Angiogenesis is a critical factor in the propagation and metastasis of many tumor types, including colorectal cancer.
Patrick Boland, MD
Assistant Professor of Medicine
Roswell Park Cancer Institute
Angiogenesis is a critical factor in the propagation and metastasis of many tumor types, including colorectal cancer. Bevacizumab, which targets vascular endothelial growth factor (VEGF)-A, has an established role in the therapy of metastatic colorectal cancer, in first-line and in second-line treatment when used in combination with chemotherapy. Recent data also support an ongoing benefit when antiangiogenic therapy is continued across lines of chemotherapy. After the initial approvals of irinotecan, oxaliplatin, bevacizumab, and the anti-EGFR antibodies, cetuximab and panitumumab, multiple agents failed to improve outcomes in colorectal cancer. In the last 2 years, aflibercept (Zaltrap), a soluble VEGF receptor, and regorafenib (Stivarga) gained approval based upon survival improvements when used in the second-line and refractory settings, respectively. These 2 drugs remain the focus of intense research, with hope that further investigation will uncover methods to optimize benefit through optimal patient selection, improved dosing strategies, and better toxicity management. Herein is a summary of the pivotal trials that resulted in the approval of these drugs, as background to their place in the treatment of advanced colorectal cancer.
Angiogenesis is necessary for normal development and cellular growth, as well as for the promotion and spread of malignant neoplasms. The concept of targeting angiogenic signaling in cancer dates back decades, at least to the initial prescient hypotheses of Folkman concerning the presence of a tumor-angiogenesis factor (TAF), which duly influences endothelial cells, stimulates angiogenesis, and ultimately promotes the growth and spread of malignancies. Folkman also observed that tumor cells seemed to likewise stimulate endothelial cell proliferation.1,2While the initially identified candidate did not itself prove to be all that was hoped, a great amount of research was later pursued addressing this concept. Further work established the presence of a vascular permeability factor (VPF), later identified as VEGF.3,4Presently, it is recognized that there are 5 structurally related molecules that constitute the VEGF family of growth factors: VEGF-A, VEGF-B, VEGF-C, VEGF-D, and placental growth factor (PIGF).5It is believed that these are crucial in the process of angiogenesis, with VEGFA and its cognate receptor VEGFR-2 representing the principal angiogenic signaling locus in vivo. However, the processes of angiogenesis and vasculogenesis remain incompletely understood and the true role of additional VEGF receptors, such as VEGFR-1, is unclear. In addition, multiple additional growth factors and cytokines clearly contribute to the process, many of which may be produced by hypoxic tumor cells and appear capable of replacing VEGF. These include but are not limited to FGF, PDGF-B, HGF, EGF, IL-8, SDF-1α, and Ang-2. Initial efforts to target neoplastic lesions by affecting the process of angiogenesis have yielded success, though modest, potentially owing to the complexity of the process and the multiple compensatory pathways.
Bevacizumab (Avastin), a recombinant humanized monoclonal antibody that targets VEGF-A, was among the first antiangiogenic agents developed and represents the first clinically approved anticancer agent specifically targeting angiogenesis.6Approval was based upon the AVF2107g study, in which patients with untreated metastatic colorectal cancer were administered either IFL (irinotecan + bolus 5-fuorouracil [5-FU]) plus bevacizumab or IFL alone. A significant survival benefit was seen: 20.3 months versus 15.6 months, P <.001.7Progression-free survival (PFS) was similarly extended, accompanied by an increase in the overall response rate (ORR), 44.8% versus 34.8%. Adverse events (AEs) were similar between the 2 arms, with an increase in overall grade 3/4 AEs, driven by hypertension. The Eastern Cooperative Oncology Group (ECOG) 3200 study evaluated bevacizumab when administered as a component of second-line therapy, after progression on 5-FU and irinotecan. Randomization was to one of 3 arms: FOLFOX-4 (5-FU + leucovorin + oxaliplatin) plus bevacizumab; FOLFOX-4 alone; or bevacizumab alone.8The median overall survival (OS) was significantly improved in the combined chemotherapy plus bevacizumab arm compared with the other 2 arms, at 12.9 months, 10.8 months, and 10.2 months, respectively (hazard ratio [HR] = 075; P = .0011). Median PFS was 7.3 months, 4.7 months, and 2.7 months, respectively. The lack of difference in OS in the latter 2 arms is presumably secondary to differences in postprogression therapy.
Following the approval of bevacizumab, the BRiTE registry was initiated in 2004 and collected data on the treatment and outcomes of patients who received first-line bevacizumab-based therapy, drawing from a less frequently selected community-based population. Upon analysis, the data suggested a marked benefit when bevacizumab was continued with subsequent therapy (ie, beyond progression). The median OS of patients who received postprogression chemotherapy with bevacizumab versus without bevacizumab was 31.8 months and 19.9 months, respectively, supporting the hypothesis that antiangiogenic therapy provided continued benefit across lines of chemotherapy.9Notably, hypertension was the only AE reported more frequently in the bevacizumabtreated group. The somewhat disappointing part of this story came when this hypothesis was actually tested, in the TML study (ML18147).10Here, patients who received first-line bevacizumab containing chemotherapy were treated with the alternate cytotoxic backbone, irinotecan or oxaliplatin, depending upon prior exposure, and were randomized to receive bevacizumab plus chemotherapy or chemotherapy alone. Median OS was improved when bevacizumab was continued with a changing chemotherapy backbone, at 11.2 months versus 9.8 months (HR = 0.81; P = .0062). Median PFS improved from 4.1 months to 5.7 months with bevacizumab (HR = 0.68; P <.0001), although responses were low in both arms, namely 5% with bevacizumab and 4% without. Thus, the hypothesis was confirmed: continued VEGF-A blockage through use of bevacizumab across chemotherapies prolongs survival. Of note, the primary benefit appears to be a modest prolongation in disease stability, which translates to several weeks of prolonged survival on averagea step forward, but a small one.
In the refractory setting, bevacizumab has been evaluated as a single agent and in combination with 5-FU. Of relevance, in the National Cancer Institute Treatment Referral Centre trial (TRC-0301), patients who were refractory to 5-FU, oxaliplatin, and irinotecan were treated with bolus or infusional 5-FU and bevacizumab. Stable disease or response was obtained in 33% of patients and a median PFS of 3.5 months was achieved (95% CI: 2-4.3 months).11However, the primary endpoint of the trial was objective response rate, the classic litmus test to gauge treatment efficacy, which stood at just 1%. An additional reported trial evaluated the combination of infusional 5-FU and bevacizumab in treatment of 48 refractory patients, all of whom had progressed on prior oxaliplatin-, irinotecan-, and cetuximab-based regimens. The median time to treatment progression (TTP) was 3.5 months, with a median OS of 7.7 months (95% CI: 3.9-11.9 mo.)12Nevertheless, objective response was again quite uncommon, at 6.25%. This regimen has not been further pursued in this setting.
Ziv-aflibercept, otherwise known as aflibercept (Zaltrap), is a fusion protein, referred to as a soluble VEGF receptor or VEGF-trap, which contains the key domains of hVEGFR-1 and hVEGFR-2 fused to the Fc portion of IgG1.13In mouse models, the VEGFtrap demonstrated near-complete blockade of tumorassociated angiogenesis and significant decrease in microvessel density.14In addition, aflibercept was demonstrated to bind VEGF-A with a much greater affinity (50-fold to 100-fold higher), with a faster association rate, further blocking the signaling of PIGF and VEGF-B through VEGFR-1.15Recent investigation has compared the activity of aflibercept with bevacizumab in patient-derived xenograft models. Complete tumor stasis was observed in the majority of xenografts through aflibercept treatment and, in a distinct minority, bevacizumab treatment.15This suggests that aflibercept has potential for a broader range of activity than does bevacizumab, a differential effect with as of yet unclear translation to the clinical setting.
CLICK TO ENLARGE
*Not yet reported.
AEs indicates adverse events; NR, not reported; OS, overall survival; PFS, progression-free survival.
In initial clinical investigation, aflibercept was shown to be tolerable across a range of doses, with the most common AEs being fatigue (63.8%), nausea (36.2%), and vomiting (27.7%). These were predominantly mild (grade 1) and resolved with discontinuation of therapy.13Toxicities associated with other antiangiogenic agents were also seen: dysphonia (46.8%), hypertension (38.3%), and proteinuria (10.6%). A sharp uptick in the incidence of hypertension, particularly grade 3/4 hypertension, was seen at doses of 4 mg/kg and greater, with maximal VEGF-bound VEGF Trap complex levels at doses of ≥2 mg/kg. Activity was witnessed in several patients, with responses in ovarian cancer and evidence of prolonged (>1 year) stable disease in renal cell carcinoma. Further studies included a phase I investigation of aflibercept combined with FOLFIRI.16Dose-limiting toxicities at the recommended phase II dose included grade 3 proteinuria as well as nephrotic syndrome. However, 6 of the 23 patients with colorectal cancer demonstrated objective response, some of whom had received prior treatment with both oxaliplatin and irinotecan. An expansion cohort was evaluated, at the 4-mg/kg dose, confirming the initial hints of activity, including a handful of responses and numerous patients attaining stable disease.17
The VELOUR study evaluated the benefit of adding aflibercept to FOLFIRI after progression during or shortly following first-line oxaliplatin-based chemotherapy in the metastatic setting.18Patients with relapse within 6 months following oxaliplatin-based adjuvant therapy were allowed. Prior bevacizumab use was permitted, but not required, and was noted in approximately 30% of patients. A statistically significant improvement in median OS was demonstrated, increasing from 12.06 months to 13.5 months (HR 0.817; P = .0032). Two-year OS rates were 28% in the aflibercept arm and 18.7% in the control arm. In addition, median PFS improved from 4.67 to 6.9 months, with RR rising from 11.1% to 19.8%. In subset analysis, there was no clear negative interaction stemming from prior bevacizumab treatment.
Increased rates of grade 3/4 AEs were observed in the experimental arm, including grade 3/4 hypertension and, to a much lesser extent, hemorrhage, arterial thrombotic events, and thromboembolic events. Unexpectedly, AEs typically associated with chemotherapy, not antiangiogenic antibodies, were also witnessed at an increased rate: diarrhea (19.3% vs 7.8%), asthenia (16.9% vs 10.6%), stomatitis and ulceration (13.7% vs 5%), infections (12.3% vs 6.9%), and palmarplantar erythrodysesthesia (2.8% vs 0.5%). As chance would have it, these results were first publicly presented at the same ASCO Annual Meeting as the results from the TML study, which similarly noted an improvement in survival from 9.8 months to 11.2 months, a difference of approximately 1.4 months. Given the near-identical difference in median survival and the suggestion of a slightly worsened AE profile, the approval of aflibercept was not greeted with open arms in the same manner as was the approval of bevacizumab.
The persistent divergence of the OS curves from the VELOUR study suggests a subset of patients who are afforded a more substantial benefit, though it is unclear whether these are patients who would have similarly benefitted from second-line therapy with chemotherapy plus bevacizumab. Post hoc subset analyses were recently conducted, examining clinical criteria that might predict for benefit or lack thereof conferred by aflibercept use in VELOUR. Patients recurring within 6 months following adjuvant oxaliplatin- based chemotherapy (n=124; approximately 10% of the study population) have been previously established to have a poorer prognosis.19On review, these patients appeared to derive no benefit from the addition of aflibercept (HR 1.12; 95% CI: 0.72-1.74). On further analysis, excluding this poor prognosis subset of patients, 2 larger subgroups were identified, including a better efficacy subgroup (ECOG PS 0 or ECOG PS 1 with <2 metastatic sites), wherein the median OS improved from 13.1 months to 16.2 months (HR = 0.73; 95% CI: 0.61-0.86) and 6-month PFS improved from 38% to 63% with the addition of aflibercept to chemotherapy. Additional analysis suggested that patients with liver only metastases were afforded particular benefit compared with the remainder of patients.20Alternatively, in the poor prognosis subgroup, the median OS was unchanged with aflibercept; in fact, there was a trend toward decreased survival, 10.4 months versus 9.6 months, which did not reach significance.
The optimal clinical application of these data is less clear. As the TML study excluded patients who progressed within 3 months following the initiation of first-line chemotherapy (also a poor prognosis subgroup), this subgroup has been cited by some as a population where aflibercept might be preferred over bevacizumab. This has not been directly examined in these subset analyses. The suggestion of limited or absent benefit in poor prognosis subgroups is potential cause to doubt aflibercept as being substantially more efficacious in said setting. For patients with rapid progression or bulky disease, inclusion of an anti-EGFR antibody in the second-line setting may be preferable, when eligible. The therapeutic combination selected in second-line therapy will ultimately depend greatly on the AE profile of therapy, the tumor’s mutational status, and the previous treatment history. Desperately needed and glaringly absent in the 10 years since the approval of bevacizumab is the identification of a reproducible biomarker for efficacy of antiangiogenic therapy. As therapeutic options expand, such an identifier has the potential to provide far more power than the clinical factors described. Given the ongoing studies testing aflibercept, it is critical that the scientific community continue to search for this biomarker.
Finally, it is important to note that the approval of aflibercept and the completed randomized studies provide concrete evidence of benefit only with the combination of FOLFIRI (leucovorin calcium, fluorouracil, irinotecan hydrochloride) in patients not previously exposed to this regimen. Thus, for patients treated with a FOLFIRI-based regimen as first-line therapy or who progress on FOLFIRI, there is no robust data to support the addition of aflibercept as a means to overcome therapeutic resistance. There are also no data to support combination with FOLFOX. Aflibercept has been studied as a single agent in refractory colorectal cancer, demonstrating limited activity: a median PFS of 2 months to 2.4 months; OS of 8.5 months to 10.4 months; and a response rate of 1%.21In summary, until further studies demonstrate a benefit outside of the currently approved setting, the utilization of aflibercept in other manners for colorectal cancer is not recommended.
Regorafenib (Stivarga) is a small molecule tyrosine kinase inhibitor (TKI) with a wide range of targets, including but not limited to VEGFR-1, VEGFR-2, PDGFRβ, TIE2, RET, FGFR1, RAF-1, and BRAF. It was initially developed as a RAF1 inhibitor, and it differs structurally from sorafenib only by the placement of one fluorine. However, regorafenib demonstrates increased pharmacologic potency compared with sorafenib, with a broader spectrum of biochemical activity.22Preclinical data have demonstrated both its ability to interfere with angiogenesis as well as to inhibit tumor growth in a wide spectrum of models, including colorectal cancer. Regorafenib is primarily metabolized in the liver by CYP3A4 and UGT1A9, with 2 active metabolites, M-2 and M-5, which exhibit similar concentrations to regorafenib at steady state. The mean elimination half-life of regorafenib is 28 hours (range 14-58), although that of its metabolite M-5 is significantly longer, at 51 hours (range 32-70). Initial phase I studies examined multiple schedules of delivery, though ultimately the 28-day cycle of 21 days on followed by 7 days off was selected as the schedule to move forward in investigation, at a dose of 160 mg daily.23A continuous regimen of 100 mg daily has also demonstrated safety, though data on efficacy are lacking.
The CORRECT study evaluated the utility of regorafenib in refractory metastatic colorectal cancer. Patients who had previously received standard available chemotherapies and progressed on or within 2 months of the most recent regimen were randomized in a 2:1 fashion to receive regorafenib or best supportive care (BSC) alone.24The primary endpoint of the trial was OS. One hundred percent of patients had previously received therapy with bevacizumab, with >80% of patients having previously progressed while receiving chemotherapy, which included bevacizumab. Median OS improved with regorafenib from 5 months to 6.4 months (HR = 0.77; P = .0052) and median PFS was improved as well, from 1.7 months to 1.9 months (HR = 0.49; P <.0001). Five patients exhibited a partial response in the experimental arm (1%) versus 1 with placebo (0.4%). Disease control (partial response plus disease stabilization) at least 6 weeks after study initiation was observed in 41% of patients on regorafenib versus 15% on BSC.
Grade 3/4 treatment-related AEs were witnessed in 54% of patients assigned to regorafenib versus 14% in the placebo arm. The most frequent of these were hand-foot skin reaction, fatigue, diarrhea, hypertension, and rash or desquamation. Hyperbilirubinemia was observed in 9% of patients in the experimental arm versus 2% with placebo. There was one fatal case compatible with regorafenib-related drug-induced liver injury. Despite these AEs, quality of life as assessed by EORTC QLQ-C30 and EQ-5D was no different between the 2 groups at baseline or the conclusion of therapy. Similar to other agents that target angiogenic signaling, the incidence of hypertension, nose bleeds, and proteinuria were elevated with regorafenib use, though reassuringly, there was no witnessed increase in thrombotic events, bowel perforation was not observed, and significant bleeding was uncommon.
Adverse events tend to present early in the course of therapy at a median of 2 weeks; thus, it has been recommended that patients be seen at least biweekly during the first 2 cycles of therapy and within a week of initiating treatment. For common hand-foot skin reactions, it has been recommended that pre-existing areas of skin damage (calluses) be addressed via exfoliation pretreatment or use of a keratolytic agent, such as a urea-based ointment, coupled with ongoing use of comfortable, cushioned footwear, moisturizing, and close monitoring. A well-written recent review has gone through specific AE management in greater detail.25
In reviewing these data, it is evident that >50% of patients will demonstrate progressive disease within 2 months of initiating regorafenib, suggesting limited overall efficacy. In practice, the attendant toxicities are readily evident upon initiation of therapy. However, a significant yet unidentified subset of patients maintains disease control at 4 months and 6 months, uncommonly witnessed with supportive care alone. Thus, it is clear that regorafenib provides significant clinical benefit and represents an advance in the treatment of patients with refractory disease; however, much work remains to be done to optimize its implementation. Many treated patients derive limited or no benefit and still incur substantial toxicities. As with other therapies, a means to select patients who are more likely to benefit would markedly increase the efficacy/toxicity ratio and thus, the clinical impact and embrace of this drug. Again, a biomarker remains elusive, though hope remains for identification. The proactive as well as reactive management of anticipated AEs may aid to lessen toxicity. Personalized dosing strategies will be the subject of future study, with the aim of improving utilization of regorafenib. At this time, the approved dose is 160 mg, without data to support a maintained benefit with lesser doses of administration. In any case, regorafenib represents a new standard, but will also serve as a benchmark upon which to improve.
Additional Antiangiogenic Agents of Interest
Ramucirumab (Cyramza) is a fully humanized IgG1 monoclonal antibody antagonist of VEGFR-2.26 Recently, ramucirumab has been evaluated in advanced gastric and gastroesophageal cancers, as a component of second-line therapy. While bevacizumab previously failed to demonstrably improve OS in the first-line treatment of gastroesophageal cancers when added to chemotherapy, ramucirumab increased OS as a single agent in the REGARD study, from 3.8 months to 5.2 months (HR 0.776; P = .047).26Hypertension was increased in the ramucirumab group (16% vs 8%) compared with placebo, although other AEs were similar. Shortly after this was reported, the RAINBOW study demonstrated a similar survival improvement when ramucirumab was added to paclitaxel compared with paclitaxel alone. Median OS increased from 9.6 months to 7.4 months (HR 0.807; P = 0.017).27Grade 3 or greater neutropenia, leukopenia, hypertension, and fatigue were all increased in the experimental arm, though febrile neutropenia was low in both groups (3% vs 2%).
The ramucirumab data are of particular interest as Eli Lilly and Company recently published press releases announcing that the RAISE study reached its primary endpoint of OS.28RAISE (NCT01183780) is a phase III, randomized second-line study in metastatic colorectal cancer, with features common to both the TML and VELOUR trials. Patients who had progressed on first-line therapy with fluoropyrimidine, oxaliplatin, and bevacizumab were randomized to FOLFIRI plus ramucirumab or FOLFIRI plus placebo. Patients must have experienced progression during first-line therapy or within 6 months after conclusion of firstline therapy. Results are to be presented at the 2015 GI Cancers Symposium. These results will be of great interest to establish what role, if any, ramucirumab might play in the treatment of metastatic colorectal cancer.In summary, in metastatic colorectal cancer, the continued inhibition of angiogenic signaling clearly improves patient outcomes, suggesting that there is a role to continue treatment beyond progression, across lines of chemotherapy. The choice to administer bevacizumab as opposed to aflibercept in concert with second-line irinotecan-based chemotherapy remains one of debate, as no head-to-head comparisons exist to establish the optimal treatment paradigm (see Table). Regorafenib represents the first TKI with proven benefit in colorectal cancer. Future studies of both may identify means to improve patient selection and limit toxicity, thus enhancing outcomes. Multiple additional agents are in development that target angiogenesis and may be capable of circumventing current mechanisms of tumor/host resistance relevant to antiangiogenic therapy. This should remain a time of great optimism in the treatment of colorectal cancer.