higher rate of potentially life-threatening TEE than has previously been reported. These complications may be most prominent in patients with brain metastases, but others receiving this regimen should be monitored carefully for these events. Our findings also raise the question of whether preventive strategies might decrease the risk of TEE. For multiple myeloma, a disease in which the benefits of thalidomide are well documented, prophylactic anticoagulation is now routinely used as part of combination regimens.45 For patients with brain metastases, however, particularly melanoma metastases that are likely to be highly vascular and that often hemorrhage spontaneously, the use of anticoagulants is controversial. Although the risk of hemorrhage has long been considered to be a relative contraindication to systemic anticoagulation in patients with CNS tumors, several studies concluded that therapeutic anticoagulation is relatively safe and more effective than inferior vena cava filter placement in brain tumor patients with DVT.22, 46–48 None of these studies included sufficient numbers of patients with melanoma brain metastases, however, to assess the safety of therapeutic anticoagulation, nor is anything known about the tolerance or efficacy of prophylactic anticoagulation in such patients. Given the high rate of spontaneous intracranial hemorrhage (7 of 26 patients) observed during an earlier study of this combination in patients with melanoma brain metastases,1 prophylactic anticoagulation would be hard to justify.
Because patients with brain tumors have a high risk of TEE, it is possible that the events we observed in this study could have occurred in the absence of treatment. We consider this unlikely, however, as all the events occurred within a short interval after starting treatment. Given the small number of patients in this study, our conclusions can be considered tentative. Nonetheless, these events, together with the absence of antitumor activity and the low overall tumor response rate observed in the earlier study of this combination,1 provide little support for further investigation of the combination of temozolomide and thalidomide for patients with melanoma brain metastases unless safe and effective methods to prevent TEE are developed.
We thank Dr. Hani Hassoun for helpful comments.
- 4Pharmacokinetic study of temozolomide penetration into CSF in a patient with dural melanoma. Ann Oncol. 1998; 9( Suppl 4): 138. Abstract 659a., , , et al.
- 5Cerebrospinal fluid levels of temozolomide as a surrogate marker for brain penetration. Proc Am Soc Clin Oncol. 2001; 20: 59a. Abstract., , , , , .
- 27Increased thromboembolic complications with concurrent tamoxifen and chemotherapy in a randomized trial of adjuvant therapy for women with breast cancer.National Cancer Institute of Canada Clinical Trials Group Breast Cancer Site Group. J Clin Oncol. 1996; 14: 2731–2737., , , , , .
- 41A phase II trial of weekly intravenous gemcitabine with prolonged continuous infusion fluorouracil and oral thalidomide in patients with metastatic renal cell carcinoma. Proc Am Soc Clin Oncol. 2001; 20: 2448. Abstract., , , et al.