A randomized, double-blind, placebo-controlled study of high-dose bosentan in patients with stage IV metastatic melanoma receiving first-line dacarbazine chemotherapy

Patients eligible for the study were at least 18 years of age with histologically-proven malignant melanoma [4] and stage IV measurable disease as defined by Response Evaluation Criteria in Solid Tumors (RECIST) [19]. No prior therapy with dacarbazine was permitted and any prior radiation therapy was required to have been >30 days before study drug administration, with indicator lesions being either outside of the field of radiation or new, non-irradiated, lesions. Eastern-Cooperative Oncology Group (ECOG) performance status was required to be ≤ 2 and life expectancy >12 weeks. Female patients were required not to be pregnant or breastfeeding, and to be either post-menopausal, surgically sterile, or practicing a reliable method of contraception.

The main criteria for exclusion from the study were: lactate dehydrogenase >1.5 times the upper limit of normal (ULN); alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) >3 times the ULN at screening or ALT and/or AST >2 times the ULN and total bilirubin >2.0 mg/dL at screening; hemoglobin >30% below the lower limit of normal; central nervous system metastases or leptomeningeal metastases; any prior chemotherapy, biological therapy or immunotherapy for stage IV metastatic disease; immunotherapy <30 days before treatment start; ocular melanoma; known hypersensitivity to any excipients of bosentan; history of other malignancy in the past 5 years, with the exception of squamous cell carcinoma of the skin treated with local resection and basal cell carcinoma; planned use (or use within 4 weeks of the start of bosentan dosing) of another investigational drug; any standard contraindications to dacarbazine. Concurrent use of calcineurin inhibition, sirolimus, fluconazole or glyburide was not permitted. The study was conducted in full compliance with the principles of the Declaration of Helsinki and Good Clinical Practice. Local institutional review boards or independent ethics committees approved the protocol and all patients gave written informed consent.

Treatment with dacarbazine 1000 mg/m² every 3 weeks starting on Day 1 of the study was mandatory for all patients. In addition, eligible patients were randomized in a 1:1 double-blind manner to receive bosentan or matching placebo. Bosentan was administered orally with or without food at a dose of 500 mg twice a day (1000 mg total daily dose). This dose, equivalent to four times the standard daily bosentan dose used in the treatment of PAH, was selected on the basis of an earlier phase II study [17] and in vitro data from melanoma cell lines [13] in order to achieve maximum potential exposure. Four tablets (125 mg per tablet) were taken in the morning and four tablets approximately 12 hours later. Down-titration to 250 mg twice a day or 125 mg twice a day was allowed if the target study medication dose was not tolerated; however, it was recommended to maintain the full dose wherever possible.

Tumor measurements including chest X-ray, computerized tomography scan, or magnetic resonance imaging were performed on Day 1, Week 6, every 6 weeks during dosing, and one week after the end of treatment. Assessment for tumor progression/response was performed at Week 6, every 6 weeks during dosing and one week after the end of treatment according to RECIST [19]. Patients underwent a series of laboratory tests including assessment of hematology (hemoglobin, hematocrit, erythrocyte count, leukocytes, platelets), liver function (ALT, AST, alkaline phosphatase, total and direct bilirubin) and chemistry (lactate dehydrogenase, sodium, potassium, blood urea nitrogen, creatinine) variables; these were performed at screening, Weeks 1-6, 8, 9, 12 and every 3 weeks thereafter during dosing, and at 1 and 4 weeks after the end of treatment. Assessments were scheduled for adverse events and serious adverse events on Day 1, at Weeks 1-6, 8, 9, 12 and every 3 weeks thereafter during dosing, and one week after the end of treatment. All new adverse events up to 24 hours after study drug discontinuation, and all new serious adverse events and grade 4/5 adverse events (National Cancer Institute Common Terminology Criteria for Adverse Events [NCI-CTCAE]) were reported up to 28 days after the end of treatment.

The primary efficacy endpoint was time to tumor progression. Time to tumor progression was measured from randomization to the start of the first occurrence of one the following: objective or subjective (as per investigator's judgment) tumor progression; death due to disease progression; and initiation of other or additional anti-tumor therapy in the absence of documented tumor progression.

In the absence of an event as described above, patients were right censored at the last date of tumor assessment or at the randomization date in absence of post-randomization measurements. This rule was followed for patients on study treatment; patients being followed on study but off study treatment; patients lost to follow-up; patients withdrawn from the trial in the absence of progressive disease; and patients who died due to non-cancer-related causes. Secondary endpoints included progression-free survival and overall survival.

Safety endpoints included the incidence of treatment-emergent adverse events (NCI-CTCAE) occurring during and up to 1 day after end of study treatment; incidence of serious adverse events up to 28 days after end of study treatment; adverse events that led to discontinuation or dose reduction; incidences of worst CTCAE grade laboratory abnormalities during and up to 28 days after the end of study treatment; incidences of hemoglobin >8 and ≤ 10 g/dL, and ≤ 8 g/dL, with decrease from baseline; liver aminotransferases ≥ 3 times the ULN; total bilirubin ≥ 2 times the ULN with ALT and/or AST ≥ 3 times the ULN; increased blood alkaline phosphatase.

A total of 66 events were required at outset to detect a bosentan vs placebo hazard ratio (HR) of 0.5 with a type-I error of 0.05 (two-sided) and 80% power using the log-rank test to compare the distribution of time to tumor progression for the two treatment groups. Assuming enrolment of 70 patients (1:1 randomization) over 70 weeks, 66 events were expected to occur within 105 weeks if the median time to tumor progression was 10 weeks in the placebo group and the HR was 0.5. Efficacy data were analyzed on the all-randomized population comprising all randomized patients, whether or not they received any study drug. Safety data were analyzed on the all-treated population comprising all randomized patients who received study drug.

The study was performed in 11 centers in Australia, with the first patient, first visit in September 2005 and the last patient, last visit in February 2008. A total of 80 patients were randomized; however, two patients in each group did not receive any study treatment due to withdrawal of consent or randomization error by the pharmacy (Figure 1), and were excluded from the safety evaluation.

Baseline demographics are summarized in Table 1. Compared with the placebo group, the bosentan group comprised a higher proportion of males and the mean age of patients was slightly higher. The median number of target and non-target lesions was five (min, 2; max, 9) in the placebo group and five (min, 1; max, 10) in the bosentan group, the most common sites being the lung (placebo, 81.6%; bosentan, 59.0%), lymph nodes (31.6%; 38.5%), adrenal glands (5.3%; 28.2%), soft tissue (21.1%; 25.6%) and liver (18.4%; 23.1%). The percent of patients with lactate dehydrogenase levels ≤ ULN and >ULN was similar between the two treatment groups, and mean and median levels were comparable.

In each group, 38 patients were treated and evaluable for exposure. For dacarbazine, the mean (± standard deviation) number of cycles administered was 5.5 (± 4.5) in the placebo plus dacarbazine treatment group and 4.8 (± 4.1) in the bosentan plus dacarbazine treatment group. Median duration of dacarbazine exposure (min; max) was 12.0 weeks (3.0; 61.0) and 7.5 weeks (3.0; 51.1) in the two treatment groups, respectively. The median daily exposure for bosentan (min; max) was 1000 mg (125 mg; 1000 mg) and the median duration of exposure (min; max) was 8.1 weeks (3.0; 71.0).

The primary endpoint, time to tumor progression, was not significantly different between the treatment groups, with a median of 2.8 months (95% CI, 1.4-4.1) for placebo and 1.6 months (95% CI, 1.3-4.0) for bosentan (HR, 1.144; 95% CI, 0.717-1.827; p = 0.5683) (Figure 2a). Supportive analysis on the all treated and per-protocol sets showed similar results (data not shown).

The secondary endpoints of median overall survival (Figure 2b) and progression-free survival were also not significantly different between treatment groups. Median overall survival was 10.6 months (95% CI, 6.9-14.7) with placebo and 13.0 months (7.8-16.6) with bosentan (HR, 1.044; 95% CI, 0.584-1.865; p = 0.8841) (Figure 2b). The Kaplan-Meier estimate for overall survival at 12 months was 42.1% (95% CI, 25.8-58.4) and 52.1% (95% CI, 35.1-69.1) for the two treatment groups, respectively. Median progression-free survival was 2.8 months (95% CI, 1.4-4.1) for placebo and 1.6 months (95% CI, 1.3-4.1) for bosentan (HR, 1.064; 95% CI, 0.664-1.704; p = 0.7944). The Kaplan-Meier estimate for progression-free survival at 6 months was 18.9% (95% CI, 6.3-31.6) and 27.0% (95% CI, 12.7-41.3) for the two treatment groups, respectively.

Adverse events occurring during and up to 1 day after the end of study treatment are shown in Table 2. All but one patient experienced an adverse event and similar proportions of patients in the two groups experienced a serious adverse event (placebo plus dacarbazine, 31.6%; bosentan plus dacarbazine, 34.2%). Anemia, thrombocytopenia, vomiting and lethargy were more common among patients receiving bosentan than placebo (all NCI CTCAE grades; ≥ 10% total difference). Conversely, fatigue, headache, dyspnea and upper respiratory tract infection were more common among patients receiving placebo.

Hematology assessment was graded according to the NCI CTCAE system and is reported in Table 3 as incidences of worst grades experienced during and up to 28 days after the end of study treatment. Grade 4 events occurred in two patients: one patient (2.6%) in the placebo plus dacarbazine group had Grade 4 reduction in neutrophils and one patient (2.6%) in the bosentan plus dacarbazine group had Grade 4 reduction in hemoglobin. Grade 3 hematology abnormalities occurred in no more than two patients (5.3%) in either treatment arm for any variable. Grades 2, 3, and 4 reductions in hemoglobin occurred with a higher incidence in the bosentan plus dacarbazine treatment group than in the placebo plus dacarbazine treatment group.

Clinically significant laboratory abnormalities were also documented. Four patients (10.5%) in each treatment group had liver aminotransferases ≥ 3 times the ULN; 3 (7.9%) in each group had liver aminotransferases ≥ 3 and <5 times the ULN; and 1 (2.6%) in each group had liver aminotransferases ≥ 5 and <8 times the ULN. No patient in either treatment group experienced aminotransferase elevations ≥ 8 times the ULN. Reductions in hemoglobin, with decreases from baseline, occurred more often in the group receiving bosentan plus dacarbazine than the group receiving placebo plus dacarbazine: hemoglobin decreased to >8 and ≤ 10 g/dL in 10 patients (26.3%) and to ≤ 8 g/dL in 3 patients (7.9%) with bosentan plus dacarbazine; the corresponding changes occurred in only one patient (2.6%) each with placebo plus dacarbazine. No patients experienced bilirubin ≥ 2 times the ULN.

There were three deaths among patients receiving bosentan (all attributed to disease progression, with additional ascites in one case and hyponatremia in another case) and two deaths among patients receiving placebo (one due to disease progression and one due to cerebrovascular accident). All of the five deaths occurred after study treatment was stopped (during the 28-day follow-up) and none were considered, by the investigators, to be related to study treatment.

Adverse events were responsible for premature treatment discontinuation in two patients in the placebo plus dacarbazine group and three patients in the bosentan plus dacarbazine treatment group. In the placebo plus dacarbazine group, one of the two patients had an adverse event denoting disease progression and the other had an increase in serum ALT and alkaline phosphatase levels that was considered to be related to treatment and led to discontinuation of both placebo and dacarbazine. In the bosentan plus dacarbazine group, two of the three patients had adverse events denoting disease progression and the other had anorexia, fatigue and nausea, with the fatigue and nausea considered to be related to both bosentan and dacarbazine.