Phase I/II Study of Atrasentan, an Endothelin A Receptor Antagonist, in Combination with Paclitaxel and Carboplatin as First-Line Therapy in Advanced Non–Small Cell Lung Cancer

Materials and Methods

Study population and patient eligibility. Patients required histologic or cytologic confirmation of advanced NSCLC (stage IIIB with malignant pleural effusion and stage IV), estimated life expectancy ≥12 weeks, Eastern Cooperative Oncology Group performance status ≤1, and one unidimensionally measurable lesion and to meet New York Heart Association class I criteria if cardiovascular disease was present. Other eligibility criteria included the following: age ≥18 years; no prior chemotherapy or biologic or immunotherapy; and adequate bone marrow (absolute neutrophil count ≥2.0 × 10⁹ cells/L, platelet count ≥100 × 10⁹ cells/L), renal (creatinine clearance ≥60 mL/min), and hepatic functions (bilirubin ≤1.5 upper limit of normal, alanine aminotransferase and/or aspartate aminotransferase ≤2.5× upper limit of normal, or ≤5× upper limit of normal with documented liver metastasis).

Brain metastases were allowed after definitive therapy (resection and/or whole brain irradiation) if asymptomatic and steroid-free for ≥2 weeks. Women who were pregnant, nursing, or of childbearing potential (on inadequate birth control method) and patients with hypersensitivity to cremophor or grade >1 neuropathy were excluded. Written informed consent was obtained from all subjects and an institutional review board reviewed and approved the protocol.

Study design. This open-label, phase I/II study was conducted at the H. Lee Moffitt Cancer Center and Research Institute. The phase I objective was to evaluate the safety of paclitaxel-carboplatin plus a fixed daily dose of atrasentan and to determine potential pharmacokinetic interactions between paclitaxel and atrasentan. The phase II goal was to further detail the safety of the regimen and to determine its antitumor activity, measuring the overall response rate, progression-free survival (PFS), and overall survival (median survival and 1-year survival).

Dose-limiting toxicities were recorded in all phase I patients during cycle 1 to assess the need for paclitaxel-carboplatin dose modifications before the phase II part. Pre- and post-atrasentan paclitaxel pharmacokinetic variables were calculated in all these patients, and based on intraperson paclitaxel pharmacokinetic variability (26), a change of >33% in the mean paclitaxel clearance was considered “clinically significant.” If a clinically significant change and/or more than three dose-limiting toxicities were observed during phase I, the chemotherapy doses were reduced following an established schema before continuing to phase II. In the absence of such findings, the phase II was initiated at phase I starting doses for all drugs.

Prestudy evaluations included a history and physical examination, Eastern Cooperative Oncology Group performance status, complete blood count, serum chemistries, coagulation profile, and urinalysis. Women of childbearing potential required a pregnancy test. Imaging of the chest, head scan, and electrocardiogram were also done. During chemotherapy, evaluations included physical examination, performance status, complete blood count, and serum chemistries before every cycle and imaging of the chest (and other target lesions) every other cycle. After chemotherapy, patients with complete response, partial response, or stable disease were followed every 2 months until progression of disease, death, or loss to follow-up. All patients were followed for a minimum of 30 days after the last dose of atrasentan.

Treatment plan. Atrasentan was supplied by Global Pharmaceutical Research and Development, Abbott Laboratories as 10 mg gelatin capsules. Paclitaxel and carboplatin were commercially available. Study medications were administered as follows. (a) On day 1 of each cycle, and following standard premedications for hypersensitivity and emesis, paclitaxel was dosed at 225 mg/m² as a 3-h i.v. infusion. Carboplatin was dosed to a target area under the curve of 6 mg × min/mL as a 30-min i.v. infusion immediately thereafter. (b) Atrasentan was administered continuously as a 10 mg daily oral dose, starting on day 4 of cycle 1 (to allow for off-atrasentan paclitaxel pharmacokinetic sampling). Treatments were repeated every 21 days for a total of four to six cycles, contingent to the return of all treatment-related adverse events to grade ≤1 or baseline. After completing four or more cycles of therapy, and in the absence of disease progression or intolerable toxicity, patients continued on daily, single-agent atrasentan.

Dose modifications and toxicity criteria. Chemotherapy dose reductions were planned for the transition from phase I to phase II, depending on the occurrence of clinically significant mean paclitaxel clearance changes and/or dose-limiting toxicities. Within study phases, dose modifications were implemented, after treatment was first delayed to allow its resumption once the adverse event had resolved or returned to baseline, based on hematologic and nonhematologic toxicities. Dose reescalation was not permitted. Treatment was stopped in the event that chemotherapy had to be delayed >2 weeks or if toxicities persisted despite maximum dose reductions. All toxicities were graded using the National Cancer Institute Common Toxicity Criteria version 2.0.

Pharmacokinetics. Paclitaxel pharmacokinetic plasma samples were obtained from all phase I patients on days 1 to 3 of cycles 1 and 2 (pre- and post-atrasentan, respectively). Samples were collected (a) before the start of the paclitaxel infusion (predose) and 1.5 and 3 h after the start of the infusion and (b) 0.5, 1, 2, 3, 5, 24, 48, and 72 h after the end of the infusion. Plasma samples were stored frozen at −80°C before assay. Paclitaxel was isolated through solid-phase extraction and all analyses were done by liquid chromatography-mass spectrometry methods described in the literature (27, 28) and validated by the Clinical Trials Laboratory Core.

Using selected ion monitoring, paclitaxel was targeted at m/z 854 a.m.u. with a linear range from 2.5 to 1,000 ng/mL. Quality control samples were checked at 30, 80, and 800 ng/mL, and the overall precisions, expressed as relative SD, were 9.1%, 8.2%, and 4.2%, respectively. Calibration curves were calculated by linear regression with 1/y weighting to determine the slopes, intercepts, and correlation coefficients. The average correlation coefficient (R²) for the assay runs was >0.9995.

Plasma paclitaxel concentration-time profiles were analyzed using standard noncompartmental methods. Derived variables included in the analyses were the maximum plasma concentration (C_MAX), the time to C_MAX (T_MAX), plasma elimination half-life (t_1/2), area under the curve in plasma, steady-state volume of distribution, and total body clearance.

Efficacy assessment. The Response Evaluation Criteria in Solid Tumors (29) method was used to evaluate antitumor activity. Only patients with tumor imaging after two cycles of therapy were evaluable. PFS was the time interval between the dates of the first chemotherapy and disease progression. Overall survival was the time interval between the dates of the first chemotherapy and death. Patients with neither disease progression nor death were censored at their last date known alive. Survival analysis (PFS, median survival, and 1-year survival) was done on an intent-to-treat population.

Statistical considerations. Ten patients were enrolled for the phase I component. The pre- and post-atrasentan mean paclitaxel pharmacokinetic variables were compared using Student's paired t test at the two-sided 0.05 significance level. Assuming a SD of 33% for the mean pharmacokinetic differences, a 10-patient sample size provides 80% power if the mean change in the pharmacokinetic variables is 33% (1 SD).

The primary objective of the phase II component was to reject the null hypothesis (median PFS, 3.5 months) and to accept the alternative hypothesis (median PFS, 6.5 months). To test this and to calculate the sample size, we assumed an exponential distribution. Thus, the estimated 6-month PFSs under null hypothesis and alternative hypothesis were 0.305 and 0.530, respectively. To achieve a power of 0.80, with α = 0.05, we calculated a sample size of 40 patients. Alternative hypothesis was accepted if ≥19 patients were progression-free at 6 months. The PFS and median survival distributions were estimated using the Kaplan-Meier (30) method.

Results

Patient characteristics. Between July 2003 and August 2005, 44 patients were enrolled (phase I, 10; phase II, 34). All patients received at least one cycle of chemotherapy and one dose of atrasentan and were evaluable for toxicity, response, and survival (intent-to-treat population). Patient demographics are shown in Table 1 . The study population included 26 (59%) men and 18 (41%) women with a median age of 61.5 years (range, 23-78). All but two patients were white (95.5%) and only nine (20.5%) had performance status of 0. The predominant histology was adenocarcinoma (21 patients, 47.7%) and 38 (86.4%) patients had stage IV disease. Seven patients with brain metastasis were enrolled (15.9%).

Treatment duration, safety, and tolerability. The total number of chemotherapy cycles administered was 166 (Table 1). Median number of cycles was 4 (range, 1-6). The median duration of treatment with single-agent atrasentan was 76.5 days (range, 11-385; mean ± SD, 95.5 ± 90.4). Twenty patients (45.5%) failed to go on to single-agent atrasentan (16 progressive disease, 2 stable disease, and 2 withdrawals).

All grade 3 and 4 treatment-related adverse events are reported in Table 2 . Two patients were removed from study due to excessive toxicity: one with intractable nausea and vomiting (the only study dose-limiting toxicity) and one for infection. Toxicity was tolerable with grade 3/4 ≥10%: lymphopenia (10 of 44, 22.7%), neutropenia (9 of 44, 20.5%), dyspnea (3 of 44, 11.4%), and hyperglycemia (3 of 44, 11.4%). There were no toxic deaths.

Pharmacokinetics. Data were obtained from all 10 patients enrolled during phase I. Neither clinically nor statistically significant changes in paclitaxel clearance (mean ± SD, 21.2 ± 4.5 L/h pre-atrasentan versus 21.3 ± 4.9 L/h post-atrasentan) or other pharmacokinetic variables (C_MAX, t_1/2, area under the curve, and volume of distribution) were observed (Fig. 2 ) and they resembled values previously described for paclitaxel (26). This allowed for transition from the phase I to the phase II components without chemotherapy dose modifications. Pre- and post-atrasentan paclitaxel plasma concentrations peaked at the end of the infusion (3 h) and mean concentrations reached >5,300 ng/mL.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Pharmacokinetics. Graph, comparison of the plasma paclitaxel concentrations (mean ± SD) versus time profiles (pre- and post-atrasentan). Symbols represent actual measured level. Pre- and post-atrasentan mean (±SD) paclitaxel pharmacokinetic variables. P values are for a paired t test of means. AUC, area under the curve; VD, volume of distribution.

Response and survival. Because all phase I patients received full doses of the three treatment drugs, they were counted toward the calculated phase II sample size. Efficacy results are shown in Table 3 . In the intent-to-treat population (44 patients), 0 and 8 patients achieved complete response and partial response, respectively, for an overall response rate of 18.2% [95% confidence interval (95% CI), 8.2-32.7]. Another 16 patients achieved stable disease and 20 had progressive disease. The PFS, median survival, and 1-year survival in the intent-to-treat population were 4.2 months (95% CI, 3.2-5.8), 10.6 months (95% CI, 8.8-14.5), and 43%, respectively (Table 3). After removal from study, 31 patients received second-line therapy (pemetrexed, 15; docetaxel, 2; gemcitabine-navelbine, 3; gefitinib/erlotinib, 7; and clinical trials, 4).

Discussion

This is the first clinical trial to address the safety and efficacy of atrasentan, a selective ET_AR antagonist, in combination with paclitaxel-carboplatin as first-line therapy for advanced NSCLC. Due to the distinctive characteristics of our trial, a phase I/II study design was chosen and this allowed for (a) the evaluation of the potential for atrasentan to inhibit paclitaxel metabolism [clearance primarily mediated by the cytochrome P450 system (CYP3A and CYP2C9); ref. 26], and (b) a timely determination of the safety and efficacy of the regimen.

Given the potential for atrasentan to inhibit paclitaxel metabolism [clearance primarily mediated by the cytochrome P450 system (CYP3A and CYP2C9); ref. 26], potential pharmacokinetic interactions between these drugs were also evaluated. However, because hepatic mechanisms of elimination are not significant for carboplatin, a pharmacokinetic interaction with atrasentan was neither anticipated (31) nor tested. Another distinctive characteristic of our study was the selection of a fixed, 10 mg atrasentan dose. This decision was based on pharmacokinetic studies showing that daily oral doses of atrasentan between 2.5 and 30 mg achieved plasma concentrations exceeding the K_i for the ET_AR and on pharmacodynamic studies showing physiologic effects in humans at such doses. Additionally, evidence of higher adverse events and of potential loss of ET_AR selectivity, with increasing doses (>20 mg) of atrasentan, also influenced the dose selection decision (32).

Because atrasentan plasma concentrations 100-fold higher (10 μmol/L) than those achieved after multiple 10 mg daily doses are required to produce a 10% inhibition of CYP-dependent enzyme activity (atrasentan concentration-dependent inhibition of CYP3A and CYP2C9 activities reached IC₅₀ values of 3 and 35 μmol/L, respectively), no atrasentan-paclitaxel interaction was expected. Therefore, to optimize the study duration, we limited the pharmacokinetic analysis to the first 10 patients (phase I). A mean paclitaxel clearance difference of >33% was considered clinically significant, but we found no effect from daily treatment with atrasentan on the pharmacokinetics of paclitaxel (Fig. 2).

A large body of evidence indicates that the cytotoxic activity of taxanes or platinum can be enhanced by combination with atrasentan (33–35). Specifically, atrasentan ET_AR blockade results in sensitization to taxane-induced (or platinum-induced) apoptosis (33), and it enhanced paclitaxel-induced vascular endothelial growth factor inhibition with additional reduction in microvessel density (34). These atrasentan properties acquire greater importance considering that a recent phase III study (E4599), where paclitaxel-carboplatin was combined with bevacizumab, an antiangiogenic monoclonal antibody against vascular endothelial growth factor, showed a superior survival in advanced NSCLC but also increased toxicities, including life-threatening pulmonary hemorrhage (36). Given the antiangiogenic and proapoptotic synergism that atrasentan has shown with taxanes or platinum and the conceptual similarities with E4599, our regimen was a logical combination to test in NSCLC.

In our study, toxicities were generally well tolerated and atrasentan did not seem to exacerbate the toxicities of paclitaxel-carboplatin. Among adverse events known to be due to atrasentan, only rhinitis and anemia were observed in >15% of patients. The former was considered drug related but with no effect on compliance. The latter was felt mostly related to chemotherapy because no major signs of hemodilution or fluid retention were seen (only six patients developed edema). Regardless of etiology, the incidence of anemia (all grades and grade 3/4) was similar to other recent studies and hematologic (neutropenia) and nonhematologic (neuropathy) toxicities observed were also in line with those described for paclitaxel-carboplatin (37, 38). No major bleeding complications were observed in our study despite allowing patients with hemoptysis, therapeutic anticoagulation, and/or brain metastasis. Only one patient developed hemoptysis while on atrasentan and an endobronchial lesion was shown as the bleeding source, which stopped spontaneously without atrasentan discontinuation.

In contrast, our efficacy results did not compare so favorably with E4599 (Table 3). Several explanations include the following: (a) the endothelin axis may not be of much importance in NSCLC and preclinical evidence from ovarian cancer studies cannot be extrapolated to NSCLC (this seems unlikely, given the vascular endothelial growth factor mode of action and its clinical importance in NSCLC); (b) although pharmacologically reasonable, 10 mg atrasentan achieves insufficient ET_AR blockade at the tumor level to enhance cytotoxicity (it is known that plasma concentrations do not reflect intratumoral concentrations and that “surrogate” determinations rarely correlate with antitumor effect); and (c) given our less selective patient population (∼16% brain metastasis), inferior results can be expected. In that regard, a more fair comparison would be with the Alpha Oncology trial (∼15% brain metastasis; ref. 39). Here, our results compare substantially better with a median survival of 2.7 months higher (10% higher 1-year survival).

In conclusion, we believe that our inability to reject the null hypothesis in this study may reflect deficiencies in our clinical trial related to the low dose of atrasentan used and that the endothelin axis remains a potentially important pathway to target in NSCLC. Future NSCLC studies with atrasentan should explore higher atrasentan doses (maximum tolerated dose) and include methodologies to test the biological activity of the drug (pre- and post-atrasentan vascular endothelial growth factor and/or ET-1 serum levels, ET-1 and/or ET_AR tumoral expression, etc.).