Previous biochemotherapy regimens for metastatic melanoma have required attenuated dosages of interleukin 2 (IL-2) that may have compromised efficacy.
Previous biochemotherapy regimens for metastatic melanoma have required attenuated dosages of interleukin 2 (IL-2) that may have compromised efficacy.
In a phase 2 study, the authors tested sequential temozolomide (75 mg/m2 per day orally for 3 weeks) followed by high-dose, IL-2 (600,000 U/kg per dose intravenously; maximum, 14 doses over 5 days).
Thirty-eight patients with treatment-naive American Joint Committee on Cancer stage IV melanoma (8 patients with M1a disease, 6 patients with M1b disease, and 24 patients with M1c disease) were enrolled. Ten patients had a history of treated brain metastases. Thirty-one patients who received at least 2 cycles of IL-2 were evaluable for response. Grade 3 toxicities included hyperbilirubinemia (9 patients), hematologic toxicities (leukopenia in 5 patients, thrombocytopenia in 3 patients), diarrhea (2 patients), and oliguria (1 patient). One patient had grade 4 nausea. The overall response rate (ORR) was 16% and included 3 complete responses that lasted 10.8 months, ≥32 months, and ≥36 months and 2 partial responses that lasted 13 months and 14 months. Responses were observed in patients with M1a disease and in patients with M1c disease. Sixteen patients had stable disease (15 patients progressed). The median progression-free survival (PFS) was 5.3 months (95% confidence interval [CI], 3.7-7.5 months). The probability of PFS at 6 months was 0.52 (95% CI, 0.33-0.67). Among 38 enrolled patients, 16 patients remained alive at a median follow-up of 6.7 months (range, 1.9-36.1 months). The median overall survival (OS) was 12.1 months (95% CI, 9.1-16.4 months), and the probability of 12-month OS was 0.54 (95% CI, 0.34-0.70 months).
The current results indicated that it is safe to administer HD IL-2 sequentially with temozolomide and that this combination has lower toxicity than previously used concurrent biochemotherapy regimens. However, The ORR and the durability of responses with this combination did not exceed those of single-agent HD IL-2. Cancer 2008. © 2008 American Cancer Society.
Metastatic melanoma remains a disease with a poor prognosis and a median survival duration <1 year. Single-agent dacarbazine or temozolomide is capable of producing responses in the range from 6.7% to 20% of patients, but long-term remissions are rare, and no impact on survival is evident.1, 2 It has not been demonstrated that combination chemotherapy is superior to single-agent chemotherapy.3–6 High-dose (HD) interleukin 2 (IL-2) produces a response rate of 16% in patients with stage IV melanoma, some of which are durable.7
Temozolomide is a cytotoxic alkylating agent that is metabolized to the same active metabolite, monomethyl triazenoimidazole carboxamide (MTIC),8, 9 as dacarbazine. A randomized phase 3 trial of oral temozolomide administered at a dose of 200 mg/m2 per day for 5 of every 28 days, compared with intravenous dacarbazine in patients with metastatic melanoma, demonstrated that temozolomide was at least equivalent to dacarbazine in terms of response rate and survival.2 However, temozolomide possesses the ability to cross the blood-brain barrier10 and has ease of oral administration with 100% bioavailability.9, 11–13
In addition, there is a potential for schedule variation with temozolomide to reduce toxicity and improve efficacy by reversing drug resistance. It is believed that the cytotoxicity of MTIC is caused primarily by alkylation at the O6 position of guanine,11 with additional alkylation at the N7 and N3 positions of lesser importance.12 O6-alkyl-guanine-DNA-alkyltransferase (AGAT) is a DNA repair protein that removes methyl adducts from DNA, thereby reversing the cytotoxicity of temozolomide by restoring DNA to its native state. Administration of temozolomide results in decreased AGAT activity, as measured in peripheral blood mononuclear cells (PBMCs)14; and it was demonstrated that continuous, prolonged, daily administration of temozolomide more effectively depleted the activity of AGAT (73% after 21 days, with the low levels persisting up to Day 28).15 In the phase 1 studies of continuous daily temozolomide, including a trial of 21 days of temozolomide of a 28-day dosing schedule, dose-limiting toxicities were hematologic, and the maximum tolerated dose was 100 mg/m2.16, 17 The dose of 75 mg/m2 per day allowed for a 2.1-fold greater exposure to drug per 28-day period compared with the 5-day schedule of 200 mg/m2 per day repeated every 28 days.
Two approaches to biochemotherapy (BCT) have involved sequential chemotherapy followed by immunotherapy in 1 approach and concurrent chemoimmunotherapy in the other. The sequential approach, testing chemotherapy with cisplatin, vinblastine, and dacarbazine (CVD) followed by IL-2 and interferon α2b (IFN), was compared with chemotherapy (CVD) alone in a randomized trial conducted at the M. D. Andersen Cancer Center. Although response rates and time-to-progression were improved for the sequential BCT group, the survival difference was only of borderline significance, and toxicity was formidable.18 A concurrent BCT regimen of CVD coupled with IL-2 and IFN was tested in a phase 2 study, and the results from that combination appeared to be equivalent to those from the sequential BCT regimen with more practicality and lower toxicity. The concurrent CVD/IL-2/IFN regimen subsequently was adopted by the US Intergroup and compared with CVD in an important randomized phase 3 trial (Eastern Cooperative Oncology Group [ECOG] trial 3695), but it failed to produce significantly better response rates, progression-free survival (PFS), or overall survival (OS) relative to chemotherapy alone.19 Two other phase 3 trials of slightly different BCT regimens conducted in Europe failed to demonstrate an improvement in response or recurrence rates or in OS.20, 21 A recent meta-analysis from 18 trials (11 trials of chemotherapy with or without IFN and 7 of trials chemotherapy with or without IFN plus IL-2) demonstrated no benefit for BCT on OS.22 Therefore, we conducted a phase 2 study to test a new approach of single-agent chemotherapy with temozolomide administered in an extended schedule followed by HD IL-2 to evaluate the safety and efficacy of this sequential combination in patients with advanced metastatic melanoma.
Patients ages 18 years to 70 years were eligible if they had histologically confirmed metastatic melanoma (American Joint Committee on Cancer [AJCC] stage IV) and measurable or evaluable disease according to the Response Evaluation Criteria in Solid Tumors (RECIST). Previous chemotherapy or immunotherapy for stage IV melanoma was not allowed, except for 1 prior therapy with vaccines or granulocyte-macrophage-colony stimulating factor or prior radiation therapy. Adjuvant therapy with immunologic agents, including IFN, IL-2, and vaccines, was permitted. Eligible patients also were required to have met the following criteria: a World Health Organization (WHO) performance status (PS) of 0 or 1, and adequate hematologic values (an absolute neutrophil count [ANC] ≥1500 cells/μL, hemoglobin ≥10 g/dL, platelet count ≥100,000/μL), hepatic function (serum bilirubin ≤2 times the upper limit of normal [ULN] and serum alanine and aspartate transferase levels ≤3 times the ULN), and renal function (serum creatinine <2 mg/dL or a calculated clearance >60 mL/minute). Patients with brain metastases were eligible, but only after they received definitive treatment, including surgical resection, stereotactic radiosurgery (SRS), whole-brain radiation therapy (WBRT), or combinations.
All patients aged >45 years or as clinically indicated underwent a cardiac stress test and pulmonary function tests within 8 weeks before registration. They were eligible if they had no significant abnormality on cardiac stress testing and if they had a forced expiratory volume in 1 second >2 L or >75% of the predicted value.
All patients provided a written informed consent. Men and women of childbearing potential were instructed to use an approved method of birth control. Premenopausal women were required to have a negative serum pregnancy test.
This was a phase 2, single-arm, safety and efficacy study of sequential temozolomide and HD IL-2 that was carried out at the University of Pittsburgh Medical Center. Eligible patients received temozolomide (75 mg/m2 per day orally) continuously for 3 weeks in an outpatient setting. HD IL-2 was administered after temozolomide during Week 4 at a dose of 600,000 U/kg per dose as a 15-minute intravenous infusion every 8 hours for 5 days (maximum, 14 doses). IL-2 therapy was carried out in a specialized inpatient oncology unit (not an intensive care unit). Administration of HD IL-2 followed the guidelines established by the National Cancer Institute (NCI). Each cycle of therapy lasted 4 weeks. The cycle was repeated, and disease assessment was carried out first after 2 cycles of therapy (Table 1). A patient had to complete at least 2 cycles of HD IL-2 to be considered evaluable for response assessment in this phase 2 study.
|Day 1||Day 21||Day 28||Days 49-56|
|Temozolomide 75 mg/m2/ d orally continuously for 3 wk followed by||IL-2 600,000 U/kg per dose iv every 8 h×5 d (maximum, 14 doses)||Temozolomide 75 mg/m2 per d orally continuously for 3 wk followed by||IL-2 600,000 U/kg per dose iv every 8 h×5 d (maximum, 14 doses)|
Patients without evidence of disease progression at response assessment (partial response [PR], complete response [CR], minor response, or stable disease [SD]) were offered additional cycles of therapy 2 weeks after they completed an even-numbered cycle. All toxicities were required to resolve to grade 2 or less before reinitiation of therapy. A delay of an additional week (maximum, 3 weeks after each even-numbered cycle) was allowed for patient convenience or recovery from toxicity. Institutional Review Board approval was obtained.
The NCI's Common Toxicity Criteria version 3.0 (available at: http://ctep.cancer.gov/forms/CTCAEv3.pdf accessed on February 20, 2008) was used for grading toxicities. Patients were monitored for toxicity every 1 to 2 weeks on temozolomide and on Days 1 through 5 of each HD IL-2 treatment cycle.
Systemic computed tomography scans and gadolinium-enhanced magnetic resonance imaging studies of the brain were obtained at baseline. Radiologic studies were repeated at the completion of every even-numbered cycle to assess response. RECIST criteria were used to determine the levels of response.
Dose adjustments during temozolomide treatment were required to be done according to hematologic or nonhematologic toxicities assessed at least every 2 weeks. Stopping treatment was required if blood counts tested at 3 weeks remained unsatisfactory. The reduced dose, based on toxicity, was considered the starting dose of temozolomide for subsequent cycles.
For HD IL-2, established NCI guidelines for administration were followed. Patients were assessed continuously, and a decision whether or not to administer a dose of IL-2 was made at each 8-hour time point.
This was a single-arm, phase 2 clinical trial. A 2-stage Simon-type phase 2 study design23 was used. The primary data analysis was estimation with 95% confidence intervals (CIs) of the objective response rate. The response rates were calculated for patients with measurable disease. For the analysis of secondary objectives, toxicities were summarized by grade, frequency, and duration. PFS and OS were estimated by using the Kaplan-Meier method.24
In total, 38 patients with AJCC stage IV melanoma were enrolled between March 2004 and April 2006. There were 24 men and 14 women. The median age was 50 years (age range, 27-74 years). Twenty-patients (73.7%) had a WHO PS of 0 (normal), and 10 patients (26.3%) had a WHO PS of 1 (ambulatory). Eight patients (21%) had AJCC stage M1a disease, 6 patients (15.8%) had M1b disease, and 24 patients (63.2%) had M1c disease.25 Ten patients had a history of treated brain metastases. One patient who had brain metastases underwent surgery and received WBRT; 1 patient underwent surgery, received WBRT, then underwent SRS at recurrence; and 8 patients underwent SRS only. Baseline patient and disease characteristics are shown in Table 2.
|Variable||No. of Patients (%)|
|Performance status (WHO)|
|Peptide vaccine||7 (18.4)|
|GM-CSF or placebo||4 (10.5)|
|History of treated brain metastases*||10 (26.3)|
All 38 patients received temozolomide, and no patient required a temozolomide dose reduction. Four patients did not receive any IL-2 because of rapid clinical disease progression (PD) and declining PS. Three patients received only 1 cycle of IL-2 because of declining PS (1 patient), rapid PD after 1 cycle (1 patient), and patient decision (1 patient). Thirty-one patients received at least 2 cycles (1 course) of IL-2. Table 3 summarizes the treatment details by course, cycle, and dosage.
|Course||Cycle||Total No. of Patients Treated*||No. Patients Off Study After Treatment||Median No. of IL-2 Doses per Cycle (Range)|
In total, 123 cycles were administered (median, 2 cycles per patient). No dose modifications were made for IL-2. Either the entire dose was given or it was held based on the presence or absence of relative and absolute criteria according to the published NCI guidelines. In some patients, a dose was held based on patient request even in the absence of relative or absolute criteria.
Thirty-one patients received at least 2 cycles of therapy according to protocol and were evaluable for response assessment. The overall response rate was 16.1% (95% CI, 3.2%-29.1%). In April 2007, 3 patients (9.7%) had a CR that lasted 10.8 months (M1c disease), ≥32 months (M1c disease), and ≥36 months (M1a disease) from the first tumor response assessment point after at least 2 cycles of therapy. Two patients (6.5%) had PRs that lasted 13 months (M1a disease) and 14 months (M1c disease), respectively, from the first tumor response assessment. Sixteen patients (51.6%) had SD that lasted from 2 months to 11 months, but all eventually progressed except 1 patient with M1c disease who had SD that still was ongoing at the last follow-up (≥11 months) (Table 4).
|Response||No. of Patients (%)||Response Duration, mo|
|Overall response rate||5 (16.1)||10.8 to ≥36|
|Complete response||3 (9.7)*|
|Partial response||2 (6.5)†|
|Stable disease||16 (51.6)|
The 3 patients who did not have PD were observed at 11 months, 32 months, and 36 months. The median PFS was 5.3 months (95% CI, 3.7-7.5 months), and the probability of PFS at 6 months was 0.52 (95% CI, 0.33-0.67). Figure 1 shows the Kaplan-Meier plot of the probability of PFS.
Among 38 patients who were enrolled on the study, 16 patients remained alive at a median follow-up of 6.7 months (range 1.9-36.1 months). The median OS was 12.1 months (95% CI, 9.1-16.4 months), and the probability of 12-month OS was 0.54 (95% CI, 0.34-0.70). Figure 2 shows the Kaplan-Meier plot of the probability of OS.
Among 10 patients with treated brain metastases, 1 patient received 1 cycle of IL-2 and was not evaluable for response. Three patients had PD after 2 cycles. Six patients had SD that lasted from 2 months to 7 months. Among 9 evaluable patients who had treated brain metastases on enrollment, 5 patients had recurrent brain metastases on progression. Among the remaining response-evaluable patients, 3 patients had evidence of brain metastases on progression.
The most frequent nonhematologic grade 3 or 4 adverse events were high bilirubin in 9 patients (24%), transient hypotension associated with IL-2 infusion in 3 patients (8%), diarrhea in 2 patients (5%), and oliguria in 1 patient (3%). Hematologic grade 3 or 4 adverse events included leukopenia (5 patients; 13.2%) and thrombocytopenia (3 patients; 7.9%). No patient had treatment terminated because of adverse events. The most common reason for treatment discontinuation was disease progression (25 patients). Of the 22 deaths reported as of April 2007, no deaths were related to treatment toxicity. Table 5 summarizes the most frequently reported adverse events by severity.
|Adverse Event||All Grades||Grade 3/4|
|No. Patients||%||No. Patients||%|
|Rash, dry skin||10||26.3||0||0|
This study explored the safety and efficacy of sequential BCT using the 2 best-known agents in advanced metastatic melanoma. The objective was to improve the efficacy and tolerability of this regimen over previously tested BCT regimens. It is apparent that temozolomide produces single-agent results that are equivalent to those produced by dacarbazine2 but with certain advantages, including oral administration and the ability to cross the blood-brain barrier.10 In terms of immunotherapeutic approaches to metastatic melanoma, to our knowledge, the only agent that has demonstrated durable responses in at least a small percentage of patients is HD IL-2.7 Therefore, in the current study, we combined single-agent chemotherapy with temozolomide and single-agent immunotherapy with HD IL-2, both administered without dose attenuation, and evaluated the efficacy of this sequential combination in patients with AJCC stage IV melanoma. The phase 3 M. D. Anderson Cancer Center trial of sequential CVD chemotherapy followed by IL-2 and IFN produced a doubling of the response rate and the time to progression and a borderline survival advantage, albeit at the expense of major grade 3 and 4 toxicity.18 Conversely, the concurrent approach using these same drugs reportedly was ineffective.19 This may hint that the myelotoxic effects of concurrent therapy compromise the ability of IL-2 to induce an effective immune response. It is noteworthy that our sequential BCT regimen delivered IL-2 on Days 21 and 49, unlike standard HD IL-2 monotherapy, as reported by Atkins et al.7
The overall response rate in the current study was 16% (95% CI, (3.18%-29.08%). All 5 responders had durable responses that lasted >10 months, and 2 of those responses were ongoing for ≥32 months and ≥36 months at last follow-up. Responses were observed in both M1a disease and M1c disease. One patient with M1c disease had SD that was ongoing for ≥11 months at last follow-up. Although the response rate was comparable, the quality of the responses was superior to the quality of responses reported with BCT in ECOG 3695.19 The median PFS was 5.3 months with 95% CI (3.7-7.5 months) in the current study compared with ECOG 3695, in which the median PFS was 3.6 months for the chemotherapy-treated controls (CVD) and 5.3 months for the BCT arm (CVD plus IL-2 and IFN). The median OS was 12.1 months (95% CI, 9.1-16.4 months) in the current study compared with 8.7 months for the chemotherapy arm and 8.4 months for the BCT arm in ECOG 3695. Conversely, the overall response rate and the durability of responses do not exceed those reported with single-agent HD IL-2.7 Our own data from a clinical trial at the University of Pittsburgh Cancer Institute with HD IL-2 after failure on BCT produced an overall response rate of 19.2% in a heavily pretreated population with advanced disease.26 Responses included 4 patients with CR, including 2 CRs that were ongoing at ≥26 months and at ≥41 months, suggesting that HD IL-2 is active in patients who have either failed treatment or progressed after treatment with lower dose IL-2-based BCT. Although no randomized clinical trials were conducted in patients with melanoma, a randomized phase 3 trial in patients with metastatic renal cell carcinoma demonstrated that HD IL-2 was superior to low-dose IL-2 and IFN.27 These data may support the hypothesis that HD IL-2 is superior to continuous-infusion IL-2 at the intermediate dose used in the BCT regimen that was tested in the ECOG 3695 intergroup trial and may suggest that future IL-2-based regimens in melanoma must consider the importance of dose intensity with IL-2 therapy.
Researchers recently have started to explore the utility of OS or PFS as benchmarks for efficacy in phase 2 studies instead of the traditional objective tumor response rate.28 On the basis of a recent meta-analysis of previously collected data from 42 melanoma phase 2 trials conducted by the Southwest Oncology Group, ECOG, Cancer and Leukemia Group B, the North Central Cancer Treatment Group, and the National Cancer Institute of Canada Clinical Trials Group in the years 1975 through 2005, Korn et al have suggested the use of 1-year OS or 6-month PFS as benchmarks for future phase 2 studies.
That meta-analysis, which was based on 1278 patients, provided an estimate of the 1-year OS rate (25.5%) and 1-year OS rates for 24 prognostic classes (ranging from 5.5% to 63.8%) defined by 4 statistically significant, independent prognostic factors that included PS, the presence of visceral metastasis, sex, and the exclusion of patients with brain metastasis. The authors of the meta-analysis acknowledged that future trials may have different survival rates than in the past because of patient mixes that differ in terms of prognostic variables. To address this, they considered defining the null hypothesis target for a phase 2 trial based on the prognostic variables recorded in the trial. They provided a table that contained the relevant information for a trial using a 1-year OS rate as the endpoint. These predicted values were based on a logistic regression analysis with effects included for PS, sex, the presence of visceral metastasis, and whether a trial allowed patients with brain metastasis. We used this model for our study because our mix of patients with the same PS, incidence of visceral disease, and sex distribution and because we included patients with brain metastasis. To account for the immaturity of follow-up for some patients, we eliminated 10 of the total 38 patients who were alive with <1 year of follow-up. By including only the remaining 28 patients, we observed that 6 patients were alive at 1 year for a 1-year survival rate of 21.4% (90% CI, 9.8%-38%), whereas the expected 1-year survival rate was 26.9%. By excluding patients who were not evaluable for response (those who did not complete 2 cycles of HD IL-2), the expected 1-year survival rate became 27.6%, essentially identical to the original estimate of 26.9%. According to this reasoning, we conclude that there is no evidence that our regimen is promising or worthy of phase 3 testing.
Brain metastases in patients with stage IV melanoma have been reported in at least 18% to 46% of patients,29, 30 with roughly twice this prevalence reported in autopsy series.30–33 Brain metastases are a major cause of morbidity and mortality, leading directly to death in as many as 95% of patients with melanoma who have central nervous system spread of the disease.30, 33, 34 Definitive local treatment can be achieved with surgery or SRS with or without WBRT in carefully selected patients who have limited disease and may prolong survival.35–40 Among 9 patients with treated brain metastases on enrollment who completed at least 2 cycles of therapy, 6 patients had SD that lasted from 2 months to 7 months. Five patients had recurrent brain metastases on progression. Among the 22 remaining response-evaluable patients, 3 patients had evidence of brain metastases on progression. According to a before-and-after paired analysis (Table 6), only 7 patients changed with therapy. Four patients with brain metastasis before therapy progressed, but not to the brain, and 3 patients without brain metastasis before therapy had brain metastasis at progression after therapy. The other 24 patients did not change their before-after brain metastasis state and did not contribute to that paired analysis. Therefore, there was no significant change in the incidence of brain metastasis because of the therapy or among the cohort of patients who were treated.
|No. of Patients After Therapy|
|Brain Metastases||No Brain Metastases|
|No. of patients before therapy|
|No brain metastases||3||19|
A recent retrospective analysis has suggested that the outcome of BCT is comparable in patients with and without brain metastases if brain metastases are controlled with multidisciplinary treatment. Among 70 patients who received BCT (consisting of either dacarbazine or temozolomide in combination with a 96-hour continuous intravenous infusion of IL-2 and subcutaneous IFN) for metastatic melanoma, 20 patients had recently diagnosed brain metastases. The median survival from the time of diagnosis with metastatic melanoma was 15.8 months for patients who had brain metastases and 11.1 months for those without central nervous system involvement (P = .26; log-rank test; P = .075; Gehan-Wilcoxon test).
With regard to the safety of this regimen, the combination of single-agent, oral, extended-schedule temozolomide with HD IL-2 was tolerated well. No patient had treatment terminated because of adverse events. No patient required the delay of an IL-2 treatment cycle because of toxicities related to temozolomide. There were no treatment-related deaths.
HD IL-2 continues to be a rare therapeutic approach that offers durable, although infrequent, responses in advanced metastatic melanoma, and combinations with this agent at high dosage are reasonable in the effort to improve efficacy. In a recent phase 1/2 clinical trial of HD IL-2 and escalating doses of anticytotoxic T-lymphocyte antigen 4 (anti-CTLA-4) antibody (ipilimumab, MDX-010; Medarex, Inc./Bristol-Myers Squibb Company) in 36 patients with metastatic melanoma, 22% experienced objective tumor responses, including 3 CRs and 5 PRs.41 Although the response rate observed in that study was not superior to the responses observed with HD IL-2 alone, further exploration using modified, sequential approaches is reasonable with anti-CTLA4 blockade or other emerging novel agents.
The current results indicated that it is safe to administer HD IL-2 sequentially in combination with single-agent temozolomide in an extended schedule and that the combination has lower toxicity than previously used BCT regimens. However, the overall response rate and the number of durable responses induced by this combination did not exceed the results obtained with single-agent HD IL-2. The marginal activity of current single agents may be a part of this problem of limited activity, and exploration of combinations that have greater impact on this disease with HD IL-2 using new and promising biotherapeutic agents may induce more clinically effective cell-mediated immune responses.