Abstract
Purpose: To assess the efficacy and safety of dual antiangiogenesis agents, bevacizumab plus trebananib, without chemotherapy, in first-line treatment of metastatic colorectal cancer (mCRC).
Patients and Methods: This open-label phase II study enrolled patients with unresectable mCRC with no prior systemic treatment. All patients received bevacizumab 7.5 mg/kg 3-weekly and trebananib 15 mg/kg weekly. The primary endpoint was disease control [stable disease, partial response (PR), or complete response (CR)] at 6 months (DC6m). Secondary endpoints included toxicity, overall response rate (ORR), progression-free survival (PFS), and overall survival (OS). Exploratory biomarkers in plasma angiogenesis–related proteins, tumor gene expression, and plasma antibodies to tumor antigens were examined.
Results: Forty-five patients were enrolled from four Australian sites. DC6m was 63% [95% confidence interval (CI), 47–77]. ORR was 17% (95% CI, 7–32), comprising of seven PRs. Median duration of response was 20 months (range, 10–48 months). Median PFS was 8.4 months and median OS 31.4 months. Grade 1–2 peripheral edema and joint-related symptoms were common. Overall incidence of grade 3–4 adverse events (AE) of any type was 33% (n = 15). Expected AEs of bevacizumab treatment did not appear to be increased by the addition of trebananib.
Conclusions: In a first-line mCRC population, the dual antiangiogenic combination, bevacizumab plus trebananib, without chemotherapy, was efficacious with durable responses. The toxicity profile of the combination was manageable and did not exceed that expected with bevacizumab +/− chemotherapy. Exploratory biomarker results raise the hypothesis that the antiangiogenic combination may enable the antitumor immune response in immunotolerant colorectal cancer.
Translational Relevance
Targeting tumor angiogenesis with bevacizumab, used in combination with chemotherapy, is standard-of-care treatment for metastatic colorectal cancer (mCRC). Nevertheless, tumors ultimately develop resistance to this treatment, necessitating alternate approaches to inhibiting angiogenesis and new combinations that may enhance the efficacy of bevacizumab. This phase II clinical trial showed that combined inhibition of the proangiogenic molecules, VEGF and angiopoietin (Ang), with bevacizumab and trebananib, respectively, without chemotherapy, was tolerable and efficacious with durable responses in patients with previously untreated mCRC. Exploratory biomarker analyses found that low baseline CD8 expression was associated with tumor response and antibodies to common tumor antigens were globally increased during therapy, raising the hypothesis that the antiangiogenic combination may provide clinical benefit by enabling the antitumor immune response, particularly in immunotolerant colorectal cancer. Future trials combining VEGF-Ang blockade with immune-targeted therapies may be of interest in testing this hypothesis.
Introduction
Targeting tumor angiogenesis with bevacizumab, a mAb inhibiting VEGF-A, administered with chemotherapy, has been shown to provide disease control and may improve survival in patients with metastatic colorectal cancer (mCRC; refs. 1–3). Although bevacizumab in combination with chemotherapy has become a common first-line treatment for mCRC, resistance ultimately develops leading to treatment failure. Optimizing the benefits of angiogenesis inhibition may require targeting alternate angiogenic pathways in combination with bevacizumab.
The Tie receptors and their angiopoietin (Ang) ligands constitute an alternate angiogenesis signaling pathway that has a distinct and complementary role to VEGF and VEGFRs. In tumor models, Ang-2 has been implicated in mediating adaptive tumor resistance to VEGF blockade, and combined blockade of both VEGF-A and Ang-2 synergistically reduced angiogenesis and tumor growth in human colorectal cancer cell line xenograft models (4, 5). In correlative clinical studies, elevated levels of circulating and tumor Ang-2 have been implicated as negative prognostic factors in colorectal cancer (6–8), and high circulating Ang-2 levels have been linked to treatment resistance in mCRC treated with bevacizumab (9). Trebananib is an investigational agent that targets angiogenesis by inhibiting the binding of Ang-1 and Ang-2 to the Tie2 receptor. We hypothesized that combined blockade of VEGF-A and Ang-2, with bevacizumab and trebananib, respectively, would have enhanced clinical activity in mCRC.
The use of bevacizumab and trebananib without chemotherapy allows the determination of efficacy of the biological combination without confounding effects of chemotherapy. In Australia, during the period that this study was conducted, patients with asymptomatic and unresectable metastatic disease may be offered alternative options to combination chemotherapy, including observation and single-agent chemotherapy. This practice is supported by randomized studies that showed no difference in survival for initial versus delayed use of chemotherapy, or sequential versus combination chemotherapy, in patients with asymptomatic and unresectable metastatic disease (10, 11). As such, offering a novel combination of biological agents was considered a reasonable alternative for these patients who may also seek to defer toxicities of chemotherapy. We conducted an open-label, single-arm, phase II study assessing the efficacy and tolerability of bevacizumab plus trebananib without chemotherapy as first-line treatment of mCRC (ClinicalTrials.gov identifier: NCT01249521). Here, we report the primary results of the study, as well as results of exploratory biomarker analyses.
Patients and Methods
Patients
Eligible patients were aged ≥18 years, with Eastern Cooperative Oncology Group (ECOG) performance status ≤2; had mCRC that was unresectable, with no prior chemotherapy (except for adjuvant chemotherapy); and for whom the investigator considered immediate cytotoxic chemotherapy was not required. Measurable or nonmeasurable disease was allowed.
Patients were excluded if they had prior treatment with VEGF inhibitors or Ang inhibitors; immunosuppressive drugs, such as cyclosporine or tacrolimus, within the previous 4 weeks; central nervous system metastases; uncontrolled hypertension; active bleeding disorders within the last 6 months; uncontrolled clinically significant cardiac disease, arrhythmias, or angina pectoris; history of arterial or venous thrombosis within the last 12 months; serious nonhealing wound or ulcer; 24-hour urinary protein >1 g/24 hours (performed if urine dipstick >1+); and were pregnant or lactating.
Study design and procedures
All patients received bevacizumab 7.5 mg/kg 3-weekly and trebananib 15 mg/kg weekly, until disease progression, unacceptable toxicity, withdrawal of consent, or patient/physician request. Dose omission (not reduction) of either drug was permitted for management of toxicities.
Clinical review and assessment for toxicity were undertaken every 3 weeks. Tumor response was measured by CT scan of chest, abdomen, and pelvis every 6 weeks and assessed according to RECIST v1.0 guidelines.
The primary endpoint of the study was disease control rate at 6 months (DC6m), defined as the proportion of patients who had stable disease (SD), partial response (PR), or complete response (CR) at 6 months.
Secondary endpoints included the incidence of grade 3 or 4 adverse events (AE) according to NCI Common Terminology Criteria for Adverse Events version 3; investigator-assessed overall response rate (ORR), defined as CR or PR per RECIST criteria for patients with baseline measurable disease; progression-free survival (PFS), measured from time of enrollment to disease progression, clinical progression, or death; overall survival (OS), measured from time of enrollment to death from any cause; and correlation of exploratory biomarkers with clinical outcomes.
KRAS and BRAF V600E mutations were not assessed as part of study procedures based on clinical practice at the time of study recruitment, but this information was collected retrospectively, where available, from clinical testing. Mismatch repair (MMR) status was assessed by IHC for MLH1, PMS2, MSH4, and MSH6 performed on archival primary tumor sections.
Blood samples were collected at baseline and archival formalin-fixed, paraffin-embedded (FFPE) primary tumor blocks were retrieved, where available, for subsequent biomarker analyses.
All patients provided written informed consent for study participation, including donation of samples for biomarker studies. The study was conducted in accordance with the principles of the Declaration of Helsinki and complied with ICH Good Clinical Practice guidelines. All study procedures, including biomarker exploration, were approved by the Human Research and Ethics Committee at Austin Health (Melbourne, Victoria, Australia; the coordinating site).
Exploratory biomarkers
Plasma angiogenesis–related proteins
Plasma isolated from patient's blood samples taken at baseline was analyzed for 10 candidate proteins in the angiogenesis pathway: Ang-1; Ang-2; soluble Tie2; VEGF-A, B, C, and D; placental growth factor (PLGF); soluble vascular cell adhesion molecule 1 (VCAM-1); and intercellular adhesion molecule-1 (ICAM-1), using a multiplex ELISA. The sensitivity of detection for all analytes assessed was in the pg/mL range, as determined by standard curve analyses for each analyte. The association between levels of each plasma protein and patient outcomes was analyzed as described in the statistical analysis methods.
Tumor gene expression
To explore differences in gene expression biomarkers in response and PFS subgroups, mRNA was extracted from cores of FFPE archival primary tumor specimens, where available. A custom NanoString panel was used to assess the expression of genes of interest in angiogenesis and immune pathways (Supplementary Table S1). Differential expression of these genes was assessed between patients who had tumor response (PR) versus all others (SD/PD), and between patients with PFS above versus below the median.
Plasma antibodies to tumor antigens
To investigate a possible immune-mediated effect of the study treatment, we measured the induction of plasma antibodies against cognate tumor antigens at baseline and during treatment.
Paired plasma samples taken at baseline and 6 weeks after treatment (where available) were screened for tumor antigen–specific antibodies using a custom tumor antigen array (12). Pooled plasma from healthy individuals was also assayed to verify cancer specificity of the array. The array content was selected from the Immunome Protein Array List (Sengenics Corporation), and consisted of 99 functional tumor antigens (mainly cancer-testis antigens; Supplementary Table S2). Individual arrays were incubated with a plasma sample, followed by a fluorescence-labeled anti-human IgG detection antibody to measure antibody titers. Unsupervised clustering of overall antibody profiles was performed as described in statistical analysis methods below.
Statistical analysis
Sample size for the clinical study was proposed on the basis of Simon two-stage design with DC6m as the primary endpoint. It was assumed that DC6m of 50% would be of interest and DC6m of 25% would be of little interest. To reject the null hypothesis with 90% power and 95% statistical significance, the study would proceed past the first stage if >5 of 17 patients achieved DC6m. The study would be considered positive if >13 of 37 patients achieved DC6m.
Confidence intervals (CI) for DC6m and ORR were calculated by the Pearson–Clopper method. Median PFS and OS were estimated by the Kaplan–Meier method.
For biomarker analyses, Fisher exact tests were used to determine the association between dichotomized (median cutoff) biomarker subsets and tumor response groups. Kaplan–Meier methods were used to generate survival curves for dichotomized biomarker subsets.
Cox regression models were used to estimate the HR for PFS and OS associated with baseline plasma protein levels. Where appropriate, multivariable models were fitted, and adjusted for known baseline prognostic factors in the study population and potential confounders that changed the HR estimate by 10% or greater. Kaplan–Meier survival curves were generated to demonstrate the impact of dichotomized biomarker subsets.
NanoString data were analyzed using the NanoString nSolver Software (version 4.0.70, NanoString Technologies). Samples identified by nSolver as having poor quality were excluded from the analysis. nSolver advanced analysis (version 2.0.115) was used to normalize data and analyze differential expression using the simplified negative binomial model, including the RNA extraction batch as a covariate.
Antibody titers from the custom tumor antigen array were processed using the protein microarray analyzer software (13), and represented as average relative fluorescence units. Average antibody titers for each antigen and across all antigens were calculated and fold change was determined between timepoints. Hierarchical clustering of patient's antibody profiles using Spearman rank correlation method with average linkage was performed using Multiple Experiment Viewer (14).
All reported P values are two-sided with P < 0.05 considered statistically significant. As these analyses were exploratory in nature, no adjustment for multiple testing was made.
Results
Study population
A total of 45 patients were enrolled from four Australian sites from September 2010 to August 2013. The median age of the population was 68 years; 55% had a left-sided primary, 38% had a right-sided primary, and 1 patient had both (synchronous primaries). Patients had an average of two metastatic sites, predominantly in lung (69%), liver (47%), and lymph nodes (42%; Table 1). The incidence of KRAS exon 2 mutation (55%), BRAF V600E mutation (18%), and deficient MMR (dMMR; 3%), reflected the expected rates of these molecular subgroups in a metastatic population (Table 1).
Baseline characteristics of study population.
The median duration on study was 36 weeks (ranging from 4 to 209 weeks), during which patients received a median of 32 weeks of study treatment, equivalent to 32 doses of trebananib and 10 doses of bevacizumab.
Reasons for study discontinuation were disease progression (n = 34, 76%), AEs (n = 7, 15%), and patient or investigator choice (n = 4, 9%). Following the study, 34 (76%) patients received subsequent systemic therapies, which included standard chemotherapy agents and targeted therapies used in mCRC (Supplementary Table S3). The majority of patients (82%) went on to receive further bevacizumab after study, and all did so in combination with chemotherapy.
Efficacy
Of the 45 patients enrolled, 2 were deemed nonassessable for DC6m and ORR as both patients ceased study protocol before the first tumor response assessment, one due to intracranial hemorrhage and the other due to clinical progression. The former patient was also excluded from PFS analyses. Two patients with nonmeasurable disease were also excluded from ORR assessment.
The primary endpoint, DC6m, was 63% (n = 27/43; 95% CI, 47–77). Response rate in the study was 17% (n = 7/41; 95% CI, 7–32), with seven PRs and no CRs (Fig. 1A). The majority of responses on study treatment were durable, with a median duration of response of 20 months (range, 10–48 months) in patients who achieved PR (Fig. 1B). Twelve (27%) patients had PFS of 12 months or longer.
Waterfall plot of best response in patients with measurable disease (n = 41; A) and swimmer plot for all patients (n = 45; B). Dotted horizontal lines represent RECIST criteria for defining PR and progressive disease (PD).
Median follow-up time in this study was 41.5 months, with 35 events for PFS and 33 events for OS. Median PFS was 8.4 months (95% CI, 6–11.8) and median OS 31.4 months (95% CI, 25.4–41.4; Fig. 2).
Kaplan–Meier estimates of PFS (A) and OS (B) in the study.
Safety
Safety analysis included all patients who received ≥1 dose of the study treatment. Fifteen patients (33%) had a grade 3 or 4 AE of any type (Table 2). Six patients had grade 3 gastrointestinal obstruction, which was attributed to their malignancy rather than treatment. The following AEs, each occurring in 1 patient, led to study discontinuation: lower gastrointestinal tract bleed, gastrointestinal perforation, angina, myocardial infarction, transient ischemic attack, intracranial bleed, and delayed wound healing. There were no treatment-related deaths.
Incidence of all grade 3–4 AEs.
For specific interest, AEs of any grade related to study treatments are presented in Table 3. Notably, AEs relating to bevacizumab treatment did not seem to be increased by the addition of trebananib. Low-grade (grade 1 only) peripheral edema occurred frequently and can be attributed to trebananib, as reported previously (15–18). Notably, grade 1–2 joint-related symptoms (arthralgia and arthritis) were common in this study. Most patients with these symptoms were easily managed with simple analgesics or nonsteroidal anti-inflammatory medication. However, 2 patients who remained on study treatment for more than 3 years were subsequently diagnosed with inflammatory musculoskeletal conditions: one with polymyalgia rheumatica, which responded to oral corticosteroid therapy, and the other with inflammatory arthritis in the hip, thought to be related to study treatment.
Incidence of treatment-related AEs (all grades).
Exploratory biomarkers
Plasma angiogenesis–related proteins
Baseline plasma samples were available from all patients (n = 45) for assessment of angiogenesis-related protein biomarkers. Among the 10 candidate proteins assessed, VEGF-C showed a significant association with PFS, where high plasma VEGF-C levels (at or above the median) were associated with improved PFS (unadjusted HR, 0.44; P = 0.02; adjusted HR, 0.38; 95% CI, 0.19–0.77; P = 0.01; Supplementary Fig. S1A). With regards to OS, Ang-2 showed a significant association, where high plasma Ang-2 level was associated with worse survival (unadjusted HR, 2.04; P = 0.05; adjusted HR, 2.58; 95% CI, 1.21–5.47; P = 0.01; Supplementary Fig. S1B). No associations were found with tumor response.
Tumor gene expression
Tumor mRNA was available from 36 patients for gene expression analysis by a custom NanoString assay. Of the panel of angiogenesis and immune pathway genes assessed, CD8 expression was lower in patients with tumor responses compared with those without tumor responses (or otherwise represented as CD8 higher in nonresponders than responders on the volcano plot; P = 0.03; Fig. 3A and C). Ang-2 expression was significantly higher in patients with longer PFS (above the median vs. below median; P = 0.03; Fig. 3B).
Differences in expression of angiogenesis and immune pathway genes in PFS or response groups. A, Comparing patients with tumor response (PR, responders) versus all others (SD/PD, nonresponders). B, Comparing patients with PFS above or below median. Genes with significant differential expression are highlighted in purple. C, Ranked CD8 expression in all patients assessed (n = 36). Patients with tumor responses (PR) are highlighted.
Notably, the case with an outlying high CD8 expression was a tumor with dMMR (the single case in our cohort), providing internal validation of the NanoString results (Fig. 3C).
Plasma antibodies to tumor antigens
Paired plasma samples at baseline and 6 weeks posttreatment were available from 8 patients. In these patients, global increase in antibody titers was observed after 6 weeks of treatment compared with baseline (Supplementary Fig. S2A). Specifically, de novo antibodies were induced against several members of the cancer-testis antigen family (CSAG1, Cxorf61, SAGE1, SPACA3, SPO11, THEG, and XAGE4), as well as other tumor antigens (FGFR2 and PIK3R1; Supplementary Fig. S2B).
Antibody profiles after 6 weeks of treatment were further analyzed by a hierarchical clustering method to group samples by similarities in antibody profiles. The resulting dendrogram showed separation of an exceptional long-term responder (05, PR and PFS, 48 months) from all other patients (Supplementary Fig. S3). The next cluster separated patients who progressed rapidly (3 months) from those who did not (11, 20, or 41 months), with the exception of 1 patient (07).
Discussion
We report the efficacy and safety of bevacizumab in combination with trebananib, to simultaneously target the proangiogenic molecules, VEGF-A and Ang-2, as first-line treatment for mCRC. This study is unique as it assessed the efficacy and safety of combining antiangiogenic agents in the absence of chemotherapy.
Trebananib (AMG386), an investigational antiangiogenic agent, was studied previously in combination with chemotherapy for subsequent treatment of mCRC after failure of one line of chemotherapy treatment, but trebananib did not prolong PFS compared with placebo plus chemotherapy (18). Trebananib remains in active clinical studies across various tumor types. This study used trebananib in combination with bevacizumab to test the approach of dual inhibition of angiogenesis pathways without chemotherapy.
The primary endpoint was achieved, with a 6-month PFS rate of 63%. Acknowledging the caveats of cross-study comparisons, the durability of disease control is also noteworthy, with a median PFS of 8.4 months, which far exceeds that expected in patients with mCRC who receive bevacizumab monotherapy (median PFS, 2.7 months) or those who receive best supportive care alone (median OS, 5 months; refs. 19, 20). Radiological responses (PR) were also durable, with 3 of the 7 patients who achieved PR maintained response for more than 3 years. Importantly, the response rate of 17% observed in this study is greater than that reported previously for either agent as monotherapy (3%; refs. 15, 20). This finding is consistent with previous preclinical studies that demonstrated additive or synergistic antitumor activity upon dual blockade of VEGF and Ang pathways (5, 21–23).
Recently, the bispecific mAb, vanucizumab, which targets VEGF-A and Ang-2, was tested in a randomized phase II trial comparing vanucizumab plus FOLFOX chemotherapy versus bevacizumab plus FOLFOX in first-line mCRC treatment (24). Vanucizumab failed to demonstrate a PFS benefit over bevacizumab and was associated with an increased incidence of grade 3 hypertension. Comparatively, we did not observe such rates of hypertension in our study, which used the combination of bevacizumab and trebananib without chemotherapy. There appeared to be an increased incidence of joint symptoms of an inflammatory nature, especially in patients who remained on bevacizumab and trebananib treatment for long periods. These symptoms were not observed at this frequency in previous studies of either bevacizumab or trebananib, thus may be a particular side effect of the combination. Overall, the combination of bevacizumab and trebananib was tolerable and did not appear to increase toxicities in excess of those expected with bevacizumab with or without chemotherapy.
The long duration of response and disease control for some patients in this study raised the possibility that the efficacy of this antiangiogenic combination may at least, in part, be immune mediated. In this regard, there is increasing evidence for the impact of angiogenic factors on the immune response in the tumor microenvironment to promote tumor survival. Specifically, VEGF inhibits dendritic cell maturation, antigen presentation, and tumor infiltration by lymphocytes, while promoting immunosuppressive cells, such as regulatory T cell (Treg) and myeloid-derived suppressor cell expansion in the tumor microenvironment (25–28). Furthermore, VEGF promotes immune tolerance in tumors by establishing an immune barrier in the endothelium and enhancing expression of inhibitory checkpoints, such as PD-1, in tumor-infiltrating T cells (29, 30). Ang-2 has also been shown to promote immunosuppression in tumor models by recruiting Tie-2–expressing monocytes that suppress T-cell activation and promote Treg expansion (31). Ang-2 also modulates the expression of immunogenic antigens on tumor cells, thereby allowing immune evasion (32). There is already significant interest in combining antiangiogenic agents with immune checkpoint blockade, and further testing of dual VEGF-Ang2 blockade in this context warrants investigation.
We took the opportunity to conduct exploratory biomarker analyses using biosamples that were available from the study population. In assessing gene expression, we found that CD8 expression was low in the majority of tumors, consistent with previous reports that most colorectal cancers lack an immune cell infiltrate. Nevertheless, in assessing differentially expressed genes between response groups, we observed that patients who experienced PRs to the study treatment had lower levels of CD8 expression in their baseline (untreated) tumors compared with the rest of the study population. A possible explanation for this finding may be that patients with lower tumor-infiltrating CD8 T cells at baseline are those who benefit most from the immune-modulating effects of bevacizumab and trebananib to facilitate infiltration of immune cells into an otherwise “cold” tumor. The possibility that tumors with low CD8 preferentially respond, or conversely those with high CD8 are less likely to respond to antiangiogenic agents, is consistent with findings from our recent biomarker analysis of the phase III MAX trial, which demonstrated that patients with consensus molecular subtype (CMS) 2, which by current understanding are poorly immunogenic with low immune infiltrate, preferentially benefit from first-line bevacizumab treatment compared with CMS1, the subtype characterized by high immunogenicity and high immune infiltrate (33–35).
We were interested in exploring the immune effects of patients on study treatment, but only a small number of patients had stored sequential blood samples on study, and these were not appropriate for assessing circulating immune cell subsets. So, instead, we measured levels of plasma antibodies against cognate tumor antigens as a measure of B-cell–mediated antitumor response. We observed a global increase in tumor antigen–specific antibody titers in all patients after 6 weeks of treatment, which we speculate could be due to a more permissive tumor microenvironment enabled by study treatment, facilitating B-cell maturation and antibody production. However, without comparative samples from patients without treatment, this is purely a hypothesis and we cannot exclude a general antibody response unrelated to the study treatment. Interestingly, however, the cluster analysis of antibody profiles at 6 weeks of treatment suggested a separation of treatment response and PFS groups, which may be proof of concept of a possible treatment-related tumor-specific immune effect.
Finally, our finding that higher tumor Ang-2 expression was associated with longer PFS is consistent with the on-target effect of trebananib. Conversely, higher circulating plasma Ang-2 at baseline was associated with worse OS. We note that previous studies have consistently identified high circulating Ang-2 as a poor prognostic factor across multiple tumor types (6, 36–38). Although treatment targeting Ang-2, as given in this study, could explain the different correlations of tumor and circulating Ang-2, further studies are required to clarify the prognostic and predictive associations of tumor and circulating Ang-2.
Overall, the biomarker analyses in this study were limited by small numbers, the lack of a control arm to distinguish between prognostic and predictive effects, and the lack of sequential, posttreatment samples to assess dynamic changes in biomarkers. Furthermore, with respect to the biomarkers studied in archival tumor tissue, these tumors were primary tumor specimens and as such may not accurately reflect the metastatic disease state at the time of the study. The reported findings are, therefore, exploratory and interpreted as hypothesis generating.
The clinical findings of this study are positive and show enhanced efficacy of combined VEGF-A and Ang-2 inhibition with bevacizumab and trebananib. A criticism may be the inclusion criteria of patients not in immediate need of cytotoxic chemotherapy could have selected a population with better prognosis. Nonetheless, the incidence of known prognostic factors, including dMMR, KRAS/BRAF mutation, sites, and burden of metastatic disease (Table 1), would support that the study population was representative of a standard first-line mCRC population. Furthermore, the median OS in this study population, 31.4 months (95% CI, 25.4–41.4), is consistent with that of contemporary first-line mCRC trials involving biologics (39–43). In addition, in applying a published prognostic nomogram that uses baseline clinicopathologic variables to estimate prognosis of patients with first-line mCRC (44), we found that the predicted median OS of the study cohort was 24 months, which would be within the estimates of a contemporary mCRC population.
In conclusion, this study showed that the dual antiangiogenic combination, bevacizumab plus trebananib, in first-line treatment of mCRC was tolerable and efficacious. This biological combination avoids chemotherapy-associated toxicities while providing durable disease control. Exploratory biomarker findings in the study raise the hypothesis that the antiangiogenic combination may help to enable the antitumor immune response, particularly in immunotolerant colorectal cancer. Future trials combining VEGF-Ang blockade with immune-targeted therapies may be of interest in testing this hypothesis.
Authors' Disclosures
P. Savas reports serving as an uncompensated consultant for Roche Genentech. G. Chong reports other from Amgen, Merck Serono, Bristol Myers Squibb, Pharmacyclics, AstraZeneca, Hutchison Medipharma, BeiGene, and Regeneron outside the submitted work. T.J. Price reports grants from Amgen outside the submitted work. N.C. Tebbutt reports grants and personal fees from Amgen during the conduct of the study. N.C. Tebbutt also reports personal fees and nonfinancial support from Roche and personal fees from Bristol Myers Squibb outside the submitted work, as well as a patent submitted and held by Amgen pending. No disclosures were reported by the other authors.
Authors' Contributions
J. Mooi: Conceptualization, data curation, formal analysis, investigation, visualization, writing–original draft, project administration, writing–review and editing. F. Chionh: Data curation, formal analysis. P. Savas: Resources, data curation, formal analysis, supervision, investigation, writing–review and editing. J. Da Gama Duarte: Data curation, formal analysis, investigation, writing–review and editing. G. Chong: Investigation, writing–review and editing. S. Brown: Data curation, formal analysis, investigation, writing–review and editing. R. Wong: Data curation, formal analysis, investigation, writing–review and editing. T.J. Price: Conceptualization, resources, supervision, investigation, writing–review and editing. A. Wann: Conceptualization, resources, data curation, formal analysis, supervision, investigation, writing–review and editing. E. Skrinos: Conceptualization, resources, data curation, formal analysis, supervision, investigation, project administration, writing–review and editing. J.M. Mariadason: Conceptualization, resources, data curation, formal analysis, supervision, investigation, project administration, writing–review and editing. N.C. Tebbutt: Conceptualization, resources, data curation, formal analysis, supervision, investigation, writing–review and editing.
Acknowledgments
We thank Amgen for supplying the study treatments. J. Da Gama Duarte was funded by Cure Cancer Australia through the Cancer Australia Priority-Driven Cancer Research Scheme. J.M. Mariadason was funded by NHMRC Senior Research Fellowships (1046092). The Olivia Newton-John Cancer Research Institute acknowledges the support of the Operational Infrastructure Support Program, Victorian Government, Australia.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Footnotes
Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).
Clin Cancer Res 2021;27:2159–67
- Received July 20, 2020.
- Revision received September 13, 2020.
- Accepted January 27, 2021.
- Published first January 29, 2021.
- ©2021 American Association for Cancer Research.
References
Volume 27, Issue 8
