The trial was completed by the Adult Brain Tumor Consortium (previously the New Approaches to Brain Tumor Therapy [NABTT] Central Nervous System Consortium), the University of Alabama at Birmingham (Birmingham, AL), University Hospital and University of Lausanne (Lausanne, Switzerland), and the Cancer Therapy Evaluation Program (National Cancer Institute, Bethesda, MD).
Cilengitide is a selective integrin inhibitor that is well tolerated and has demonstrated biologic activity in patients with recurrent malignant glioma. The primary objectives of this randomized phase 2 trial were to determine the safety and efficacy of cilengitide when combined with radiation and temozolomide for patients with newly diagnosed glioblastoma multiforme and to select a dose for comparative clinical testing.
In total, 112 patients were accrued. Eighteen patients received standard radiation and temozolomide with cilengitide in a safety run-in phase followed by a randomized phase 2 trial with 94 patients assigned to either a 500 mg dose group or 2000 mg dose group. The trial was designed to estimate overall survival benefit compared with a New Approaches to Brain Tumor Therapy (NABTT) Consortium internal historic control and data from the published European Organization for Research and Treatment of Cancer (EORTC) trial EORTC 26981.
Cilengitide at all doses studied was well tolerated with radiation and temozolomide. The median survival was 19.7 months for all patients, 17.4 months for the patients in the 500 mg dose group, 20.8 months for patients in the 2000 mg dose group, 30 months for patients who had methylated O6-methylguanine-DNA methyltransferase (MGMT) status, and 17.4 months for patients who had unmethylated MGMT status. For patients aged ≤70 years, the median survival and survival at 24 months was superior to what was observed in the EORTC trial (20.7 months vs 14.6 months and 41% vs 27%, respectively; P = .008).
Cilengitide (cyclo-L-Arg-Gly-L-Asp-D-Phe-N [Me] L-Val) (EMD 121974; Merck KGaA, Darmstadt, Germany) is a cyclic arginine-glycine-aspartic acid-containing peptide that binds to αvβ3 and αvβ5 with nanomolar affinity.1 In cell-based assays, cilengitide inhibited both αvβ3-mediated and αvβ5-mediated function with 50% inhibitory concentration values in the low-micromolar range and also inhibited angiogenesis in in vitro model systems.2, 3 In preclinical animal studies of glioma in both mice and rat models, cilengitide demonstrated tumor control and survival advantages.4-7
The experience of cilengitide in patients with glioma also has demonstrated biologic activity as measured by responses, progression-free survival (PFS), and overall survival (OS). The initial phase 1 study in recurrent malignant glioma by Nabors et al failed to define a maximum tolerated dose but did suggest that higher doses were more active by measuring improvements in tumor blood flow with perfusion magnetic resonance imaging.8 A randomized phase 2 trial in patients with recurrent disease evaluated low and high doses of the drug and likewise suggested greater biologic activity at the higher dose (2000 mg twice weekly).9 In newly diagnosed patients with glioblastoma multiforme (GBM), cilengitide at a dose of 500 mg twice weekly was combined with standard therapy: radiation and temozolomide (RT+TMZ).10 Compared with historic controls, those authors noted improvements in both PFS and OS that appeared to be enhanced in the O6-methylguanine-DNA methyltransferase (MGMT) methylated patient population.
Thus, the results detailed above provide a rationale for the study of integrin antagonists in the setting of primary brain malignancies. The current study was undertaken to determine the toxicities of cilengitide in patients with newly diagnosed GBM and to determine the OS of patients who receive different doses of the agent.
MATERIALS AND METHODS
This study was sponsored by the Cancer Therapy Evaluation Program at the National Cancer Institute and was conducted by the New Approaches to Brain Tumor Therapy (NABTT) Central Nervous System Consortium. The protocol was reviewed and approved by the institutional review board at each participating institution and all patients signed informed consent.
Patients who were eligible for enrollment met the following criteria: aged ≥18 years; newly diagnosed and histologically proven glioblastoma; maintained on a stable dose of corticosteriods for ≥5 days; recovered from previous surgery; a Karnofsky performance status (KPS) ≥60%; adequate hematologic, renal, and hepatic function; agreed to practice acceptable birth control method; a mini-mental status examination score of ≥15; and capable of providing informed consent.
This trial was designed as an open-label, randomized, phase 2 study to evaluate the safety and efficacy of cilengitide in terms of OS. In addition, it was designed to pick 1 of 2 cilengitide doses in combination with standard RT+TMZ to be used in future clinical trials in this patient population. The phase 2 study was preceded by a safety run-in combining cilengitide with standard chemoradiation for patients with newly diagnosed glioblastoma. In total, 3 predefined doses (500 mg, 1000 mg, and 2000 mg twice weekly) were selected. The study drug was infused intravenously over a 1-hour period on a twice-weekly schedule with a minimum of 72 hours between infusions for 4 weeks. This 4-week interval was considered a treatment cycle. The starting dose and schedule were selected based on previous studies of cilengitide in patients with malignant glioma.8, 9 Toxicity was evaluated according to the National Cancer Institute Common Toxicity Criteria version 3. Temozolomide was administered daily at 75 mg/M2 concurrent with radiation therapy followed by a maintenance phase of 150 to 200 mg/M2 for 5 consecutive days of every 28-day cycle. Radiation therapy consisted of the conventional treatment used by the Radiation Therapy Oncology Group. This consisted of treatment to the tumor plus a generous margin for a total dose of 6000 centigrays in 30 fractions. During the randomized phase 2 portion, patients were randomized into 1 of 2 treatment groups (500 mg or 2000 mg) using a stratified and randomly permuted blocks method. The stratification was based on common prognostic factors of the disease to include age (≤50 years or >50 years), KPS (≤80% or >80%), and tumor status (measurable disease or unmeasurable disease). Patients were received adjuvant temozolomide for 6 months and received cilengitide until disease progression, unacceptable toxicity, or voluntary withdrawal.
Tumor tissue from the patient's surgery at diagnosis was collected to determine MGMT methylation status. Patients underwent evaluation of measurable disease by cranial magnetic resonance imaging 4 weeks after they completed cilengitide plus chemoradiation and again 8 weeks thereafter. Laboratory studies, including a chemistry profile and liver function enzymes, were determined at baseline and before each odd-numbered maintenance cycle. Complete blood counts were obtained at baseline, weekly during radiation therapy, and monthly during maintenance therapy. A urinalysis, electrocardiogram, and chest x-ray were obtained at baseline.
Patient samples were tested for MGMT promoter methylation status using previously described procedures11 by MDxHealth (Liege, Belgium). Briefly, 40 microns of a formalin-fixed, paraffin-embedded tumor sample were deparaffinized and digested, and the genomic DNA was isolated using the standard phenol-chloroform method. Up to 1500 ng of DNA were then treated with sodium bisulfite, resulting in the deamination of nonmethylated cytosine bases to uracil bases. Quantitative polymerase chain reaction analysis for the methylated version of the MGMT gene and for the β-actin gene (ACTB) was then performed on the sample. The β-actin gene was used to establish sample quality and also was used to normalize the test output. A threshold of 1250 copies of ACTB was necessary to properly interpret the sample. The normalized ratio of methylated MGMT (mMGMT) copy numbers to ACTB copy numbers was then calculated (mMGMT/ACTB × 1000). Samples with an mMGMT:ACTB ratio of ≥2 were considered methylated; below this ratio, the samples were considered unmethylated.
The trial was designed to accrue 94 patients, and the analysis required 63 deaths. This would provide 94% power to detect a reduction ≥30% in the hazard of death (the hazard rate) compared with the NABTT internal historic control at an αalpha level of .1. Comparison with European Organization for Research and Treatment of Cancer (EORTC) data also was proposed. Survival was calculated from the date of histologic diagnosis to the date of death from any cause, and surviving patients were censored at the date of last follow-up. The trial also was designed to further evaluate drug safety and to identify either the 500-mg or the 2000-mg dose as the best for future clinical trials in this setting. This trial design was chosen as a consequence of our previous NABTT phase 1 study of this agent, which suggested that toxicities at the 2 dose levels would be similar and that activity occurred across the range of doses tested.8 The overall hazard rate of death was defined as the number of deaths per person-years of follow-up. A decision rule was included in the trial design specifying that a reduction ≥30% in the hazard rate compared with the NABTT internal historic control (which did not include temozolomide-treated patients) would provide strong evidence to proceed to a formal comparative trial.
Patient characteristics and toxicities were summarized using appropriate descriptive statistics. The chi-square statistic was used for proportion comparison, and the Student t test was used for comparisons of continuous data.
The method of Kaplan and Meier12 was used to estimate OS. Confidence intervals (CIs) were calculated using standard methods. A Cox proportional-hazards regression model was used to estimate the hazard ratio for death attributable to the treatment and to prognostic factors. To pick a more efficacious dose in terms of an OS advantage, the 2 doses were compared using a nonparametric log-rank test. All reported P values are 2-sided, and all analyses were performed using the SAS statistical software package (version 9.1.3; SAS Institute, Inc., Cary, NC).
Between April 2005 and December 2007, 112 patients were enrolled into the study. Their characteristics are described in Table 1. All patients had undergone previous surgery and were diagnosed histologically with glioblastoma. The patient characteristics and prognostic factors were well balanced between treatment groups and were comparable to those of historic controls. At the time of data cutoff (May 30, 2010), 6 patients were continuing to receive study drug infusions.
Patient infusion with cilengitide did not produce any acute toxicities. During the safety run-in portion of the study, there were no dose-limiting toxicities at any dose level that were sufficient to define a maximum tolerated dose. During the randomized phase, hematologic and nonhematologic grade 3 or 4 adverse events were distributed relatively evenly, with 48 such events occurring in the 500-mg cohort and 35 occurring in the 2000-mg cohort. Hematologic toxicities comprised 58% of the adverse events among patients in the 500-mg dose cohort and 34% of the adverse events among patients in the 2000-mg cohort. These adverse events were attributed uniformly to RT+TMZ. The nonhematologic toxicities also were spread fairly evenly across the dose groups and exhibited no consistent, concerning toxicity profiles. Table 2 presents the range of adverse events observed.
Table 2. Nonhematologic Grade 3 or 4 Adverse Events
Table 3. Comparison of Baseline and Overall Survival for All Patients and for Patients Ages 18 to 70 Years Only in the Combined Radiation, Temozolomide, and Cilengitide European Organization for Research and Treatment of Cancer Phase 3 Cohort
Ages 18-70 Years: No. of Patients (%)
RT+TMZ+EMD Trial, n = 101
EORTC Phase 3 Trial, n = 287
Abbreviations: CI, confidence interval; EORTC, European Organization for Research and Treatment of Cancer; GBM, glioblastoma; KPS, Karnofsky performance status; OS, overall survival; RT+TMZ+EMD, combined radiation, temozolomide, and cilengitide.
The estimated median survival for all 112 patients enrolled on this study was 19.7 months (95% CI, 16.6-21.9.months). The overall hazard rate of death was 0.494 (95% CI, 0.4-0.6) per person-year of follow-up (Fig. 1), and the estimated 2 year OS rate was 38% (95% CI, 29%-48%). No patients were lost to follow-up.
Compared with the NABTT historic control group, which did not include patients who received treatment with temozolomide, the unadjusted hazard ratio of death was 0.41 (95% CI, 0.32-0.53; P = .0001), which constitutes a 59% reduction in the hazard of death. After adjusting for known prognostic factors (age (P = .0003), KPS (P = .004), and surgical procedure (P = .003), the adjusted hazard ratio of death was 0.39 (95% CI, 0.3-0.5; P < .0001). In this trial, 101 patients were aged ≤70 years. Their median survival was 20.7 months (95% CI, 17.2-24.0 months), which was better than the similarly aged patients in the EORTC study, in which RT+TMZ also was used and the median survival was 14.6 months (95% CI, 13.2-16.8 months). The 24-month survival rate with cilengitide was 41% (95% CI, 31%-51%) compared with the EORTC results of 27% (95% CI, 21%-32%; P = .008). This comparison is summarized in Table 3.
In the randomized portion of the trial for dose selection, 94 patients were assigned to either the 500-mg dose group (46 patients) or the 2000-mg dose group (48 patients). The estimated median OS in the 500-mg dose group was 17.4 months (95% CI, 13.3-21.6 months) compared with a median OS of 20.8 months in the 2000-mg dose group (Fig. 2). The unadjusted hazard ratio of death was 0.83 (95% CI, 0.5-1.3; P = .4), representing a 17% reduction in the hazard of death for patients in the 2000-mg dose group compared with patients in the 500-mg dose group. After adjusting for stratification factors and well known prognostic factors, such as age, KPS, surgical procedure, and MGMT methylation status at baseline, the adjusted hazard ratio of death was 0.8 (95% CI, 0.5-1.28), representing a 20% reduction in the hazard of death for patients in the 2000-mg dose group compared with patients in the 500-mg dose group. The 2-year survival rates among the randomized patients were 28% (95% CI, 16%-44%) and 44% (95% CI, 30%-59%) for the 500-mg dose group and the 2000-mg dose group, respectively.
PFS was not a primary objective of this trial. However, disease progression was recorded on the patient follow-up form as an off-treatment criterion. PFS was calculated from the date of diagnosis to the date of disease progression. In total, 103 of 112 patients (92%) developed progressive disease or died by the data cutoff date. The median PFS was 9.97 months (95% CI, 8.3-11.1 months) for all adult patients and 10.6 months (95% CI, 8.5-13.2 months) for patients ages 18 to 70 years. Among the 94 randomized patients, PFS was similar between the 2 dose groups. The median PFS was 9.5 months (95% CI, 6.9-12.2 months) for the 500-mg dose group and 9.3 months (95% CI, 7.7-11.1 months) for the 2000-mg dose group. After adjusting for age, KPS, surgical procedure, and MGMT methylation status at baseline, the adjusted hazard ratio of PFS was 1.07 (95% CI, 0.67-1.68) for patients in the 2000-mg dose group versus the 500-mg dose group.
O6-methylguanine-DNA methyltransferase status
In total, 88 of 112 patients (79%) had tumor tissue collected at baseline. A neuropathologic review determined that 79 patients had adequate tissue for baseline MGMT assessment. At baseline, 21 of 112 patients (19%) had methylated MGMT status, 48 patients (43%) had unmethylated MGMT status, and 43 patients (38%) had unknown MGMT methylation status; hence, the frequency of methylated MGMT status among evaluable patients was 30% (21 of 69 patients). The patients with unknown MGMT methylation status included those who were tested but were judged invalid either because of an insufficient number of ACTB gene copies or because tumor tissue was not collected or was not available. OS stratified only according to baseline MGMT methylation status for all 112 evaluable patients is illustrated in Figure 3. The estimated median OS of patients who had methylated, unmethylated, and unknown MGMT status was 30 months (95% CI, 16.6-35.4 months), 19.1 months (95% CI, 14.6-21.9 months), and 17.4 months (95% CI, 13.7-24.2 months), respectively.
In this report, we extend our previous studies with the integrin antagonist cilengitide to the newly diagnosed glioblastoma population and augment the safety and toxicity description of the compound with standard radiation and chemotherapy. Overall, cilengitide was well tolerated without new or unexpected toxicities in this patient population when combined with RT+TMZ. Most of the adverse events noted in the trial were consistent with the expected toxicities of standard chemoradiation. Specific events that may be induced or attributed by an agent with antiangiogenic or anti-invasive properties, such as hemorrhage, were not noted. These findings are consistent with other studies of cilengitide for newly diagnosed GBM.10
The primary endpoints of this study were to evaluate OS in patients with newly diagnosed GBM who received treatment with cilengitide in conjunction with standard chemoradiation and to select a dose for comparative clinical trial testing. The trial achieved a 61% reduction in the hazard rate of death compared with the NABTT historic control database, exceeding the protocol decision rule of a reduction ≥30% in the failure rate. Because the NABTT historic database did not include the receipt of temozolomide administered with and after radiation, the analysis planned before the 2005 publication of the EORTC results is of marginal importance in the current therapeutic environment. However, the observed median OS for all 112 patients enrolled on the trial was 19.7 months, which compares very favorably with the 16.1 months reported from a recent study of low-dose cilengitide (500 mg) in the same setting10 and with the 14.6 months noted in the EORTC phase 3 trial,13 which included a better prognostic group of only patients aged ≤70 years. In addition, the patients who received cilengitide had improved 2-year survival for similarly aged patients (41% vs 27%).
Secondary objectives of this study were to examine toxicity and survival in patients treated with low doses (500 mg) and high doses (2000 mg) of cilengitide and to choose the dose that should be studied in future phase 3 trials. The toxicities were similar. However, the median OS was 20.8 months for the 2000-mg group and 17.4 months for the 500-mg group (adjusted hazard ratio of death, 0.8), and the 2-year survival rate was 44% in the 2000-mg group and 28% in the 500-mg group. Although these differences were not statistically different using a Cox regression model (data not shown), the current trial was not intended or powered to conduct such a comparison, as our goal was simply to select the best dose to study in the future. These data strongly suggest that cilengitide at a dose of 2000 mg twice weekly with standard radiation and chemotherapy is the superior choice for phase 3 trials in patients with newly diagnosed GBM.
It has been demonstrated that determining MGMT methylation status in GBM can predict survival.14 We were able to determine MGMT status in 62% of the patients on study and identified 21 patients (30%) with methylated MGMT of 69 patients who had valid test results. In Figure 3, Kaplan-Meier survival curves illustrate improved survival for patients who had methylated MGMT relative to those with unmethylated and unknown MGMT status. This observation is in keeping with other studies. It was not possible for us to make a reasonable analysis of the effect MGMT status on survival based on the dose of cilengitide because of the small numbers of evaluable patients. However, it appears that cilengitide had an impact on survival for patients with unmethylated and methylated MGMT. We observed a median OS of 30 months for the methylated group and 19.1 months for the unmethylated group. This indicates an improvement for both genotypes compared with the results reported from EORTC 26981, in which patients with methylated MGMT who received RT+TMZ had a median OS of 21.7 months compared with 12.7 months for patients with unmethylated MGMT.
The use of an integrin antagonist in the treatment of malignant glioma is emerging as a strong complement to RT+TMZ in patients with newly diagnosed GBM. The ability to exploit the novel integrin-mediated signaling cascade brings to bear a new class of compounds for the treatment of this disease. The randomized nature of the current study provides strong support for use of the 2000-mg dose and lays a foundation for the current comparative trials ongoing with cilengitide (CENTRIC). In addition, the different mechanism of action and excellent tolerability may permit use the combination of cilengitide with antivascular endothelial growth factor agents to treat patients with newly diagnosed GBM.15
This work was supported by a grant from the National Institutes of Health (CA-62475).
CONFLICT OF INTEREST DISCLOSURES
T. Mikkelsen received research funding support from Merck KGaA, and T. Batchelor has acted as a consultant to EMD-Serono (the biopharmaceutical division of Merck KGaA).