Antiangiogenesis therapy in second line metastatic colorectal cancer: similar but different

Abstract

Recent updates in the second-line treatment of patients with metastatic colorectal cancer (mCRC) have indicated the benefit of adding aflibercept (plus FOLFIRI) after an oxaliplatin-containing regimen (with or without bevacizumab) or a bevacizumab-containing regimen after progression on a different bevacizumab-containing regimen. This has led some oncologists to compare the trials and equate the drugs, with the assumption that they confer equal efficacy, but the background of this must be adequately assessed before reaching this conclusion. In fact, the publication of these trial results has left oncologists with much to consider for their mCRC patients who will eventually progress on a first-line chemotherapy/bevacizumab regimen. The author comments on this issue and emphasizes the need for additional research into antiangiogenic biomarkers to identify patients who are likely to respond to one drug more than another.

Keywords::

• aflibercept
• bevacizumab
• colon cancer
• VEGF

1. Introduction

Recent updates in the clinical guidance for second-line treatment of patients with metastatic colorectal cancer (mCRC) have provided new options for patients, but have also left some physicians stymied as to what to do with these choices. Specifically, the National Comprehensive Cancer Network (NCCN) has added both a newly approved agent, aflibercept (known as ziv-aflibercept in the USA), and the well-known agent bevacizumab as antiangiogenic therapy options for patients with mCRC progressing on a first-line regimen that included antiangiogenic therapy with bevacizumab [1]. From a mechanistic standpoint, it is important to distinguish these two antiangiogenic agents, in that unlike bevacizumab, which is a monoclonal antibody targeted specifically to vascular endothelial growth factor-A (VEGF-A), aflibercept acts as a soluble decoy receptor, which binds and inhibits not only VEGF-A, but also VEGF-B and placental growth factor (PlGF) [2-4]. Previously, Phase III data supporting the use of aflibercept were not yet available, and bevacizumab was not recommended for patients with mCRC who progressed on a first-line combination therapy that included bevacizumab (outside of a clinical trial) [5]. All of this changed in 2012 with the reporting of results from two well-designed trials; the VELOUR trial for aflibercept [6] and the ML18147 trial for bevacizumab [7]. The trials had different objectives and varying designs, but much attention has centered on the difference in overall survival between the experimental and control arms of each trial, which, coincidentally happened to be identical in the final analysis [6,7]. This has led some oncologists to compare the trials and equate the drugs, with the assumption that they confer equal efficacy [8], but the background of this must be adequately assessed before reaching this conclusion. In fact, the publication of these trial results has left oncologists with much to consider for their mCRC patients who will, eventually, progress on a first-line chemotherapy/bevacizumab regimen.

The results of VELOUR, published in 2012, demonstrate that the addition of aflibercept to FOLFIRI chemotherapy (irinotecan with leucovorin, followed by infusional fluorouracil, repeated every 2 weeks) significantly improved overall survival, progression-free survival and response rates compared with placebo for patients, inclusive of those who received prior bevacizumab, with mCRC and disease progression following an oxaliplatin-containing regimen [6]. In the primary analysis, the hazard ratio (HR) for overall survival with aflibercept relative to placebo was 0.817 (p = 0.0032) with a median survival of 13.50 versus 12.06 months in the respective groups (net gain of 1.4 months) [6]. Of note, in VELOUR, a pre-specified subanalysis of patients with prior bevacizumab therapy compared with those with no prior bevacizumab therapy demonstrated no significant treatment interaction effects (i.e., similar benefit of aflibercept therapy regardless of whether the patients had, or had not, received prior bevacizumab) [3,9]. In patients who received no prior bevacizumab, the HR for overall survival was 0.788 (0.669 – 0.927), whereas for those with prior bevacizumab use, the HR was 0.862 (0.673 – 1.104). Importantly, this was a subanalysis, and not sufficiently powered to detect significant differences in overall or progression-free survival [9].

The overall results of ML18147, published in early 2013, demonstrated a significant improvement in overall as well as progression-free survival for patients who continued bevacizumab treatment (with an accompanying change in chemotherapy regimen) following progression on a first-line regimen which included bevacizumab [7]. In the primary analysis, the HR for overall survival with continued bevacizumab, versus chemotherapy alone was 0.81 (p = 0.0062), with a median survival of 11.2 and 9.8 months in the respective groups (net gain of 1.4 months) [7]. The NCCN 2013 guidance now allows for either option, aflibercept or bevacizumab, as a component of therapy for patients who received a first-line bevacizumab-containing regimen [1].

Before we evaluate results from these trials, we first must understand differences in the drugs themselves. Bevacizumab, the first antiangiogenesis therapy to be used in mCRC, is a monoclonal antibody that specifically targets what is believed to be the most important mediator of tumor angiogenesis, VEGF-A [1,10]. Inhibition of VEGF-A interaction with its cellular receptor, VEGF receptor 2 (VEGFR2) is believed to cause both an inhibition of tumor vascularization, as well as ‘normalization' of the tumor vasculature, although the precise mechanism of its anti-tumor action when used in combination with chemotherapy is not fully understood [11]. By comparison, aflibercept is a recombinantly produced fusion protein consisting of the high affinity ligand-binding domains of VEGFR2, as well as VEGFR1, and, as noted earlier, acts as a soluble decoy receptor [2,10]. Aflibercept binds VEGF-A, and two additional related ligands that interact with VEGFR1, VEGF-B and PlGF. We now know that the pathways which regulate angiogenesis are complex and dynamic, with multiple overlapping pathways that are capable of crosstalk and regulatory feedback mechanisms with each other [2]. Thus, one rationale behind the design of aflibercept is that in the face of ongoing inhibition of the VEGF-A/VEGFR2 axis with bevacizumab, alternative adaptive mechanisms of angiogenesis can emerge, some of which may be mediated through less-understood VEGFR1-mediated effects, including, among other possible pathways, PlGF-mediated recruitment of proangiogenic myeloid cells, PlGF-stimulated proliferation and/or migration of tumor cells and VEGFR1-mediated seeding at new metastatic sites [12]. Increased PlGF expression has also been associated with poor prognosis in CRC [13], and intriguingly, while the generalizability of the findings are limited, it has been shown that PlGF levels are increased with bevacizumab treatment, with bevacizumab and chemotherapy treatment, and just prior to the onset of progressive disease in mCRC patients [14]. These findings raise the possibility of a contributing effect of PlGF to antiangiogenic escape, which could involve other factors in addition to VEGF-A. Additional preclinical findings also point to some important differences in binding to VEGF-A and corresponding inhibition of VEGF-A-mediated biologic effects [15]. For example, aflibercept's binding to VEGF-A has been reported to be between 10- and 100-fold more efficient than that of bevacizumab in in vitro studies [15,16], while fivefold greater concentrations of bevacizumab than aflibercept were utilized for equivalent inhibition of VEGF-A-induced human endothelial-cell migration [16]. The potential clinical relevance of these differences has not been definitively determined. It could be argued, however, that patients who progress on a prior antiangiogenesis therapy (in this case bevacizumab) would benefit from an agent with a different mechanism of action (in this case aflibercept).

It may be inappropriate to conduct cross-study comparisons when trial designs are different, but it is important to truly understand the substantial differences between them (Table 1). The purpose of the VELOUR trial was to evaluate the safety and efficacy of aflibercept following disease progression on a first-line oxaliplatin-containing regimen. Prior to VELOUR, no therapy had been shown to provide a significant survival benefit in the second line for patients receiving a prior oxaliplatin-containing regimen who were treated with FOLFIRI [6]. The main objective of ML18147 was to evaluate the effect on overall survival of continuing bevacizumab after progression on a bevacizumab-containing first-line regimen. Rationale for the trial was in part provided by preliminary data from non-randomized observational cohort studies (BRiTE and ARIES) suggesting a benefit of continuing bevacizumab therapy beyond progression [7]. This clear distinction between the overall objectives of each trial are important to bear in mind when considering other aspects of the trial design, as well as the results.

Table 1. VELOUR and ML18147 trials: key differences [6,7].

	VELOUR trial	ML18147 trial
Study design	Prospective, randomized, double-blind, placebo control	Prospective, randomized, open label, no placebo control
Antiangiogenic study treatment	Aflibercept versus placebo	Bevacizumab versus no treatment
Chemotherapy backbone	Full dose FOLFIRI	Physician's discretion on dose and regimen of chemotherapy
Study objective	Evaluate the safety and efficacy of aflibercept versus placebo in mCRC patients with progression on a prior oxaliplatin-containing regimen	Evaluate whether bevacizumab could be continued in patients progressing on a first-line regimen containing bevacizumab
Study population	∼ 30% bevacizumab experienced	100% bevacizumab experienced
Study key exclusion criteria	Prior irinotecan; history of major surgery within 28 days; (no exclusion based on timing of progression)	Progressive disease for > 3 months after last bevacizumab dose; first line PFS < 3 months; first line (consecutive) bevacizumab treatment of < 3 months
Study conclusions	Addition of aflibercept to FOLFIRI improved PFS and OS over placebo	Continuation of bevacizumab (with a change in CT) after 1st-progression improved PFS and OS over no treatment (with a change in CT)

CT: Chemotherapy; FOLFIRI: Folinic acid, fluorouracil, and irinotecan; mCRC: Metastatic colorectal cancer; OS: Overall survival; PFS: Progression-free survival; VELOUR: A Multinational, Randomized, Double-blind Study, Comparing the Efficacy of Aflibercept Once Every 2 Weeks Versus Placebo in Patients With Metastatic Colorectal Cancer (MCRC) Treated With Irinotecan/5-FU Combination (FOLFIRI) After Failure of an Oxaliplatin Based Regimen.

In line with its objective of being a first pivotal trial for aflibercept in mCRC patients, VELOUR was designed as a randomized, double-blinded, parallel-arm, Phase III study in which patients were randomized (1:1) to treatment with FOLFIRI in combination with either aflibercept or placebo [6]. Patients were stratified according to their prior bevacizumab use, as well as Eastern Cooperative Oncology Group (ECOG) performance status; the primary end point was overall survival, and the trial was also appropriately designed and powered to assess secondary end points, including progression-free survival and adverse events [6]. The final intent-to-treat (ITT) population in VELOUR (n = 1,226 randomized) was consistent with that of a standard Phase III registration trial population, and was reflective of an ‘all-takers' recruitment approach. All patients had documented progression while on or following an oxaliplatin-containing regimen, but they were not selected based on the timing of that progression [6].

ML18147 was a randomized, prospective, intergroup, open label, Phase III trial. Because the objective was to compare continuation of bevacizumab versus no bevacizumab after progression, patients were randomized (1:1) to a change in chemotherapy with or without bevacizumab, with no placebo comparator. Most patients (∼ 77%) were 42 days or less from their last bevacizumab dose, and the distribution in use of oxaliplatin-based (∼ 42%) or irinotecan-based (∼ 58%) first-line chemotherapy was balanced between the treatment arms [7]. The choice of chemotherapy (to be used in combination with bevacizumab, or not) was varied, and based on the patient's prior regimen, with all approved second-line treatments allowed at the investigator's discretion. Although the distribution of individual chemotherapy regimens was balanced between the treatment arms, a lack of uniformity among the types of chemotherapy used (in as much as these were at the investigator's discretion) is an important limitation to recognize [7]. The primary end point was overall survival, with secondary end points such as progression-free survival also assessed [7]. The final ITT population of ML18147 (n = 819) included adult patients with mCRC and progressive disease following first-line treatment with bevacizumab plus a standard chemotherapy regimen.

The ML18147 population had a number of notable exclusions, including those with progressive disease for > 3 months following the last bevacizumab administration, those with fewer than 3 months of consecutive first-line bevacizumab treatment, as well as those with a first-line progression-free survival of < 3 months [7]. In the latter excluded group, it could be argued that the exclusion of this early recurrence or ‘fast relapse' group might have selected a more favorable outcome population who were more likely to respond to bevacizumab therapy. The trial would also, by virtue of these criteria, have excluded those unable to tolerate antiangiogenic therapy with bevacizumab [7].

2. Implications for clinical practice

The overall conclusions from VELOUR showed that the combination of aflibercept with FOLFIRI significantly improved overall survival and other efficacy outcomes for patients previously treated with an oxaliplatin-containing regimen who are receiving FOLFIRI [6]. Additional subanalysis from VELOUR showed a benefit of aflibercept regardless of prior bevacizumab use, and thus, the results of VELOUR support its use as a second-line therapy in patients receiving a first-line bevacizumab-containing regimen [9]. The overall conclusions of ML18147 suggest that the continuation of bevacizumab (with a change in chemotherapy) improves survival over a change in chemotherapy alone in patients previously receiving a bevacizumab-containing regimen [7]. While results of both trials are welcomed and now provide for further options in the second line, the individual results of the two trials cannot be compared directly. It is important to note that the coincidental 1.4-month gain in overall survival that occurred in both trials is actually the median of a vast number of different patient types (all takers) in the VELOUR trial while it is the median of a highly selected and relatively healthy population in ML18147. In fact, in a substudy (n = 1,102) of VELOUR which only included patients who received oxaliplatin as first-line for mCRC and excluded patients who received oxaliplatin as adjuvant therapy and who progressed within 6 months (n = 124), the median overall survival (95% confidence interval) for placebo/FOLFIRI was 11.9 months compared with 13.8 months for aflibercept/FOLFIRI; a difference of 1.9 months [17]. Lastly, in terms of safety, the data show that aflibercept appears to have an adverse event profile similar to that of bevacizumab and consistent with VEGF inhibition (e.g., hypertension, proteinuria) [4,18]. Any increase in toxicity, for example increased diarrhea in VELOUR over that of ML18147, was likely related to VELOUR patients having had an oxaliplatin regimen followed by a full dose of FOLFIRI, while in the ML18147 trial, initial starting doses as well as subsequent modification of the chemotherapy backbone were at the physician's discretion. Because the two trials were not conducted on the same chemotherapy backbone, there is no basis to conclude that one agent is superior to the other with respect to safety.

With the current range of available therapies, the median survival for unresectable mCRC has improved to 24 months or longer, and by optimizing the sequential use of these treatments, there is room for further improvement [2,3]. Ultimately, decisions regarding second-line therapy must be individualized based on prior therapy and other patient factors. Although a benefit of continuing bevacizumab past first progression was demonstrated in ML18147, we should not assume that continuing bevacizumab is the best option for all patients with first-line progression. One certainly should consider a change in antiangiogenic therapy for the ‘fast-relapse' patients with fewer than 3 months of progression-free survival on a first-line bevacizumab-containing regimen; a population that was entirely excluded from ML18147 [7] or even those who may have had a modest progression-free survival (∼ 6 months) on a first-line chemotherapy-bevacizumab regimen. And, while continuing bevacizumab with a change in chemotherapy backbone would be a viable option for many of these patients, a switch to aflibercept, by virtue of its distinct mechanism of action and higher affinity binding to VEGF-A than bevacizumab, could allow for continued antiangiogenesis therapy and the potential for additional survival benefit. Unfortunately, bevacizumab is used in combination with chemotherapy, and it is not possible at present to unequivocally discern, through the use of biomarkers or otherwise, which patients will continue to respond to antiangiogenic therapy with bevacizumab, and which patients may have developed bevacizumab resistance [2]. In the latter case, such patients would be better candidates for aflibercept treatment as the second-line antiangiogenic therapy. In this regard, there is an urgent need for continued research into biomarkers to help identify in advance those patients who are likely to respond to one antiangiogenesis therapy versus another, and indeed there is precedent for success using such an approach [3]. For example, while initially used in unselected populations, we now know that a benefit of epidermal growth factor receptor targeted therapies (i.e., cetuximab and panitumumab) is observed only in mCRC patients with wild type KRAS status, and thus recommendation for their use is now limited to this subset of patients [1-3]. One can envision a comparative trial of bevacizumab or aflibercept in the second line setting, in a population of high-risk mCRC patients, for example, those with early progression (fewer than 3 months on first line), utilizing biomarkers to identify patient subsets which might benefit from one therapy versus the other. In the absence of such data, however, we must rely on a fair assessment of the available data from VELOUR and ML18147, which both show the benefit of continuing antiangiogenesis in the second line, albeit in two very different trials with varying patient populations.

Declaration of interest

The author has been part of speaker's bureaus for Genentech and Sanofi, no funding was received in preparation of this article.