A phase 2, single-arm study of ramucirumab in patients with metastatic renal cell carcinoma with disease progression on or intolerance to tyrosine kinase inhibitor therapy
Jorge A. Garcia MD,
Cleveland Clinic Taussig Cancer Institute, Cleveland, Ohio
Corresponding author: Jorge A. Garcia, MD, Departments of Solid Tumor Oncology and Urology, Cleveland Clinic Taussig Cancer Institute, Cleveland Clinic Glickman Urological & Kidney Institute, CCF/CWRU Lerner College of Medicine, 9500 Euclid Avenue/R35, Cleveland, OH 44195; Fax: (216) 444-9464; firstname.lastname@example.org
See related editorial on pages 1604–1607, this issue.
Presented in part as a poster at the 2009 European Cancer Organization/European Society for Medical Oncology Meeting; September 20-24, 2009; Berlin, Germany (Abstract 7108) and at the 2010 ASCO Genitourinary Symposium; March 5-7, 2010; San Francisco, California (Abstract 326).
We acknowledge the patients, their families, and the study personnel who participated in this clinical trial. Writing support was provided by Lori Kornberg (inVentiv Health Clinical) and Patricia Zimmer (ImClone Systems LLC, a wholly owned subsidiary of Eli Lilly and Company).
Multitargeted tyrosine kinase inhibitors (TKIs) have antitumor activity in metastatic renal cell carcinoma (mRCC). Resistance to these agents develops frequently, and their use is often limited by intolerance. Ramucirumab is a recombinant human monoclonal antibody directed against human vascular endothelial growth factor receptor-2. For this study, the authors investigated the clinical efficacy and safety of ramucirumab in patients with TKI-resistant/intolerant mRCC.
In this single-arm phase 2 trial, patients received ramucirumab 8 mg/kg every 2 weeks until they developed disease progression or intolerable toxicity. The primary endpoint was the best objective response rate (ORR); additional endpoints included the disease control rate (DCR), progression-free survival (PFS), the median duration of overall response, and safety.
Thirty-nine patients with RCC received ramucirumab monotherapy. Prior TKI therapy included sunitinib (59% of patients), sunitinib and sorafenib (30.8% of patients), and sorafenib (10.3% of patients). The ORR was 5.1% (95% confidence interval [CI], 0.6%-17.3%). The 12-week DCR was 64.1% (95% CI, 47.2%-78.8%). The median PFS was 7.1 months (95% CI, 4.1-9.7 months), and the median overall survival was 24.8 months (95% CI, 18.9-32.6 months). Grade 3 or higher adverse events that occurred in ≥5% of patients included grade 3 hypertension (7.7%) and proteinuria (5.1%). There was 1 on-study death from multiorgan failure.
Recent biologic insights regarding clear cell renal cell carcinoma (RCC) have led to therapeutic advances, especially regarding frequent alterations in the von Hippel-Lindau (VHL) tumor suppressor gene/signaling pathway.[1, 2] The elucidation of vascular endothelial growth factor (VEGF) up-regulation as a critical effect of VHL inactivation in RCC has led to the development of VEGF pathway inhibitors for this disease. Despite the availability of multiple inhibitors of VEGF, VEGF receptor (VEGFR), and mammalian target of rapamycin (mTOR) in metastatic RCC (mRCC),3-8 complete responses (CRs) are uncommon, and the duration of therapy is often limited by drug intolerance or disease progression. Among approved agents, axitinib and everolimus are the only US Food and Drug Administration (FDA)-approved therapies for patients who have progressive disease after initial systemic therapy.[6, 8]
Ramucirumab (IMC-1121B; LY3009806) is a recombinant human monoclonal receptor-targeted antibody (MoAb) of the immunoglobulin G subclass 1 that specifically binds to the extracellular domain of the human VEGFR-2 with high affinity. Ramucirumab blocks the interaction of VEGFR-2 and VEGF ligands (including VEGF-A) and inhibits VEGF-stimulated endothelial proliferation and migration.[10, 11] In 2 phase 1 trials, ramucirumab was evaluated at doses and schedules ranging from 2 mg/kg per week to 20 mg/kg every 3 weeks.[9, 12] Disease control in excess of 6 months was observed in 30% of patients with diverse, resistant solid malignancies; dose-limiting toxicities were observed infrequently and consisted of hypertension and deep vein thrombosis.
Preclinical models have indicated that antibody-mediated VEGFR-2 inhibition confers disease control in tumor xenografts resistant to VEGFR-2 inhibitory tyrosine kinase inhibitors (TKIs). IN the current, single-arm, phase 2 trial, we examined the efficacy and safety of ramucirumab in patients who had mRCC with resistance or intolerance to oral VEGFR TKIs. The primary endpoint of the trial was the best objective response rate (ORR). Secondary endpoints included safety, progression-free survival (PFS), the 12-week disease control rate (DCR) (CR + partial response [PR] + stable disease [SD]), and the median duration of overall response. Exploratory objectives included serum-based analysis of potentially relevant biomarkers and preliminary correlation of biomarker levels with efficacy.
MATERIALS AND METHODS
Patients aged ≥18 years who had measurable disease (according to Response Evaluation Criteria in Solid Tumors version 1.0 [RECIST v1.0]) and clear cell mRCC were eligible. Other eligibility criteria included prior sunitinib and/or sorafenib therapy with either progressive disease (PD) or intolerance (as determined by the investigator), prior nephrectomy, an Eastern Cooperative Oncology Group performance status from 0 to 1 or a Karnofsky performance status ≥80%, and a serum creatinine level ≤2.2 mg/dL. Exclusion criteria included brain or leptomeningeal metastases, prior bevacizumab, >1 prior immunotherapy regimen (interleukin [IL]-2 or interferon-α), and >2 prior cytotoxic chemotherapy regimens.
Each center's institutional review board approved the protocol in accordance with an assurance filed with and approved by the United States Department of Health and Human Services. This study was conducted in accordance with good clinical practices and the Declaration of Helsinki. Patients provided written informed consent before undergoing study procedures or receiving study treatment.
Treatment and Dose Adjustments
The first dose of ramucirumab was administered within 7 days of enrollment. Patients received 8 mg/kg of ramucirumab infused intravenously over 1 hour on day 1 of each 14-day cycle. Treatment continued until patients developed PD or intolerable toxicity. A phase 2 dose of 8 mg/kg every 2 weeks was selected, because this was associated with minimum drug concentrations at which biologic activity was observed in preclinical models. In addition, in phase 1 studies, antitumor activity was observed at or below this dosage, and a pharmacokinetic evaluation indicated that doses at and above 8 mg/kg were associated with pharmacokinetic profiles suggesting receptor saturation.[9, 12]
Dose modifications were permitted for nonlife-threatening, reversible grade 3 and 4 adverse events (AEs) that resolved to grade ≤1 within 2 weeks and for unresolved, nonlife-threatening hypertension and proteinuria. The infusion rate of ramucirumab was reduced for grade 1 and 2 infusion reactions. Ramucirumab was discontinued for grade 3 and 4 infusion reactions, arterial or venous thromboembolic events, bleeding or hemorrhagic events; grade 4 or poorly controlled hypertension; proteinuria exceeding 3 g/24 hours; or persistent, recurrent proteinuria >2 g/24 hours.
Efficacy and Safety Assessments
Patients were evaluated for response according to RECIST v1.0. Baseline tumor burden was assessed within 21 days before the first dose. Disease imaging occurred every 6 weeks. Assessment of response required confirmation. AEs were graded according to National Cancer Institute Common Terminology Criteria for Adverse Events version 3.
Circulating Biomarker Measurements
The patient serum samples that were assayed for this biomarker analysis were collected before infusion at cycles 1, 4, and 7 and at 30-day follow-up. Samples were used to assay circulating biomarkers using an electrochemiluminescence assay for VEGF/VEGF-A, placental growth factor (PlGF), soluble fms-like tyrosine kinase-1/soluble VEGFR-1 (sFlt-1 [sVEGFR1]), soluble kinase insert domain receptor (sKDR/sVEGFR2), VEGF-C, VEGF-D, hepatocyte growth factor (HGF), basic fibroblast growth factor (bFGF), stromal cell-derived factor (SDF-1A), E-selectin, P-selectin, vascular cell adhesion molecule (VCAM-1), intercellular cell-adhesion molecule-1 (ICAM-1), soluble intercellular adhesion molecule (sICAM-3), C-reactive protein (CRP), angiopoietin-1, angiopoietin-2, mast/stem cell growth factor receptor (c-KIT), IL-4, IL-8, IL-12, serum amyloid A (SAA), and thrombomodulin or mass spectrometry for ceramide (Cer) (C16:0), Cer (C18:0), Cer (C20:0), Cer (C22:0), Cer (C24:0), Cer (24:1), sphinganine-1-phospate, sphingosine, and spingosine-1-phosphate.
Data and Statistical Analysis
This was a phase 2, single-arm, multicenter study (ClinicalTrials.gov NCT 00515697). Efficacy and safety analyses were performed on the intent-to-treat population and included all enrolled patients who received ramucirumab.
The primary efficacy endpoint was best ORR using RECIST v1.0 criteria. The planned sample size was 36 patients to detect an ORR ≥15% under a null hypothesis that a response rate ≤3% would have no interest. This sample size provided an 80% power to demonstrate the assumed response rate at the 5% significance level. Secondary endpoints included safety, PFS, 12-week DCR (CR + PR + SD), and median duration of overall response. PFS was defined as the time from first therapy until radiologic progressive disease or death from any cause.
For exploratory analyses involving correlation of biomarkers with efficacy (overall survival [OS], PFS, binary 12-week disease control [CR, PR, SD vs PD; patients nonevaluable for tumor response were excluded], and best percentage change in target lesion measurements), biomarkers were used as covariates in statistical models as follows: If at least 20% of the results were above or below the limits of quantification, then the biomarkers were treated as binary covariates dichotomized at the upper or lower limit of quantification, respectively. Otherwise, they were treated as: 1) continuous covariates, and, separately as 2) continuous covariates treated as binary with automatic cutoff point selection using an applicable maximal chi-square technique. Cox proportional hazard models, logistic regression models, and linear regression models were used for time-to-event, binary, and continuous efficacy outcomes, respectively.
Patient Characteristics and Treatment
From November 2007 to February 2009, 40 patients were enrolled. At completion of enrollment, several patients were undergoing screening; these patients were subsequently deemed eligible and requested participation. The investigators and the study sponsor concurred that their participation was appropriate. Overall, 39 patients received investigational treatment and are included in the primary efficacy analysis. Table 1 lists the baseline patient characteristics. Fifty-nine percent of patients had received prior sunitinib, 30.8% had received both prior sunitinib and prior sorafenib, and 10.3% had received prior sorafenib. Eighty-five percent of patients discontinued TKIs because of PD. Thirty-one percent of patients had received 1 prior line of systemic therapy, and 69% had received 2 or more prior lines of systemic therapy (range, 1-5 prior lines of systemic therapy). The most frequent sites of metastases were lung (79.5%), lymph nodes (71.8%), and liver (30.8%). Overall, 47.5% of patients were categorized with intermediate-risk disease according to the Memorial Sloan Kettering Cancer Center risk group, as were 62.5% of patients according to the International Metastatic Renal Cell Carcinoma Database Consortium model.
Table 1. Patient Characteristics
No. of Patients (%), N = 39
Abbreviations: ECOG, Eastern Cooperative Oncology Group; IMRCCDC, International Metastatic Renal-Cell Carcinoma Database Consortium model; MSKCC, Memorial Sloan-Kettering Cancer Center; N, population size; NA, not assessable because of missing measurement(s); RCC, renal cell carcinoma; TKI, tyrosine kinase inhibitor.
Disease duration was calculated using (date of informed consent − date of confirmation of RCC + 1)/30.4375.
Other consisted of 1 each of the left posterior triangle, pancreas, pericardial effusion, psoas, spleen, stomach, subcutaneous, and thyroid.
A prior line of therapy was defined as any single agent or combination of agents administered concurrently. A treatment that was temporarily stopped and then restarted was considered a single line of therapy.
The observed ORR was 5.1% (95% confidence interval [CI], 0.6%-17.3%; 1 CR and 1 PR). Twenty-five of 39 patients (64.1%) had SD. One patient achieved a CR approximately 28 months after initial therapy; this patient first achieved a PR approximately 4 months after starting study therapy. A second patient achieved a PR after approximately 5.5 months; the duration of response for these patients was 27.6 months and 2.8 months, respectively. Figure 1 illustrates tumor burden reduction (at best response assessment). The 12-week DCR was 64.1% (95% CI, 47.2%-78.8%). At 12 weeks, 25 patients had SD with no responses observed.
The median PFS was 7.1 months (95% CI, 4.1-9.7 months), and 28% of patients were progression-free at 12 months (95% CI, 14.6%-43.1%) (Fig. 2). For patients who had disease progression on prior TKI therapy, the median PFS was 6.9 months (95% CI, 4.1-10.3 months;) and, for those who discontinued prior TKIs for other reasons, the median PFS was 9.7 months (95% CI, 1.3-13.2 months). The median PFS based on International Metastatic Renal Cell Carcinoma Database Consortium model prognostic score at the time of study entry was as follows: poor, 2.6 months (95% CI, 1.2-6.2 months); intermediate, 4.9 months (95% CI, 1.8-8.5 months); and favorable, 12.9 months (95% CI, 5.5 19.8 months). A post-hoc analysis indicated that the median OS was 24.8 months (95% CI, 18.9-32.6 months).
Subgroup analyses were performed to examine potential associations between hypertension and efficacy (PFS or OS). For the 8 patients who experienced a treatment-emergent AE of hypertension (any grade), the median PFS was 15.9 months (95% CI, 1.6-19.8 months), and the median OS was 30.4 months (95% CI, from 18.9 months to not applicable). For the 31 patients who did not experience a treatment-emergent AE of hypertension, the median PFS was 5.5 months (95% CI, 2.6-8.3 months), and the median OS was 24.3 months (95% CI, 13.8-32.6 months).
Dose Modifications, Duration of Therapy, and Safety
The median duration of therapy was 24 weeks (range, 2-136 weeks), and the median number of infusions was 12 (range, 1-65 infusions). The median relative dose intensity was 98.6% (range, 76.2%-106.7%). Three patients (7.7%) had a dose reduction, and 16 patients (41%) required a dose delay. The duration of therapy on prior sunitinib and/or sorafenib and on ramucirumab on-study, was examined in a post-hoc analysis (Fig. 3). Twenty patients received ramucirumab for more than 5 months without PD, and 14 patients received ramucirumab for a greater duration than they received the immediately preceding TKI. Thirty-three patients (84.6%) received subsequent systemic therapy, including mTOR inhibitors (66.7%), bevacizumab (33.3%), and pazopanib (30.8%).
AEs that resulted in dose modifications/delays in >1 patient were proteinuria (n = 3), acute renal failure (n = 2), hemoptysis (n = 2), hypertension (n = 2), increased blood creatinine (n = 2), and infusion reaction (n = 2). Overall, 37 patients (94.9%) experienced a drug-related AE, defined as an event that was considered possibly, probably, or definitely related to investigational therapy. Of these, 10 patients (25.6%) experienced a grade 3 or 4 AE. The most common drug-related AEs were fatigue (35.9%; 2.6% grade 3), headache (33.3%; 2.6% grade 3), and epistaxis (30.8%; 0% grade 3). The incidence of hypertension was 15.4% (7.7% grade 3) (Table 2).
Table 2. Drug-Related Adverse Events Occurring in ≥10% of Patientsa
These were adverse events that the investigator believed were possibly, probably, or definitely drug-related.
Multiple occurrences of the same preferred term within a system organ class for each patient were counted only once at the highest grade.
Blood and lymphatic system disorders
General disorders and administration site conditions
Metabolism and nutrition disorders
Nervous system disorders
Renal and urinary disorders
Respiratory, thoracic, and mediastinal disorders
Skin and subcutaneous tissue disorders
Six patients (15.4%) experienced a drug-related serious AE (cardiorespiratory arrest, cerebral ischemia, myocardial infarction, proteinuria, hypertensive crisis, and hemoptysis), and 7 patients (17.9%) discontinued because of a drug-related AE (cerebral ischemia, myocardial infarction, elevated liver function test, acute renal failure, proteinuria, hypertensive crisis, hemoptysis). There were 4 serious arteriothromboembolic events (ATEs) or potential ATEs, including hypertensive crisis, myocardial infarction, and cerebral ischemia, and 1 patient died of multiorgan failure (considered unrelated to study treatment) preceded by cardiorespiratory arrest. All of these patients were men and had underlying cardiovascular disease. Three of these events occurred after >1 year of ramucirumab therapy.
Exploratory analyses of potential associations between several biomarkers and efficacy were conducted (data not shown). Higher baseline levels of sFlt-1 (sVEGFR1) demonstrated weak associations with shorter PFS and OS and the absence of 12-week disease control (P = .05, P = .02, and P = .007, respectively, using continuous biomarker values). Additional analyses indicated that higher levels of baseline VEGF-A, VCAM-1, and ICAM-1 also had weak associations with shorter PFS (P = .09, P = .02, and P = .08, respectively, using continuous biomarker values, binary values from maximal chi-square analysis, and continuous biomarker values, respectively). Other correlations also were identified with statistically significant P values; however, plots revealed that these were primarily driven by few data points or were particularly weak correlations. It should be noted that P values were not adjusted for multiple comparisons across markers, clinical endpoints, or analytic methodologies in these exploratory analyses.
Management options for patients with mRCC who have PD or intolerance to VEGFR-TKIs remain limited. Even with the availability of 2 FDA-approved agents in this setting, the precise clinical utility of second-line mTOR or VEGF pathway-directed therapy remains uncertain.
Here, we report the results from a single-arm, phase 2 trial examining the efficacy and safety of ramucirumab in nephrectomized patients with clear cell mRCC who predominantly had resistance to VEGFR-targeted TKIs. Although the observed ORR did not meet the predefined threshold for a successful result, the antitumor activity was clinically meaningful, and the median PFS observed in the study was comparable to that observed with other agents in resistant settings,[6, 8] indicating that the ORR may not have been the optimal endpoint for this trial. The limitations of response as an endpoint in resistant cancer settings are supported by data from the recently presented phase 3 REGARD study (ramucirumab monotherapy for previously treated advanced gastric or gastroesophageal junction adenocarcinoma) indicating that, despite a modest response rate, treatment with second-line, single-agent ramucirumab improved OS and PFS versus placebo in patients with gastric/gastroesophageal junction cancer.
Our results strengthen the concept of sequential VEGFR inhibition in mRCC.,21-23 Patients who are able to receive second-line and subsequent lines of therapy have an inherently favorable prognosis relative to those who are not candidates for further therapy beyond the first line. It is likely that the disease control observed with ramucirumab and the use of subsequent systemic therapies with documented efficacy in RCC (predominantly mTOR inhibitors, bevacizumab, and pazopanib) contributed to the relatively favorable survival observed in this group of substantially pretreated patients. Recent results from the INTORSECT study (a phase 3 trial comparing the efficacy and safety of temsirolimus with sorafenib in patients with mRCC who failed prior therapy with sunitinib malate) indicated an improvement in OS with a VEGFR inhibitor (sorafenib) after documented progression on sunitinib. Although comparisons across studies are inherently limited, our study shares several features with the RECORD-18 (RCC treatment with daily oral everolimus) and AXIS (axitinib vs sorafenib in advanced RCC) trials, which established the efficacy of everolimus and axitinib in the second-line setting. Both of those randomized phase 3 trials had several limitations, including heterogeneity with respect to prior treatment and the issue of drug intolerance versus PD. Our trial population was similar, in that several patients' enrollment was because of drug intolerance to primary therapy and not because of PD. Despite a limited understanding regarding resistance to VEGF/VEGFR inhibitors, it is clear that intolerance to an agent does not reflect the development of biologic resistance in this disease.
Of particular interest is the AXIS trial, which compared the efficacy of axitinib versus sorafenib in patients who had disease progression after front-line therapy. Both agents led to modest RECIST-defined responses that ranged from 9% to 19% in favor of axitinib. In addition, there was a PFS improvement in favor of axitinib. Of interest was the observation that sorafenib in the second-line setting outperformed its own efficacy from a previous study. In AXIS, 54% of the population had received prior sunitinib, and 35% had received cytokine-based therapy without prior VEGFR-inhibitory TKI therapy. On the basis of a subset analysis of prior therapy versus independent review committee assessment, the efficacy of axinitib (relative to sorafenib, especially as assessed by the independent review committee) was observed predominantly in the cytokine-pretreated population as opposed to patients who had received prior sunitinib.
In light of the modest PFS benefits conferred by second-line therapy with available agents, tolerability and safety become important. The AE profile of ramucirumab is quite different from that of any mTOR-targeted or VEGFR-targeted TKI. In our study, we observed AEs similar to those reported with bevacizumab-based therapy; the most frequently observed drug-related AEs were fatigue, headache, and epistaxis. The grade ≥3 incidence of these AEs was generally less than 5%, with the exception of grade 3 hypertension and grade 3 proteinuria, which were observed in 7.7% and 5.1% of patients, respectively. The incidence of ATEs in the current study was higher than that observed in other phase 1 and 2 studies investigating ramucirumab[9, 12, 20] but was similar to that reported by some studies involving the MoAb bevacizumab with chemotherapy.[24, 25] Of note, these events occurred predominantly in patients with underlying cardiovascular risk factors who had received ramucirumab for more than 1 year. Ongoing randomized phase 3 trials will more definitively delineate the incidence of ATEs in patients receiving ramucirumab.
Exploratory biomarker analyses suggested that higher baseline sFlt-1 (sVEGFR1) may have weak associations with shorter PFS and OS and the absence of 12-week disease control. Similarly, higher baseline levels of VEGF-A, VCAM-1, and ICAM-1 had evidence of weak associations with shorter PFS. Because of the single-arm nature of this study, these findings are hypothesis-generating, and no conclusions can be drawn regarding their value as prognostic markers or as markers predictive of ramucirumab efficacy. In addition, P values were not adjusted for multiple comparisons in these exploratory analyses.
Preliminary findings from our study suggest a potential relation between hypertension and longer PFS, as has been demonstrated with other VEGF inhibitors in mRCC.[26, 27] However, analyses investigating an association between treatment-emergent hypertension on ramucirumab and the drug's efficacy are exploratory, and their interpretation is limited both by the lack of an appropriate control, which could mean that any association is confounded by independent patient characteristics associated with a better prognosis, and by the small sample size. Additional study limitations include the single-arm, nonrandomized design; the use of ORR as the primary endpoint; and the use of RECIST criteria to define previous failure on prior therapy.
In conclusion, ramucirumab had a limited impact in the RECIST-defined response of patients with mRCC who either had disease progression on or developed drug intolerance to VEGFR inhibitors. The median and 1-year PFS observed in the study are encouraging. Additional investigation of ramucirumab in mRCC, either as monotherapy or in combination with agents that inhibit mTOR, cMet, or immunomodulatory therapy, is under consideration.
This study was supported by ImClone Systems LLC, a wholly owned subsidiary of Eli Lilly and Company.
CONFLICT OF INTEREST DISCLOSURES
Dr. Choueiri serves on advisory boards for Pfizer, GlaxoSmithKline, Novartis, and Aveo, and has received research funding from Pfizer. Dr. Hozak is a full-time employee of Eli Lilly and Company and owns stock in the company. Dr. Joshi is a full-time employee of Eli Lilly and Company. Drs. Schwartz and Xu are full-time employees of ImClone Systems LLC, a wholly owned subsidiary of Eli Lilly and Company, and they own stock in the company.