- © 2004 by American Society of Clinical Oncology
Temozolomide for the Treatment of Brain Metastases Associated With Metastatic Melanoma: A Phase II Study
- Sanjiv S. Agarwala,
- John M. Kirkwood,
- Martin Gore,
- Brigitte Dreno,
- Nicholas Thatcher,
- Beate Czarnetski,
- Michael Atkins,
- Antonio Buzaid,
- Dimosthenis Skarlos and
- Elaine M. Rankin
- From the University of Pittsburgh Cancer Institute, Pittsburgh, PA; the Royal Marsden, National Health Service Trust, London, and Christie Hospital, National Health Service Trust, Withington, Manchester, United Kingdom; Clinique Dermatologique, Nantes Cedex, France; Virchow Klinikum, Universitat zu Berlin, Berlin, Germany; New England Medical Center, Boston, MA; Hospital Sirio-Libanês, São Paulo, Brazil; Agioi Anarguroi Hospital, Athens, Greece; and Netherlands Cancer Institute, Amsterdam, the Netherlands
- Address reprint requests to Sanjiv S. Agarwala, MD, Division of Hematology Oncology, University of Pittsburgh Medical Center Pavilion, 5150 Centre Ave, Pittsburgh, PA 15232; e-mail: agarwalass{at}msx.upmc.edu
Abstract
Purpose Temozolomide is a well-tolerated oral alkylating agent with activity in the CNS. A multicenter, open-label, phase II study was conducted to assess the safety and efficacy of temozolomide in patients with brain metastases from metastatic melanoma (MM) who did not require immediate radiotherapy.
Patients and Methods Eligible patients had histologically confirmed MM to the brain, and no prior radiotherapy or radiosurgery for brain metastases. Previously untreated patients received temozolomide at 200 mg/m2/d × 5 days; previously treated patients received 150 mg/m2/d × 5 days every 28 days. Treatment continued for 1 year or until disease progression or unacceptable toxicity.
Results Of 151 patients enrolled, 117 had received no prior systemic chemotherapy, and 34 had received prior chemotherapy for MM. Among previously untreated patients, 25% had more than four brain lesions, eight (7%) achieved an objective response (one complete and seven partial), and 34 (29%) had stable disease in brain metastases. Median overall survival was 3.5 months. Among previously treated patients, 21% had more than four brain lesions, one had a partial response, and six (18%) had stable disease in brain metastases. Median overall survival was 2.2 months. Temozolomide was well tolerated, with four (3%) patients discontinuing because of adverse events. Grade 3/4 hematologic toxicities included thrombocytopenia (3%), neutropenia (2%), and leukopenia (1%). Headache (9%) and vomiting (8%) were the most common nonhematologic grade 3/4 adverse events.
Conclusion Temozolomide was well tolerated and demonstrated activity in the treatment of brain metastases from MM. Further evaluation of temozolomide combination therapy is warranted.
INTRODUCTION
Advanced metastatic melanoma (MM) is associated with a poor prognosis, and effective treatment options are limited. Patients with stage IV melanoma generally have a median survival of only 6 to 9 months, and a low probability (10% to 20%) of 5-year survival.1 The CNS is a common site of disease involvement; up to 75% of patients with MM develop brain metastases during the course of their disease.2 In fact, patients with MM who respond to aggressive systemic therapy often relapse with metastases in the CNS.
Once patients develop brain metastases, treatment is palliative. Surgery and radiosurgery can produce effective palliation in selected cases but are usually restricted to patients with solitary lesions. Radiation therapy is the current standard of care for multiple brain metastases; it can improve neurologic symptoms but does not alter disease outcome. Chemotherapy has been studied in patients with brain metastases associated with MM, but objective responses are rarely observed, and median survival is short.3–5
The failure of systemic therapy for MM may be related to the inability of commonly used chemotherapeutic and biologic agents to penetrate the CNS effectively. Temozolomide is a novel oral alkylating agent with excellent CNS penetration6 that has demonstrated activity in patients with recurrent high-grade gliomas and with MM.7–10 A randomized trial comparing temozolomide with dacarbazine (DTIC) in patients with previously untreated American Joint Committee on Cancer stage IV melanoma showed response rate and survival with the two drugs to be equivalent, with temozolomide-treated patients demonstrating an improved quality of life.9 Unfortunately, the effect of temozolomide on CNS lesions was not evaluated in this trial. In patients with brain metastases from lung and other tumors, objective responses to temozolomide and improvements in neurologic status have been observed.11,12 In patients with melanoma, it has been suggested that temozolomide may reduce the risk of CNS relapse compared with DTIC-based therapy.13,14 A retrospective study compared the incidence of CNS relapse in 41 patients who responded to temozolomide versus those who responded to DTIC. At a median follow-up of 19 months, there were significantly fewer patients with CNS relapse in the temozolomide group than in the DTIC group (two v nine patients).13 In addition to demonstrating activity in brain metastases, oral temozolomide is well tolerated and convenient to administer, does not require dietary restrictions, and may improve patient quality of life.
To further evaluate the safety and efficacy of temozolomide in brain metastases from melanoma, a multicenter, international study of temozolomide was conducted in the United States and Europe.
PATIENTS AND METHODS
Patients
Patients were enrolled if they had histologically confirmed melanoma and newly diagnosed brain metastases documented by gadolinium-enhanced magnetic resonance imaging (Gd-MRI), but did not require immediate radiation therapy to relieve symptoms. Eligible patients also met the following criteria: age ≥ 18 years; WHO performance status of 0 or 1; life expectancy of 12 weeks or longer; and adequate hepatic function (serum bilirubin < 1.5× the upper limit of normal [ULN] and serum ALT and serum AST ≤ 3× the ULN), renal function (serum creatinine < 1.5× the ULN), and hematologic values (absolute neutrophil count [ANC] ≥ 1,500 cells/μL, hemoglobin ≥ 10 g/dL, platelet count ≥ 100,000/μL). Steroids were permitted as long as the dose was stable for at least 7 days before enrollment. Previous therapy (ie, chemotherapy and/or immunotherapy) for systemic extracranial disease was permitted, and patients were stratified according to previous exposure to systemic chemotherapy.
Patients were ineligible if they had been previously treated for brain metastases with radiotherapy, interstitial radiotherapy, or stereotactic radiosurgery; had received prior treatment with temozolomide; or had received immunotherapy, biologic therapy, or chemotherapy (excluding nitrosoureas, vincristine, or mitomycin C) within 4 weeks, vincristine within 2 weeks, or a nitrosourea or mitomycin C within 6 weeks before temozolomide administration. All acute toxicities from previous therapies must have resolved. Patients with known carcinomatous meningitis, intractable vomiting, or any medical condition that could interfere with oral drug administration were also excluded. All patients provided written informed consent.
Study Design and Treatment
This international, multicenter, open-label, phase II study stratified patients into two groups: patients who had received no prior systemic chemotherapy and patients who had received prior systemic chemotherapy for MM outside the CNS. Patients who had not received prior chemotherapy were administered oral temozolomide 200 mg/m2/d × 5 days every 28 days. Patients who had received prior chemotherapy were administered oral temozolomide 150 mg/m2/d × 5 days every 28 days. Daily doses were rounded to the nearest 5 mg. Treatment was continued for a maximum of 1 year or until disease progression or unacceptable toxicity. Antiemetics were administered at the discretion of the treating physician.
The primary end points of the study were response in brain metastases and toxicity; each group was evaluated separately. Secondary end points included progression-free and overall survival, and time to progression of brain metastases.
Toxicity and Response Assessments
Toxicities were graded according to the National Cancer Institute's Common Toxicity Criteria. Patients were monitored for toxicity on day 1 of each treatment cycle, with a complete history and physical examination that included a neurologic examination, vital signs, a laboratory assessment (including CBCs and serum chemistry), and urinalysis. In addition, a CBC was performed on day 22 of each treatment cycle.
Gd-MRI and systemic computed tomography scans were performed at baseline and at the completion of every even-numbered treatment cycle to assess response in brain metastases and extracranial sites of disease. All radiologic studies were reviewed and independently confirmed by a central radiology facility, which was blinded to patient characteristics and stratification group. For responding patients, radiologic studies of all involved sites were performed 4 weeks later to confirm response.
Levels of response, both intra- and extracranial, were determined separately at two consecutive examinations at least 4 weeks apart according to the following criteria: a complete response (CR) was defined as the disappearance of all clinically detectable malignant disease; a partial response (PR) was defined for bidimensionally measurable disease as a ≥ 50% decrease in the sum of the products of the largest perpendicular diameters of all measurable lesions, and for unidimensionally measurable disease as a ≥ 50% decrease in the sum of the largest diameters of all lesions; and stable disease was defined for bidimensionally measurable disease as a less than 50% decrease or a less than 25% increase in the sum of the products of the largest perpendicular diameters of all measurable lesions, and for unidimensionally measurable disease as a less than 50% decrease or a less than 25% increase in the sum of the diameters of all lesions. Each of these response categories was valid only in the absence of development of new CNS lesions.
Dose Modifications
Dose modifications were required in the presence of significant hematologic or nonhematologic adverse events. For patients with an ANC of less than 1,500 cells/μL (≥ grade 2) or a platelet count of less than 100,000/μL (≥ grade 1), temozolomide administration was delayed ≤ 3 weeks until recovery to an ANC of ≥ 1,500 cells/μL and platelet count of ≥ 100,000/μL. Once these criteria were met, a 25% dose reduction was indicated for patients who had experienced an ANC nadir of less than 1,000 cells/μL or a platelet nadir of less than 50,000/μL. Patients requiring dose level reductions to less than 100 mg/m2/d were discontinued from the study. If after 3 weeks patients' ANCs remained less than 1,500 cells/μL or their platelet counts remained less than 100,000/μL, they were discontinued from study treatment. For patients with nonhematologic grade 2, 3, and 4 toxicities, the temozolomide dose was withheld until the toxicity resolved to no greater than grade 2 (or to baseline for transaminase and alkaline phosphatase levels), and then temozolomide was restarted at the same dose level or a reduced dose at the investigator's discretion.
Statistical Methods
Patients were accrued in this study using a two-stage design15 to optimize the probability of observing a real response. Statistical analyses for baseline demographics, response rates, and adverse events were descriptive. Survival rates were estimated using the Kaplan-Meier method.16
RESULTS
Patient Characteristics
A total of 155 patients were enrolled between September 1996 and March 1999. Four patients did not meet protocol eligibility requirements and never received study medication; these patients were excluded from all analyses. Baseline patient and disease characteristics for the 151 eligible patients included in the efficacy and safety analysis are shown in Table 1. Among 117 patients who had not received prior chemotherapy, 23 (20%) received prior immunotherapy. Among 34 patients who had received prior chemotherapy for systemic extracranial disease, 21 (62%) also received prior immunotherapy. The median time from diagnosis of melanoma to development of brain metastasis was 2 years. The median number of brain lesions was three in both groups (range, one to 15 for previously untreated patients; one to eight for previously treated patients). Approximately 40% of patients had ≤ two brain lesions.
Patient Demographics and Baseline Disease Characteristics
Response
Objective tumor response in brain metastases is presented in Table 2. Among the 117 patients who had received no prior chemotherapy, one patient (1%) had a CR; seven patients (6%) had a PR, and 34 patients (29%) had stable disease in the brain. Median duration of stable disease was two cycles (range, one to eight cycles). Therefore, among patients with no prior chemotherapy, 42 (36%) achieved clinical benefit. Among the 34 patients who had received prior chemotherapy, one patient achieved a PR, and six patients (18%) had stable disease in the brain. Median duration of stable disease was one cycle (range, one to three cycles).
Summary of Best Response to Temozolomide in Brain Metastases
The Kaplan-Meier estimate of progression-free survival for each group is shown in Figure 1. Median progression-free survival was 1.2 months (range, 0 to 10.9 months) for patients who had not received prior chemotherapy and 1.0 month (range, 0 to 6.1 months) for patients who had received prior chemotherapy. Kaplan-Meier estimates of overall survival are shown in Figure 2. Median overall survival was 3.5 months (range, 0 to 22.9 months) for patients who had not received prior chemotherapy and 2.2 months (range, 0 to 41.8 months) for patients who had received prior chemotherapy. The median overall survival for the entire eligible patient population (N = 151) was 3.2 months (range, 0 to 41.8 months).
Kaplan-Meier estimate of progression-free survival for patients who had not received prior chemotherapy (n = 117) and for patients who had received prior chemotherapy (n = 34).
Kaplan-Meier estimate of overall survival for patients who had not received prior chemotherapy (n = 117) and for patients who had received prior chemotherapy (n = 34).
Safety
The most frequently reported adverse events by severity and treatment arm are summarized in Table 3. The most frequently reported adverse events in patients who had not received prior chemotherapy were headache (50%), nausea (39%), and vomiting (39%). In patients who had received prior chemotherapy, the most frequent adverse events were headache (32%), nausea (32%), asthenia (29%), and pain (29%). Among 151 treated patients, 71 (47%) reported grade 3/4 adverse events; the most common were headache (9%) and vomiting (8%). Hematologic grade 3/4 adverse events were rare, including thrombocytopenia in four patients (3%) and neutropenia in three patients (2%).
Summary of Adverse Events by Severity
The number of patients with a hematologic nadir event and their respective median time to recovery, are summarized in Table 4. The thresholds for nadirs were defined as the following: hemoglobin, 10 g/dL; ANC, 1,500 cells/μL; platelet count, 100,000/μL; and WBC count, 3,000 cells/μL. Hemoglobin nadirs were the most common hematologic nadir events, occurring in 48 patients (32%). In addition, hemoglobin nadirs had the longest median recovery time (22 days). Platelet, WBC, and differential neutrophil count nadir events occurred in 15% to 25% of patients and had short median recovery times (7 to 9 days). During the study, eight patients (5%) had dose reductions, and seven patients (5%) had dose delays, primarily because of hematologic toxicity.
Hematologic Nadir Events
Disease progression was the most common reason for treatment discontinuation (74% of cases). Only four patients discontinued study treatment because of adverse events. One patient discontinued because of grade 3 abdominal pain, possibly related to study drug; one patient because of grade 4 thrombocytopenia, considered related to study drug; one patient because of grade 3 pain from a fall unrelated to study drug; and one patient because of grade 3 headache and hydrocephalus unrelated to study drug. Of the 32 reported deaths during the study, one was from sepsis (considered related to study drug), and one was from cerebral bleeding (considered probably related to study drug).
DISCUSSION
Patients with melanoma brain metastases have limited treatment options and a poor prognosis. Survival depends on the extent of cranial disease and the degree to which it can be controlled. Whole-brain radiotherapy is the standard of care in patients with multiple or inoperable brain metastases, and radiosurgery is becoming an increasingly attractive option for patients with up to four lesions.17,18 Chemotherapy options are limited and are generally used as salvage therapy in patients who fail to respond to whole-brain radiotherapy or radiosurgery.19–21 However, melanoma brain metastases typically do not respond to chemotherapy.3–5
Temozolomide readily crosses the blood-brain barrier, has activity in MM,9,10 and has recently been shown to have activity in patients with brain metastases from various malignancies.11,12,22 However, there has been limited experience with temozolomide in melanoma patients with brain metastases. Christodoulou et al11 studied the safety and efficacy of temozolomide in 28 heavily pretreated patients with recurrent brain metastases. A PR was achieved in one (4%) of 24 assessable patients, and stable disease was reported in four patients (17%). However, this trial enrolled only two patients with melanoma, one of whom had stable disease in the brain and a CR in lung metastases following treatment with temozolomide. To our knowledge to date, the current study represents the largest clinical trial experience in patients with melanoma brain metastases. It is also the only large trial of its type to systematically evaluate response in the brain using a standardized Gd-MRI in all patients, with central, blinded radiology review.
The occurrence of objective responses in brain metastases using a single systemic agent without added radiation is promising and suggests that temozolomide has activity in this setting. Clinical benefit (ie, objective response or stable disease) was observed primarily in patients who had received no prior chemotherapy for brain metastases; 7% achieved an objective response, and 29% had disease stabilization. However, one patient previously treated with DTIC achieved a PR in the brain, and 18% of patients who received prior chemotherapy had stable disease for a median of 28 days. Therefore, temozolomide has activity in patients who develop CNS relapse following chemotherapy (including DTIC) for systemic MM. It is noteworthy that single-agent therapy with DTIC or temozolomide produces systemic response rates of 10% to 15% in patients with advanced MM, but these occur primarily in nonvisceral sites.9,23 Responses to these agents in visceral sites of disease occur in less than 10% of patients. Indeed, the 6% overall response rate in brain metastases observed in this study is comparable to the response rate in visceral metastases, suggesting that melanoma brain metastases may be as responsive to temozolomide as are visceral metastases. This suggests that the failure of previous chemotherapy regimens to effectively treat melanoma brain metastases is related more to their failure to effectively penetrate the CNS, rather than chemoresistance of brain metastases.
The activity of temozolomide in the current study, albeit modest, suggests two potential applications for temozolomide: in combination with radiotherapy or other agents as treatment for brain metastases, and to prevent relapse in the brain in patients with advanced systemic disease. The latter approach was investigated by Atkins et al,24 who substituted temozolomide for DTIC in an aggressive biochemotherapy regimen of temozolomide, cisplatin, vinblastine, interleukin-2, and interferon alfa-2b. Among the 47 assessable patients, a 47% systemic response rate was observed and, importantly, only two responding patients developed an isolated CNS relapse.
In addition, combination therapy with temozolomide and radiotherapy has demonstrated promising results in patients with brain metastases. Thirty-one patients with brain metastases associated with melanoma were treated with temozolomide (75 mg/m2/d for 6 weeks and repeated every 10 weeks) in combination with whole-brain radiotherapy (30 Gy delivered over 10 fractions [days 1 to 5, and 8 to 12]).14 Three patients (10%) experienced an objective response, including one CR lasting 4.5 months, and two patients had PRs lasting 2 months and 7 months, respectively. A study by Antonadou et al12 investigated the efficacy and safety of temozolomide in combination with concurrent radiotherapy in patients with previously untreated brain metastases associated primarily with lung and breast cancer. Patients were treated with radiotherapy alone for 4 weeks or with daily oral temozolomide (75 mg/m2/d) combined with fractionated radiotherapy for 4 weeks, followed by maintenance therapy with temozolomide (200 mg/m2/d for 5 days × 28 days for an additional six cycles). The objective response rate in patients treated with temozolomide plus radiotherapy was significantly higher than in patients receiving radiotherapy alone (96% v 67%; P = .017). CRs were reported in nine patients (38%), and PRs were seen in 14 patients (67%) treated with temozolomide plus radiotherapy, versus seven CRs (33%) and seven PRs (33%) in patients treated with radiotherapy only. In addition, there was a marked neurologic improvement in patients receiving temozolomide, and temozolomide plus radiotherapy was safe and well tolerated.
Both of these studies used a continuous low-dose schedule of temozolomide,12,14 which may provide greater efficacy in the CNS based on the potential to achieve greater drug exposure over time (compared with the standard schedule) and to prevent the occurrence of drug resistance.25–29 Ongoing trials with temozolomide in patients with brain metastases are attempting to exploit the potentially greater activity of this alternative dosing schedule and use it in combination with radiotherapy and other agents. For example, Hwu et al30 have demonstrated that combination therapy with daily temozolomide and thalidomide is well tolerated and has antitumor activity in patients with advanced MM. A recent phase II study of this combination also demonstrated two CRs and one PR in brain metastases among 11 assessable melanoma patients.31 Based on these studies, an ongoing trial coordinated by the Cytokine Working Group is investigating an extended daily schedule of temozolomide plus thalidomide, with and without radiotherapy, in patients with brain metastases from MM.
In conclusion, temozolomide was well tolerated in this large, multicenter, phase II trial, and demonstrated modest antitumor activity as a single agent (using the standard 5-day dosing schedule) for the treatment of melanoma brain metastases, including patients with multiple brain lesions. The ease of oral administration of temozolomide and the possible benefits of combination therapy, particularly with alternative dosing schedules, warrant further evaluation of this agent in the treatment of brain metastases associated with MM. Ongoing studies at several institutions are attempting to optimize the use of temozolomide for the treatment of established disease, or to prevent metastasis to the brain, a primary site of relapse for patients with MM who respond to treatment for systemic disease.
Authors' Disclosures of Potential Conflicts of Interest
The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Acted as a consultant within the last 2 years: Sanjiv S. Agarwala, Schering-Plough; Nicholas Thatcher, Schering-Plough; Martin Gore, Schering-Plough. Received more than $2,000 a year from a company for either of the last 2 years: Sanjiv S. Agarwala, Schering-Plough.
Acknowledgments
We wish to acknowledge the contributions of all the participating investigators, study coordinators, and nurses.
Footnotes
-
Authors' disclosures of potential conflicts of interest are found at the end of this article.
- Received November 7, 2003.
- Accepted March 10, 2004.
