Clinical responses to selumetinib (AZD6244; ARRY-142886)-based combination therapy stratified by gene mutations in patients with metastatic melanoma

Authors


Abstract

BACKGROUND:

The high prevalence of v-raf murine sarcoma viral oncogene homolog B1 (BRAF) and neuroblastoma v-ras oncogene homolog (NRAS) mutations in melanoma provides a strong rationale to test the clinical efficacy of mitogen-activated protein kinase kinase (MEK) inhibition in this disease. The authors hypothesized that the presence of BRAF or NRAS mutations would correlate with clinical benefit among patients who received treatment with combination regimens that included the MEK inhibitor selumetinib.

METHODS:

BRAF and NRAS mutation status was determined retrospectively in available tissue specimens from patients with melanoma who were enrolled in a phase 1 trial of selumetinib in combination with 1 of 4 drugs (dacarbazine, docetaxel, temsirolimus, or erlotinib). The clinical response rate and the time to progression (TTP) were assessed as a function of BRAF and NRAS mutation status.

RESULTS:

Among 18 patients analyzed, 9 patients (50%) harbored a BRAF mutation (8 had a valine-to-glutamic acid substitution at residue 600 [V600E]; 1 had an arginine nonsense mutation at residue 603 [R603]), 4 patients (22%) harbored an NRAS mutation (2 had a glutamine-to-arginine substitution at residue 61 [Q61R], 1 had a glutamine-to-lysine substitution at residue 61 [Q61K], and 1 had a glycine-to-lysine substitution at residue 12 [G12S]), and 5 patient (28%) had the wild type of both genes. These mutations were mutually exclusive. Among the 9 patients who had BRAF mutations, 5 patients (56%) achieved a partial response, and 4 patients (44%) achieved stable disease for at least 6 weeks. No patient with the wild-type BRAF gene achieved a clinical response (P = .01 vs patients with BRAF mutations). The presence of an NRAS mutation did not correlate with the clinical response rate. The presence of a BRAF mutation was correlated significantly with the TTP in a multivariate model (hazard ratio, 0.22; P = .02 vs wild-type BRAF).

CONCLUSIONS:

Higher response rates and longer TTP were observed with selumetinib-containing regimens in patients who had tumors that harbored a BRAF mutation compared with patients who had wild-type BRAF. Cancer 2013. © 2012 American Cancer Society.

INTRODUCTION

Melanoma is the most aggressive form of skin cancer. In 2011, an estimated 166,000 new cases will be diagnosed in developed countries.1 The median survival of patients with stage IV melanoma is only 6 to 8 months. This is largely because of the chemoresistance of the disease. US Food and Drug Administration-approved agents for metastatic melanoma include dacarbazine, interleukin-2, and ipilimumab, which provide response rates in the range from 7% to 15%.2-4 Vemurafenib, a selective RAF (proto-oncogene serine/threonine protein) inhibitor, recently was approved for the treatment of patients who have melanoma with the v-raf murine sarcoma viral oncogene homolog B1 (BRAF) valine-to-glutamic acid substitution at residue 600 (V600E) mutation because of its clinically significant survival benefit. The median progression-free survival for these patients is only approximately 6 months, however, and the majority of patients ultimately have disease progression within 1 year after starting treatment.5-7

Genetic sequencing reveals that the majority of melanomas harbor activating mutations in genes of the mitogen-activated protein kinase (MAPK) pathway.8 The most common mutation, which occurs in approximately 50% of cutaneous melanomas, results in an amino acid change in the BRAF gene at codon 600. This point mutation leads to a change in amino acid from valine to glutamic acid, less often to lysine, and occasionally to aspartic acid, and results in constitutive activation of BRAF kinase activity.9, 10 Another 15% to 20% of melanomas have a mutation in neuroblastoma v-ras oncogene homolog (NRAS), which is upstream of BRAF in the MAPK signaling pathway. For the most part, these 2 gene mutations are mutually exclusive.9 These findings have generated much interest in targeting the MAPK pathway, by inhibiting either BRAF or downstream MAPK kinase (MEK). Pharmacologic RAS (rat sarcoma) inhibition has proved unsuccessful to date.

Selumetinib (AZD6244; ARRY-142886) is an oral, selective MEK1 and MEK2 inhibitor that does not compete with adenosine triphosphate for binding to the MEK proteins. In preclinical testing at concentrations up to 10 μM/L, this agent was exquisitely selective for MEK protein and did not significantly inhibit up to 40 other kinases.11 Moreover, in studies of cell lines from melanoma, colon cancer, and nonsmall cell lung cancer, those lines that harbored a RAS or RAF mutation appeared to be more sensitive to selumetinib than cell lines that were wild type for these genes.12 Subsequently, selumetinib (hydrogen-sulfate formulation) was evaluated in a phase 1 trial in which it reportedly was well tolerated at a dose of 75 mg twice daily (the phase 2 recommended dose).13

In vitro studies have demonstrated synergistic or additive activity for selumetinib in combination with various cytotoxic or molecularly targeted drugs. In a v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS)-mutant colorectal cancer cell line, selumetinib plus irinotecan reportedly was additive, whereas the combination of selumetinib plus docetaxel was synergistic.12 Haass and colleagues similarly demonstrated that the combination of selumetinib and docetaxel produced synergistic growth inhibition and apoptosis induction in a BRAF V600E-mutant melanoma xenograft model.14

On the basis of the promising preclinical activity of selumetinib-containing combination regimens for various tumor types, including melanoma, a phase 1 study of selumetinib in combination with 1 of 4 different drugs (dacarbazine, docetaxel, erlotinib, or temsirolimus) in patients with advanced solid tumors was conducted (National Clinical Trial NCT00600496). We used available tumor samples from patients who were enrolled in that trial to test the hypothesis that the presence of activating BRAF and NRAS mutations correlates with clinical benefit from these selumetinib-containing combination regimens.

MATERIALS AND METHODS

An open-label, multicenter, phase 1 study (National Clinical Trial NCT00600496) was conducted to evaluate the safety, tolerability and pharmacokinetics of twice-daily, oral doses of selumetinib in combination with dacarbazine, docetaxel, erlotinib, or temsirolimus. The protocol for this trial was approved by the institutional review boards of each participating center in compliance with the Declaration of Helsinki. All patients provided written informed consent before enrollment.

Patient Selection

Eligible patients were aged ≥18 years who had refractory solid tumors and a World Health Organization (WHO) performance status of 0 or 1. Among all participating patients in this phase 1 study, we selected only patients who had metastatic melanoma for our retrospective genomic analysis. These patients had stage IV or unresectable stage III disease with at least 1 measurable lesion as defined by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.0. Patients had to have adequate organ and bone marrow function. Patients with brain metastases or spinal cord compression were excluded from the study unless they had received treatment and were stable for at least 1 month and were off glucocorticoids by the time of enrollment.

Treatment Plan

Two dose levels of selumetinib were explored in a 3 + 3 escalation design. Patients received oral selumetinib at a starting dose of 50 mg twice daily, continuously, with dose escalation to 75 mg twice daily after 3 to 6 patients in the previous cohort tolerated the starting dose. Selumetinib was combined with intravenous dacarbazine 1000 mg/m2 on day 1, intravenous docetaxel 75 mg/m2 on day 1, oral erlotinib 100 mg daily, or intravenous temsirolimus 25 mg weekly. A treatment cycle was defined as 3 weeks. The choice of the combination was at the discretion of the local investigator.

Response Evaluation

Radiologic assessment of tumor(s) was performed every 2 cycles. All patients who received at least 1 dose of the study drugs were considered evaluable for response. Clinical responses were defined according to RECIST version 1.0. The time to progression (TTP) was measured from the start of treatment to the time of documented progressive disease.

Mutation Analysis

DNA was extracted from formalin-fixed, paraffin-embedded tumor samples that had been verified by a pathologist for the presence of tumor and had been manually microdissected to enrich tumor content, usually to >75% of total sample nuclei. The pyrosequencing method, a polymerase chain reaction-based DNA-sequencing technique, which detects a bioluminescence signal caused by the release of pyrophosphate when a nucleotide is introduced into the DNA strand, was used to detect mutations in BRAF exon 15 or NRAS exons 2 and 3 (focused on codons 12, 13, and 61). The lower limit of detection of this assay was approximately 1 mutation-bearing cell per 10 cells (10%). This method also detects mutations in codons adjacent to those listed as the focus of the assay and has a higher sensitivity than Sanger sequencing. When necessary, traditional Sanger sequencing was used to clarify a result. All mutation analyses were performed at The University of Texas M. D. Anderson Cancer Center.

Statistical Analysis

For the current post hoc analysis, the Fisher exact test was used to determine the relation of gene mutation status with the clinical response rate and the dichotomized TTP (≤12 weeks vs >12 weeks). The median TTP was presented for each of the BRAF and NRAS subgroups using Kaplan-Meier analysis and was compared between the mutant and wild-type groups using the Wilcoxon rank-sum test. All P values < .05 were considered statistically significant. All analyses were performed using SAS version 9.2 (SAS Institute Inc., Cary, NC).

RESULTS

Patient Characteristics

From March 2008 to June 2009, 29 patients with metastatic cutaneous melanoma were enrolled and received treatment with 4 different selumetinib-based combinations. Eighteen of those patients had adequate tumor tissues available for mutation analysis. The baseline characteristics of these patients are provided in Table 1. Nine patients (50%) harbored a BRAF mutation (8 V600E mutations; 1 arginine nonsense mutation at residue 603 [R603]), and 4 (22%) patients harbored a mutation in NRAS (2 glutamine-to-arginine substitution at residue 61 [Q61R]; 1 glutamine-to-lysine substitution at residue 61 [Q61K]; and 1 glycine-to-lysine substitution at residue 12 [G12S]). The NRAS sequence could not be amplified on 1 patient. Overall, the patients had a good WHO performance status, and most had either M1b or M1c, stage IV disease; 3 patients had stable, treated brain metastases. The number of previously received systemic treatments was distributed evenly across all groups (Table 1). No patient had received a BRAF or MEK inhibitor before enrollment, and only 1 patient had received prior ipilimumab.

Table 1. Baseline Characteristics
 No. of Patients (%)
Patient CharacteristicTotalBRAF MutNRAS MutBRAF WTNRAS WT
  • Abbreviations: BRAF, v-raf murine sarcoma viral oncogene homolog B1; LDH, lactate dehydrogenase; Mut, mutation; NRAS, neuroblastoma v-ras oncogene homolog; WHO, World Health Organization; WT, wild type.

  • a

    Three patients received prior treatment for brain metastasis, including 1 with a BRAF mutation, 1 with an NRAS mutation, and 1 who was wild type for both.

  • b

    These were previous treatments for metastatic disease and did not include adjuvant therapy.

No. of patients1894913
Men13 (72)5 (56)4 (100)8 (89)8 (62)
Age: Median [range], y59 [39-80]52 [39-66]63 [50-80]66 [50-80]55 [39-77]
WHO performance status     
 015 (83)7 (78)4 (100)8 (89)10 (77)
 13 (17)2 (22)0 (0)1 (11)3 (23)
Disease stage     
 III2 (11)0 (0)2 (50)2 (22)0 (0)
 IV16 (89)9 (100)2 (50)7 (78)13 (100)
Metastasis classification     
 M1a1 (5.6)1 (11)0 (0)0 (0)1 (8)
 M1b5 (28)4 (44)0 (0)1 (11)5 (38)
 M1ca10 (56)4 (44)2 (50)6 (66)7 (54)
No. of prior treatmentsb     
 06 (33)3 (33)1 (25)3 (33)4 (31)
 13 (17)1 (11)2 (50)2 (22)1 (8)
 25 (28)3 (33)0 (0(2 (22)5 (38)
 ≥34 (22)2 (22)1 (25)2 (22)3 (23)
Elevated LDH4 (22)2 (22)2 (50)2 (22)1 (8)

Treatment

One patient received treatment during the dose-escalation phase of the study at a dose of 50 mg twice daily, and the remaining 17 patients received oral selumetinib 75 mg twice daily. Three of these patients required a subsequent dose reduction to 50 mg twice daily. The numbers of patients who received the different combination regimens are listed in Table 2. The majority of patients (n = 14; 77%) received the combination of selumetinib and dacarbazine, as expected. For patients who had previously received dacarbazine or who had a contraindication to dacarbazine, docetaxel was the most commonly combined agent (n = 3; 17%). One patient (6%) received the temsirolimus combination. This patient's melanoma harbored a BRAF mutation (R603 nonsense mutation), but NRAS sequencing analysis was uninformative on the tissue specimen. No patient with melanoma in our analysis received the erlotinib combination. The percentages of patients who received the dacarbazine and selumetinib combination were similar among the BRAF mutation, NRAS mutation, and wild-type groups.

Table 2. Selumetinib-Containing Regimens Grouped by Genomic Mutation Status
 No. of Patients (%)
Treatment ArmTotalBRAF MutNRAS MutBRAF WTNRAS WT
  • Abbreviations: BRAF, v-raf murine sarcoma viral oncogene homolog B1; Mut, mutation; NRAS, neuroblastoma v-ras oncogene homolog; WT, wild type.

  • a

    Status was unknown.

No. of patients1894913
Dacarbazine14 (77)7 (78)3 (75)7 (78)11 (85)
Docetaxel3 (17)1 (11)1 (25)2 (22)2 (15)
Temsirolimus1 (6)1 (11)a0 (0)a

Response by Mutation

Overall, there were 5 partial responses (28%) and no complete responses among the 18 patients. All 5 responders had a BRAF mutation (Table 3). The remaining 4 patients with a BRAF mutation had stable disease as their best response; none of the patients with a BRAF mutation had disease progression within the first 12 weeks. In contrast, no patient with a wild-type BRAF tumor achieved a complete or partial response, and 4 of those 9 patients had disease progression before week 12. The difference in the response rate was statistically significant between the 2 groups (P = .01). In contrast, there was no statistically significant difference in the clinical response between patients with and without an activating NRAS mutation (P = .76) (Table 3).

Table 3. Clinical Responses Stratified by Genomic Mutation Status
 Response: No. of Patients 
Mutation StatusPartial ResponseStable DiseaseProgressive DiseasePa
  • Abbreviations: BRAF, v-raf murine sarcoma viral oncogene homolog B1; NRAS, neuroblastoma v-ras oncogene homolog; WT, wild type.

  • a

    P values were determined using the Fisher exact test.

BRAF   .01
 Mutation (n = 9)540 
 WT (n = 9)054 
NRAS   .76
 Mutation (n = 4)031 
 WT (n = 13)463 
Overall, n = 18594 

The median TTP was 19.5 weeks for all 18 patients. On univariate analysis, there was a trend toward an improvement in the TTP for patients who had a BRAF mutation (median, 51 weeks) compared with patients who had the wild-type BRAF gene (median, 12 weeks; P = .08) (Table 4). However, in a multivariate Cox proportional-hazards model, the association between the median TTP and the presence of a BRAF mutation was statistically significant (hazard ratio, 0.22; 95% confidence interval, 0.06-0.82; P = .02) (Table 5). There was no statistically significant correlation between the TTP and NRAS mutation status, sex, age, WHO performance status, or disease stage. Kaplan-Meier curves for overall survival and the TTP according to BRAF status are provided in Figure 1.

Figure 1.

These Kaplan-Meier curves illustrate (A) overall survival (OS) and (B) the time to progression (TTP) according to v-raf murine sarcoma viral oncogene homolog B1 (BRAF) mutation status. WT indicates wild type.

Table 4. Time to Progression Stratified by Genomic Mutation Status
 TTP, wk   
Mutation Status>12≤12PaMedian TTP, wkPb
  • Abbreviations: BRAF, v-raf murine sarcoma viral oncogene homolog B1; NRAS, neuroblastoma v-ras oncogene B1; TTP, time to progression; WT, wild type.

  • a

    These P values were determined using the Fisher exact test.

  • b

    These P values were determined using the Wilcoxon rank-sum test.

BRAF     
 Mutation (n = 9)90.0351.08
 WT (n = 9)45 12 
NRAS     
 Mutation (n = 4)22.5415.5.96
 WT(n = 13)103 20 
Table 5. Multivariate Analysis of the Time to Progressiona
VariableNo. of Patients With ProgressionMedian TTP, wkHR95% CI for HRMultivariate P
  • Abbreviations: BRAF, v-raf murine sarcoma viral oncogene homolog B1; CI, confidence interval; HR, hazard ratio; NRAS, neuroblastoma v-ras oncogene; TTP, time to progression; WT, wild type.

  • a

    The multivariate Cox proportional hazards model suggested that BRAF status is an important predictor of TTP. Patients with WT BRAF were more likely to experience disease progression than patients with BRAF mutations. There were no statistically significant correlations between TTP and sex, age, World Health Organization performance status, or disease stage.

BRAF status     
 WT (n = 9)8121.00
 Mutation (n = 9)8510.220.06-0.82.02
NRAS status     
 WT (n = 13)12201.00 
 Mutation (n = 4)315.50.310.07-1.49.14

At the time of the current retrospective analysis with a median follow-up duration of 91 weeks, the median overall survival for patients who had a BRAF mutation was not reached, whereas the median overall survival was 60 weeks for patients with wild-type BRAF tumors.

DISCUSSION

Before clinical development of the selumetinib hydrogen-sulfate formulation, another formulation (selumetinib freebase) was evaluated in clinical trials. In a phase 1 study of oral selumetinib freebase in patients with metastatic solid tumors, the dose of 100 mg twice daily was established as the maximum tolerated dose. This dose was evaluated in a randomized phase 2 study in chemotherapy-naive patients with metastatic melanoma in which selumetinib freebase was compared with temozolomide.15 For the primary endpoint of progression-free survival, there was no apparent difference in efficacy between the 2 drugs; and, in a subset analysis, there was no difference in the clinical response rate between patients with BRAF mutations (11% response rate in both arms). It was noted, however, that 5 of the 6 patients who had a partial response to selumetinib freebase had BRAF-mutant tumors, whereas 3 of 9 responders to temozolomide had BRAF-mutant tumors. Because monotherapy with selumetinib freebase failed to demonstrate superiority to the standard of care in patients with advanced melanoma, and given evidence in preclinical models that the activity of selumetinib can be enhanced in combination with a variety of agents, the phase 1 study evaluating selumetinib with dacarbazine, docetaxel, erlotinib, or temsirolimus was conducted.

In the current study, we retrospectively analyzed tissue from patients with metastatic melanoma who were participating in a phase 1 study of selumetinib in combination with dacarbazine, docetaxel, erlotinib, or temsirolimus. Our hypothesis was that the presence of activating BRAF and/or NRAS mutations would correlate with clinical benefit. In our analysis of a subset of patients with metastatic melanoma who had tumor tissue samples available for BRAF sequencing analysis, we observed that patients who had melanoma harboring a BRAF mutation had a higher response rate and a longer TTP than patients who had wild-type BRAF. All of the responders to the study regimens had a BRAF mutation, and none of the patients with a BRAF mutation had disease progression within the first 4 cycles (12 weeks) of treatment. The median TTP of the BRAF mutation group was 51 weeks. In contrast, 4 of 9 patients without a BRAF mutation had rapid disease progression within the first 12 weeks of treatment. The median TTP of the BRAF wild-type cohort was 12 weeks. Our findings suggest that BRAF status is an important predictor of the TTP based on a multivariate Cox proportional-hazards model. Without data from a randomized trial, however, the predictive versus prognostic effects of BRAF status cannot be distinguished. In the current study, we did not observe any statistically significant correlation between the TTP and other factors, such as sex, age, WHO performance status, or disease stage, probably because of the small number of patients in our analysis.

We also observed no association between clinical outcome and NRAS mutation status. The small number of patients with an NRAS mutation made it impossible to draw any meaningful conclusions regarding the clinical benefit of selumetinib-containing combination regimens in the NRAS population. It is also not known whether trametinib, another selective MEK inhibitor, has clinical activity in this patient subset. However, the results from a recent phase 1 study of MEK162, another MEK inhibitor, demonstrated that 6 of 30 patients with metastatic melanoma harboring an NRAS mutation had an objective clinical response.16 This finding suggests that MEK inhibition alone can lead to clinically meaningful activity in this patient subset. It is not clear whether the specific drug, the type of NRAS mutation, or simply the small sample size of each study can explain the discrepant results among these agents. We need further investigation involving a larger cohort of patients to evaluate the true contribution of MEK inhibition in this patient subset.

There are several selective MEK inhibitors currently in clinical development. Among them, trametinib already has demonstrated a survival benefit over chemotherapy in patients with BRAF-mutated melanoma. This was most recently demonstrated in a randomized phase 3 study, which reported a clinical response rate of 22% with trametinib.17 A recent phase 2 study of MEK162 also produced a response rate of 23% among 35 evaluable patients.16 Once again, the reasons for the higher response rates to trametinib and MEK162 compared with selumetinib remain unclear. It would be interesting to determine whether the newer formulation of selumetinib (hydrogen-sulfate) has clinical activity similar to that observed with other selective MEK inhibitors.

The observed trend toward a clinical benefit in patients with a BRAF mutation is not surprising in light of preclinical data indicating the sensitivity of BRAF-mutated tumor cells to selective MEK inhibitors and the results from a phase 1 study of a MEK inhibitor, which demonstrated a greater frequency of tumor reduction in patients with BRAF-mutated melanoma.18, 19 However, the randomized phase 2 study of selumetinib versus temozolomide failed to demonstrate a meaningful clinical benefit from selumetinib monotherapy.15 It is unlikely that the cytotoxic chemotherapy agents alone were responsible for the superior clinical outcome in patients with BRAF mutations in our analysis, because there is no evidence that melanoma cells harboring a BRAF mutation are more sensitive to cytotoxic drugs. In fact, the clinical response rate and the progression-free survival of patients with BRAF mutations who received dacarbazine treatment in a phase 3 study comparing vemurafenib and dacarbazine were similar to what was demonstrated previously in other clinical studies of dacarbazine in patients who were unselected for BRAF mutation status.7, 20, 21 Our analysis suggests that the sensitivity of BRAF-mutated melanoma to selumetinib may be enhanced in combination with certain cytotoxic or targeted drug(s), and further combination studies may be appropriate in this population.

The BRAF mutation has become recognized as an essential predictive marker for BRAF inhibitors (vemurafenib, dabrafenib).5, 6, 22 On the basis of the superior survival benefit demonstrated in a phase 3 study, vemurafenib is now a standard of care for this patient population. Unfortunately, however, a majority of patients will develop resistance to the BRAF inhibitor, and these patients will need other targeted drug treatments. MEK and the downstream extracellular regulated kinase 1 and 2 (ERK1/2) proteins become reactivated at the time of treatment resistance to a BRAF inhibitor in some melanoma cell lines,23-27 whereas these proteins continue to be inhibited at the time of treatment resistance in other cell lines.27, 28 In the MEK-dependent resistance model, the melanoma cells appear to depend on the activation of the MEK protein through several mechanisms, such as the development of a secondary NRAS mutation,27 BRAF kinase amplification or alternate splicing,26, 29 or the transactivation of other RAF kinases23 with the reversal of cell growth upon the addition of an MEK inhibitor. In addition, COT (Cancer Osaka Thyroid; also known as MAP3K8) overexpression has been implicated in MAPK activation at the time of resistance.24 MEK-independent resistance mechanisms include insulin-like growth factor pathway signaling and activation of the phosphotidylinosityl-3 kinase (PI3K) pathway.27, 28 This heterogeneity in the mechanism of resistance also has been demonstrated in patients' tumor samples after treatment with a selective BRAF inhibitor and implicates another stratagem, phosphatase and tensin homolog (PTEN) loss, as a putative manner of resistance.25 Therefore, selumetinib (or other MEK inhibitor)-based therapy potentially may be useful in patients whose BRAF-mutated melanomas are refractory or resistant to a BRAF inhibitor alone, particularly if those tumors continue to exhibit MEK dependence. Still, it is currently unknown whether clinical MEK inhibition after failure of BRAF inhibition has a role in the treatment of this disease in this particular population.

In summary, although our patient numbers were small, our retrospective analysis demonstrates that higher response rates and longer TTP were observed with selumetinib-containing regimens in patients who had tumors that harbored a BRAF mutation rather than wild-type BRAF. Although our analysis demonstrated encouraging results, our findings are not conclusive because of the limited sample size, the lack of a comparator, and the limitations of any retrospective analysis. Our findings will need to be confirmed in a large, randomized controlled study. A randomized phase 2 trial comparing the clinical efficacy of dacarbazine in combination with selumetinib versus dacarbazine alone in patients with BRAF mutation (National Clinical Trial NCT00936221) may confirm the value of therapy targeting MEK in this patient population.

FUNDING SOURCES

No specific funding was disclosed.

CONFLICT OF INTEREST DISCLOSURES

AstraZeneca sponsored the NCT00600496 trial.

Ancillary