European Journal of Cancer (2014) 50, 320– 331
`
`A v a i l a b l e a t w w w . s c i e n c e d i r e c t . c o m
`
`ScienceDirect
`
`j o u r n a l h o m e p a g e : w w w . e j c a n c e r . c o m
`
`Aflibercept versus placebo in combination with
`fluorouracil, leucovorin and irinotecan in the treatment of
`previously treated metastatic colorectal cancer:
`Prespecified subgroup analyses from the VELOUR trial
`
`Josep Tabernero a,⇑
`, Eric Van Cutsem b, Radek Lakomy´ c, Jana Prausova´ d, Paul Ruff e,
`Guy A. van Hazel f, Vladimir M. Moiseyenko g, David R. Ferry h, Joseph J. McKendrick i,
`Karen Soussan-Lazard j, Soazig Chevalier j, Carmen J. Allegra k
`
`a Vall d’Hebron University Hospital and Institute of Oncology (VHIO), Universitat Auto` noma de Barcelona, Barcelona, Spain
`b University Hospitals Leuven and KU Leuven, Belgium
`c Masaryk Memorial Cancer Institute, Brno, Czech Republic
`d Fakultni nemocnice v Motole, Praha, Czech Republic
`e University of Witwatersrand Faculty of Health Sciences, Johannesburg, South Africa
`f University of Western Australia, Western Australia, Australia
`g Oncology Research Institute Na.NN.Petrov, St-Petersburg, Russian Federation
`h Russells Hall Hospital, Dudley, West Midlands, UK
`i Monash University, Victoria, Australia
`j Sanofi, Vitry-sur-Seine, France
`k University of Florida, Gainesville, United States
`
`Available online 16 October 2013
`
`KEYWORDS
`Aflibercept
`VEGF Trap
`Bevacizumab
`Metastatic colorectal
`
`Abstract Purpose: The antiangiogenic agent aflibercept (ziv-aflibercept in the United States)
`in combination with 5-fluorouracil,
`leucovorin and irinotecan (FOLFIRI) significantly
`improved survival in a phase III study of patients with metastatic colorectal cancer (mCRC)
`previously treated with an oxaliplatin-based regimen. In the present analysis, outcomes were
`evaluated in prespecified subgroups to assess the consistency of the treatment effect.
`
`Previously presented at: The 2011 European Multidisciplinary Cancer Congress, September 24–27, 2011, Stockholm, Sweden (oral). The 2012
`ASCO Annual Meeting, June 1–5, Chicago, IL, USA (oral).
`Funding: This study was funded by Sanofi, in collaboration with Regeneron Pharmaceuticals.
`ClinicalTrials.gov number, NCT00561470.
`⇑
`Corresponding author: Address: Vall d’Hebron University Hospital, Vall d’Hebron Institute of Oncology (VHIO), Universitat Auto` noma de
`Barcelona, P. Vall d’Hebron, 119-129, 08035 Barcelona, Spain. Tel.: +34 93 274 6085; fax: +34 93 274 6059.
`E-mail address: jtabernero@vhio.net (J. Tabernero).
`
`0959-8049Ó 2013 The Authors. Published by Elsevier Ltd.
`http://dx.doi.org/10.1016/j.ejca.2013.09.013
`
`Open access under CC BY license.
`
`1
`
`JHU 2108
`Merck Sharp v. Johns Hopkins
`IPR2024-00623
`
`

`

`cancer
`Performance status
`Subgroup analysis
`
`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`321
`
`Methods: Patients were randomised to receive FOLFIRI plus aflibercept or placebo every
`2 weeks until disease progression or unacceptable toxicity occurred. Efficacy and safety out-
`comes were analysed with respect to demographic and baseline characteristics, and stratifica-
`tion factors (prior bevacizumab treatment and Eastern Cooperative Oncology Group
`performance status).
`Results: Median overall survival (OS, months [95.34% confidence interval (CI)]) for afliber-
`cept versus placebo was 12.5 (10.8–15.5) versus 11.7 (9.8–13.8) in patients with prior bev-
`acizumab treatment and 13.9 (12.7–15.6) versus 12.4 (11.2–13.5) in patients with no prior
`bevacizumab treatment. The p value for interaction was 0.5668, indicating there was no het-
`erogeneity in these subgroups. For OS and progression-free survival (PFS), there was a signif-
`icantly greater benefit (at the 2-sided 10% level) of treatment for patients with liver only
`metastases versus patients with no liver metastases/liver metastases with other organ involve-
`ment (p value for interaction: 0.0899 [OS]; 0.0076 [PFS]). There was no evidence of heteroge-
`neity in treatment effect in any of the other subgroups examined.
`Conclusions: The benefits of aflibercept in combination with FOLFIRI in patients with
`mCRC previously treated with oxaliplatin were maintained across the specified patient sub-
`groups, including in patients with or without prior bevacizumab treatment.
`Ó 2013 The Authors. Published by Elsevier Ltd.
`Open access under CC BY license.
`
`1. Introduction
`
`The majority of patients (60%) with colorectal cancer
`(CRC) have locally advanced or metastatic disease at
`the time of diagnosis [1]. Prognosis for these patients
`is poor, with a 5-year survival rate of approximately
`12% [1]. However, the administration of molecularly tar-
`geted agents in combination with standard chemother-
`apy regimens has improved the outlook for patients
`with metastatic CRC (mCRC).
`Aflibercept (known as ziv-aflibercept in the U.S.,
`and also known as VEGF Trap or AVE0005) is a
`recombinant fusion protein containing vascular endo-
`thelial growth factor (VEGF)-binding portions from
`the extracellular domains of human VEGF receptors
`1 and 2, fused to the Fc portion of human immuno-
`globulin (Ig)G1. Aflibercept blocks the activity of
`VEGFA, VEGFB and placental growth factor (PlGF),
`by acting as a high-affinity ligand trap to prevent these
`ligands from binding to their endogenous receptors
`[2,3].
`Aflibercept is the first targeted therapy to demon-
`strate a statistically significant and clinically meaningful
`survival benefit in patients previously treated with an
`oxaliplatin-based regimen and receiving the irinotecan-
`containing regimen FOLFIRI (5-fluorouracil, leucovorin
`and irinotecan) for the treatment of metastatic disease
`[4]. The VELOUR (VEGF Trap (aflibercept) with irino-
`tecan in colorectal cancer after failure of oxaliplatin
`regimen) trial evaluated the efficacy and safety of the
`combination of aflibercept and FOLFIRI in patients
`with mCRC, following disease progression while on or
`after completion of treatment with an oxaliplatin-based
`regimen. Adding aflibercept to FOLFIRI significantly
`improved overall survival (OS) relative to placebo plus
`FOLFIRI (hazard ratio [HR], 0.817; 95.34% confidence
`interval
`[CI], 0.713–0.937; p = 0.0032), with median
`
`survival times of 13.50 versus 12.06 months, respectively.
`Aflibercept also significantly improved progression-free
`survival
`(PFS; HR, 0.758; 95% CI, 0.661–0.869;
`p = 0.00007). Median PFS was 6.90 months for afliber-
`cept plus FOLFIRI and 4.67 months for placebo plus
`FOLFIRI. The response rate in VELOUR was 19.8%
`(95% CI, 16.4–23.2%) for aflibercept plus FOLFIRI
`and 11.1% (95% CI, 8.5–13.8%)
`for placebo plus
`FOLFIRI (p = 0.0001). No unexpected toxicities were
`observed in the aflibercept plus FOLFIRI treatment
`arm. As a consequence aflibercept was approved by
`the U.S. Food and Drug Administration (FDA) in
`August 2012 and by the European Medicines Agency
`(EMA) in February 2013 for use in combination with
`FOLFIRI in the treatment of patients with mCRC
`who were resistant
`to or progressed following an
`oxaliplatin-based regimen.
`The purpose of
`this analysis of data from the
`VELOUR trial was to assess the consistency of the ben-
`efit of aflibercept on OS and PFS by evaluating its effect
`in prespecified subgroups. Of particular interest was the
`analysis of outcomes in patients stratified by prior bev-
`acizumab treatment.
`In addition,
`the relationship
`between patient demographics and baseline characteris-
`tics and outcome was investigated.
`
`2. Materials and methods
`
`2.1. Patients and study design
`
`Details of patient eligibility and the design of the
`VELOUR study (a prospective multinational, random-
`ised, double-blind, parallel-arm, phase III study [Clini-
`calTrials.gov number, NCT00561470]) have been
`reported previously [4]. Eligible patients were aged at
`least 18 years; with an Eastern Cooperative Oncology
`Group (ECOG) performance status (PS) score of 0–2.
`
`2
`
`

`

`322
`
`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`Eligible patients had histologically- or cytologically-pro-
`ven colorectal adenocarcinoma with metastatic disease
`not amenable to potentially curative treatment. Patients
`were required to have undergone a single prior oxalipl-
`atin-containing chemotherapy regimen for metastatic
`disease, with documented evidence of disease progres-
`sion during or after treatment completion. Patients
`who relapsed within 6 months of completion of oxalipl-
`atin-based adjuvant therapy were also eligible for inclu-
`sion. Prior bevacizumab was permitted, but not prior
`irinotecan.
`Eligible patients were randomised 1:1 to receive
`FOLFIRI plus either aflibercept 4 mg/kg (aflibercept
`arm) or placebo (control arm) every 2 weeks, with stratifi-
`cation according to prior bevacizumab treatment (yes/no)
`and ECOG PS (0/1/2). Aflibercept or placebo was
`administered intravenously (IV) over 1 h on day 1 of
`each cycle,
`followed immediately by the FOLFIRI
`regimen (irinotecan 180 mg/m2 IV over 90 min, with leu-
`covorin 400 mg/m2 IV over 2 h, followed by 5-FU
`400 mg/m2 bolus and 5-FU 2400 mg/m2 continuous
`infusion over 46 h). Premedication with atropine and
`anti-emetics was permitted. Granulocyte-colony stimu-
`lating factor (G-CSF) was used according to the
`American Society of Clinical Oncology guidelines [5].
`Dose adjustments for each study treatment component
`individually and/or cycle delays (up to 2 weeks) were
`permitted in the event of toxicity. Patients were treated
`until the occurrence of disease progression or unaccept-
`able toxicity as judged by the physician. If FOLFIRI
`was permanently discontinued, patients could continue
`to receive aflibercept/placebo;
`if aflibercept/placebo
`was permanently discontinued, patients could continue
`to receive FOLFIRI. No crossover to aflibercept was
`permitted after progression was documented in the
`control arm.
`At each treatment cycle, patients underwent clinical
`examination and laboratory assessments (including uri-
`nalysis) before receiving study treatment. Adverse events
`(AEs) coded using The Medical Dictionary for Regula-
`tory Activities (MedDRA) version 13.1, to provide a
`system organ class (SOC) and preferred term (PT) for
`each event and graded according to National Cancer
`Institute Common Terminology Criteria for Adverse
`Events v3.0, were recorded. Concomitant medications
`were also recorded. Disease assessment was performed
`every 6 weeks until documented progression. Response
`was assessed according to Response Evaluation Criteria
`In Solid Tumors (v1.0) [6] by a third party (independent
`review committee, IRC), blinded to patient treatment.
`
`2.2. Statistical analysis
`
`included PFS, cutoff date 6th May 2010, defined as the
`interval from randomisation to the first observation of
`disease progression according to IRC review, or death
`from any cause; objective response (complete response
`and partial response), and treatment-emergent adverse
`events (TEAEs) and laboratory abnormalities.
`Prespecified subgroups were defined according to
`demographic variables (age, gender, race, geographic
`region), baseline characteristics (prior hypertension,
`number of metastatic sites [IRC evaluation], liver metas-
`tasis [IRC evaluation], location of primary tumour) and
`stratification factors at randomisation (ECOG PS, prior
`bevacizumab treatment).
`HRs and CI estimates for OS and PFS in the intention-
`to-treat (ITT) population (all randomised patients) and
`within each subgroup were estimated by a Cox propor-
`tional hazards model [4]. Interactions between treatment
`and each subgroup were tested at the 2-sided 10% level
`(i.e. a p-value >0.1 indicates no evidence of heterogeneity
`of treatment effect across the subgroups for each factor).
`Descriptive analysis of adverse events by treatment
`arm was performed within each subgroup in the safety
`population (patients who received at least one dose of
`study treatment).
`All final analyses were
`personnel.
`
`conducted by Sanofi
`
`3. Results
`
`Between November 2007 and March 2010, 1226
`patients were randomised to receive,
`in combination
`with FOLFIRI, either aflibercept (612 patients) or pla-
`cebo (614 patients). Five patients in each treatment
`arm were not treated and four patients randomised to
`placebo received at least one administration of afliber-
`cept,
`therefore the safety population included 611
`patients in the aflibercept arm and 605 in the control
`arm (Fig. 1). At the cutoff date for survival analysis
`(7th February 2011),
`the median follow-up was
`22.28 months.
`
`3.1. Efficacy in demographic subgroups
`
`OS and PFS in patients grouped by age, gender, race
`and geographic region are shown in Figs. 2A and 3A.
`For both efficacy end-points, there was no significant
`treatment interaction between treatment groups and fac-
`tors for any of the demographic subgroups examined. In
`all subgroups, a treatment effect in favour of the afliber-
`cept arm over the control arm (HR < 1.0) was seen for
`both OS and PFS.
`
`The primary end-point of the study was OS, defined
`as the time interval from randomisation to death from
`any cause. The cutoff date for OS was the date of the
`863rd event (7th February 2011). Secondary end-points
`
`3.2. Efficacy in baseline characteristic subgroups
`
`As shown in Figs. 2B and 3B, there was a significantly
`greater benefit of treatment for patients with metastases
`
`3
`
`

`

`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`323
`
`Assessed for eligibility (n=1401)
`
`Excluded (n=175)
`• Not meeƟng inclusion criteria (n=123)
`• Declined to parƟcipate (n=17)
`• Other reasons (n=35)
`
`Randomised (n=1226)
`
`Allocated to FOLFIRI/placebo (n=614)
`• Received allocated intervenƟon (n=605)
`• Did not receive allocated intervenƟon (n=9)
`— 5 paƟents not treated (3 not eligible, 1
`ill, 1 did not wish to conƟnue)
`— 4 paƟents received ≥1 dose of
`aflibercept (dosing irregulariƟes)
`
`DisconƟnued intervenƟon (n=598)
`• Adverse events (n=74)
`• Disease progression (n=437)
`• PaƟent request (n=45)
`•
`Lost to follow-up (n=2)
`• Other (n=40)
`SƟll on treatment (n=11)
`
`Analysed n= 614 (in primary analysis)
`• Excluded from analysis (n=0)
`Pre-specified sub-group analyses (n=614)
`Safety analysis (n=605)
`• Excluded from analysis (n=5, not treated) 4
`paƟents received at least one dose of
`aflibercept, analysed in aflibercept arm
`
`Allocated to FOLFIRI/aflibercept (n=612)
`• Received allocated intervenƟon (n=604)
`• Did not receive allocated intervenƟon (n=8)
`— 5 paƟents not treated (2 not eligible, 1
`PD, 1 withdrawn consent, 1 had no
`medical cover)
`— 3 paƟents received ≥1 dose of placebo
`(dosing irregulariƟes)
`
`DisconƟnued intervenƟon (n=593)
`• Adverse events (n=163)
`• Disease progression (n=305)
`• PaƟent request (n=83)
`•
`Lost to follow-up (n=0)
`• Other (n=42)
`SƟll on treatment (n=14)
`
`Analysed n= 612 (in primary analysis)
`• Excluded from analysis (n=0)
`Pre-specified sub-group analyses (n=612)
`Safety analysis (n=611)
`• Excluded from analysis (n=5, not treated)
`
`Enrollment
`
`AllocaƟon
`
`Follow-up
`
`Analysis
`
`Fig. 1. CONSORT diagram.
`
`confined to the liver versus patients with no liver metas-
`tases or
`liver and metastases at other
`sites
`(OS:
`p = 0.090; PFS: p = 0.008). A consistent treatment effect
`in favour of the aflibercept arm was seen for OS and
`PFS for all subgroups (Figs. 2B and 3B). Of note a
`revised classification was used for the primary tumour
`location (colon/rectosigmoid/other versus rectum). In
`the previous presentations and using the three prespeci-
`fied primary tumour location categories of colon, recto-
`sigmoid/other and rectum, patients with the location
`rectosigmoid/other had a less favourable outcome for
`OS with aflibercept (HR: 1.039 [0.7702, 1.4]).
`
`3.3. Efficacy by baseline stratification factors (prior
`bevacizumab treatment and ECOG PS)
`
`Baseline characteristics for patients stratified by prior
`bevacizumab treatment are shown in Table 1. Patients
`with prior bevacizumab treatment comprised 30% of
`the ITT population and the duration of bevacizumab
`use as well as the antiangiogenic-free period were well
`balanced between the control and aflibercept treatment
`
`arms. For OS, there was no significant interaction at
`the 2-sided 10% level between treatment and stratifica-
`tion levels for prior bevacizumab treatment (yes versus
`no; p = 0.5668) or ECOG PS (0 versus 1 versus 2;
`p = 0.7231), and thus no evidence of heterogeneity in
`treatment effect (Table 2). A difference in favour of aflib-
`ercept over placebo was observed in each stratification
`subgroup.
`Survival outcomes by stratification level of prior bev-
`acizumab treatment are shown in Fig. 4. For patients
`previously treated with bevacizumab, median OS was
`12.5 months in the aflibercept arm and 11.7 months in
`the control arm. Median OS in patients without prior
`exposure to bevacizumab was 13.9 months in the afliber-
`cept arm and 12.4 months in the control arm.
`The analysis of PFS by stratification level also indi-
`cated that there was no significant interaction at the
`2-sided 10% level between treatment and stratification
`levels for prior bevacizumab treatment (yes versus no;
`p = 0.1958) or ECOG PS (0 versus 1 versus 2;
`p = 0.6954) (Table 2). For patients who had received
`prior bevacizumab, median PFS was 6.7 months in the
`
`4
`
`

`

`324
`
`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`A. Demographic subgroups
`N=
`
`Subgroups
`
`HR (95.34% CI)
`
`All patients
`
`1226
`
`0.817 (0.713 to 0.937)
`
`Age <65
`
`Age >=65
`
`Male
`
`Female
`
`783
`
`443
`
`718
`
`508
`
`0.796 (0.67 to 0.945)
`
`0.853 (0.682 to 1.066)
`
`0.83 (0.696 to 0.989)
`
`0.776 (0.625 to 0.963)
`
`Caucasian/White 1071
`
`0.8 (0.691 to 0.925)
`
`Other
`
`Western Europe
`
`Eastern Europe
`
`NorthAmerica
`
`SouthAmerica
`
`Other countries
`
`155
`
`425
`
`297
`
`138
`
`118
`
`248
`
`0.831 (0.561 to 1.23)
`
`0.891 (0.712 to 1.114)
`
`0.697 (0.519 to 0.934)
`
`0.691 (0.442 to 1.079)
`
`0.838 (0.53 to 1.324)
`
`0.891 (0.67 to 1.186)
`
`B. Baseline characteristics
`
`Favours Aflibercept
`
`Favours Placebo
`
`P-value for interaction
`
`Age: 0.683
`
`Gender: 0.594
`
`Race: 0.969
`
`Region: 0.653
`
`0
`
`1
`Hazard ratio
`
`2
`
`N=
`
`HR (95.34% CI)
`
`Favours Aflibercept
`
`Favours Placebo
`P-value for interaction
`
`Prior hypertension: 0.131
`
`Number of organs
`with metastasis: 0.699
`
`Liver metastases only: 0.090
`
`Location of primary tumour: 0.890
`
`Subgroups
`
`stneitapllA
`
`No prior hypertension
`
`Prior hypertension
`
`6221
`
`0.817 (0.713 to 0.937)
`
`692
`
`534
`
`0.883 (0.74 to 1.054)
`
`0.714 (0.577 to 0.884)
`
`Number of organs with metastasis <=1
`
`535
`
`0.767 (0.618 to 0.953)
`
`Number of organs with metastasis >1
`
`691
`
`0.825 (0.692 to 0.982)
`
`No liver metastasis,
`or liver & other metastases
`
`Liver metastasis only
`
`Colon/Rectosigmoid/Other
`
`Rectum
`
`927
`
`299
`
`855
`
`371
`
`0.868 (0.742 to 1.015)
`
`0.649 (0.492 to 0.855)
`
`0.818 (0.695 to 0.963)
`
`0.806 (0.629 to 1.031)
`
`Cutoff date: 7 February 2011
`HR, hazard ratio
`
`Hazard ratio
`
`Fig. 2. Forest plots for overall survival in planned subgroup analyses by patient demographic (Panel A) and baseline characteristics (Panel B). p
`values 60.1 indicate a significant interaction between treatment and demographic subgroup factors at the 2-sided 10% level (Cox proportional
`hazard model). HR, hazard ratio.
`
`5
`
`

`

`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`325
`
`A. Demographic subgroups
`
`Subgroups
`
`N=
`
`HR (95% CI)
`
`Favours Aflibercept
`
`Favours Placebo
`
`All patients
`
`1226
`
`0.758 (0.661 to 0.869)
`
`P-value for interaction
`
`Age <65
`
`Age >=65
`
`Male
`
`Female
`
`783
`
`443
`
`718
`
`508
`
`0.765 (0.645 to 0.909)
`
`0.748 (0.598 to 0.936)
`
`0.769 (0.645 to 0.917)
`
`0.733 (0.593 to 0.908)
`
`Western Europe
`
`425
`
`0.778 (0.624 to 0.97)
`
`Eastern Europe
`
`297
`
`0.688 (0.51 to 0.929)
`
`North America
`
`138
`
`0.536 (0.344 to 0.837)
`
`South America
`
`118
`
`0.666 (0.424 to 1.045)
`
`Other countries
`
`248
`
`0.926 (0.695 to 1.235)
`
`B. Baseline characteristics
`
`Age: 0.929
`
`Gender: 0.712
`
`Region: 0.325
`
`0
`
`1
`Hazard ratio
`
`2
`
`Subgroups
`
`All patients
`
`No prior hypertension
`
`Prior hypertension
`
`Number of organs with
`metastasis <=1
`
`Number of organs with
`metastasis >1
`
`No liver metastasis,
`or liver and other metastases
`
`Liver metastasis only
`
`Colon/Rectosigmoid/Other
`
`Rectum
`
`Cutoff date: 7 February 2011
`HR, hazard ratio
`
`N=
`
`HR (95% CI)
`
`Favours Aflibercept
`
`Favours Placebo
`
`1226
`
`0.758 (0.661 to 0.869)
`
`P-value for interaction
`
`692
`
`534
`
`535
`
`691
`
`927
`
`299
`
`855
`
`371
`
`0.736 (0.616 to 0.88)
`
`0.761 (0.616 to 0.94)
`
`0.655 (0.527 to 0.815)
`
`0.822 (0.69 to 0.979)
`
`0.839 (0.719 to 0.98)
`
`0.547 (0.413 to 0.725)
`
`0.71 (0.603 to 0.836)
`
`0.881 (0.689 to 1.127)
`
`Prior hypertension: 0.788
`
`Number of organs
`with metastasis: 0.146
`
`Liver metastases
`only: 0.008
`
`Location of
`primary tumour: 0.153
`
`Hazard ratio
`
`Fig. 3. Forest plots for progression-free survival in planned subgroup analyses by patient demographic (Panel A) and baseline characteristics
`(Panel B). p values 60.1 indicate a significant interaction between treatment and baseline characteristic subgroup factors at the 2-sided 10% level
`(Cox proportional hazard model). HR, hazard ratio.
`
`6
`
`

`

`326
`
`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`Table 1
`Baseline characteristics in patients stratified by prior bevacizumab treatment.
`
`Parameter
`
`Prior bevacizumab
`
`Placebo/FOLFIRI
`(n = 187)
`
`No bevacizumab
`
`Aflibercept/FOLFIRI
`(n = 186)
`
`Placebo/FOLFIRI
`(n = 427)
`
`Aflibercept/FOLFIRI
`(n = 426)
`
`Eastern Cooperative Oncology Group performance status, n (%)
`0
`107 (57.2)
`1
`74 (39.6)
`2
`6 (3.2)
`Male, n (%)
`105 (56.1)
`Age, y, median (range)
`60.0 (27–86)
`Prior hypertension, n (%)
`97 (51.9)
`
`107 (57.5)
`74 (39.8)
`5 (2.7)
`110 (59.1)
`59.0 (32–81)
`92 (49.5)
`
`243 (56.9)
`176 (41.2)
`8 (1.9)
`248 (58.1)
`61.0 (19–84)
`171 (40.0)
`
`242 (56.8)
`176 (41.3)
`8 (1.9)
`255 (59.9)
`61.0 (21–82)
`174 (40.8)
`
`Region, n (%)
`Western Europe
`Eastern Europe
`North America
`South America
`Other countries
`
`Race, n (%)
`Caucasian/White
`Black
`Asian/Oriental
`Other
`
`Primary tumour location
`Colon/rectosigmoid/other
`Rectum
`
`70 (37.4)
`34 (18.2)
`53 (28.3)
`8 (4.3)
`22 (11.8)
`
`166 (88.8)
`7 (3.7)
`6 (3.2)
`8 (4.3)
`
`141 (75.4)
`46 (24.6)
`
`No. of metastatic organs involved at baseline, n (%)
`0
`1 (0.5)
`1
`86 (46.0)
`>1
`100 (53.5)
`
`Metastatic organs involved at baseline, n (%)
`Any site
`186 (99.5)
`Liver
`146 (78.1)
`Lung
`74 (39.6)
`Lymph nodes
`49 (26.2)
`Peritoneum
`23 (12.3)
`
`Liver metastasis, n (%)
`None, or liver and other metastases
`Liver metastasis only
`
`Duration of bevacizumab use, months,
`median (range)
`Antiangiogenic-free period, months,
`median (range)
`
`131 (70.1)
`56 (29.9)
`
`6 (0–28)
`
`2 (1–21)
`
`54 (29.0)
`47 (25.3)
`49 (26.3)
`8 (4.3)
`28 (15.1)
`
`172 (92.5)
`5 (2.7)
`4 (2.2)
`5 (2.7)
`
`140 (75.3)
`46 (24.7)
`
`1 (0.5)
`80 (43.0)
`105 (56.5)
`
`185 (99.5)
`146 (78.5)
`87 (46.8)
`46 (24.7)
`24 (12.9)
`
`134 (72.0)
`52 (28.0)
`
`6 (0–29)
`
`2 (1–33)
`
`147 (34.4)
`102 (23.9)
`22 (5.2)
`48 (11.2)
`108 (25.3)
`
`357 (83.6)
`20 (4.7)
`45 (10.5)
`5 (1.2)
`
`299 (70.0)
`128 (30.0)
`
`5 (1.2)
`185 (43.3)
`237 (55.5)
`
`422 (98.8)
`285 (66.7)
`203 (47.5)
`132 (30.9)
`65 (15.2)
`
`337 (78.9)
`90 (21.1)
`
`–
`
`–
`
`154 (36.2)
`114 (26.8)
`14 (3.3)
`54 (12.7)
`90 (21.1)
`
`376 (88.3)
`11 (2.6)
`31 (7.3)
`8 (1.9)
`
`275 (64.6)
`151 (35.4)
`
`1 (0.2)
`176 (41.3)
`249 (58.5)
`
`425 (99.8)
`313 (73.5)
`184 (43.2)
`127 (29.8)
`44 (10.3)
`
`325 (76.3)
`101 (23.7)
`
`–
`
`–
`
`Abbreviation: FOLFIRI, 5-fluorouracil, leucovorin and irinotecan.
`
`aflibercept arm and 3.9 months in the control arm. Med-
`ian PFS in patients without prior exposure to bev-
`acizumab was 6.9 months in the aflibercept arm and
`5.4 months in the control arm (Fig. 5).
`
`3.4. Adverse events in patient subgroups
`
`The frequency of TEAEs was generally consistent
`across patient subgroups. Table 3 summarises the most
`frequently reported grade 3 and grade 4 adverse events
`(including those leading to treatment discontinuation),
`as well as events typically associated with anti-VEGF
`therapy,
`in patients stratified by prior bevacizumab
`treatment. As previously reported, the use of aflibercept
`
`increased the chemotherapy-related toxicities associated
`with FOLFIRI [4]. There was no evidence of greater
`toxicity in patients previously treated with bevacizumab
`compared with those not previously treated with
`bevacizumab. Furthermore,
`the previous use of
`bevacizumab did not increase the rate of anti-VEGF-
`associated events.
`
`4. Discussion
`
`Patients with mCRC comprise a heterogeneous pop-
`ulation as represented by differences in variables such as
`age, co-morbidities, primary tumour site, extent and
`location of metastases, ECOG PS and prior treatment
`
`7
`
`

`

`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`327
`
`Table 2
`Median overall survival (months) and median progression-free survival (months) and hazard ratio estimates by stratification factors.
`
`Overall survival
`
`n
`
`Placebo/FOLFIRI median
`(95.34% CI)
`
`Aflibercept/FOLFIRI median
`(95.34% CI)
`
`Hazard ratio
`(95.34% CI)
`
`p Value for
`interaction
`
`All patients
`
`1226
`
`12.1 (11.07–13.11)
`
`13.5 (12.52–14.95)
`
`0.817 (0.713–0.937)
`
`Prior bevacizumab
`No
`Yes
`
`ECOG PS
`0
`1
`2
`
`Progression-free
`survival
`
`All patients
`Prior bevacizumab
`No
`Yes
`
`ECOG PS
`0
`1
`2
`
`853
`373
`
`699
`500
`27
`
`12.4 (11.17–13.54)
`11.7 (9.82–13.77)
`
`13.9 (12.71–15.64)
`12.5 (10.78–15.51)
`
`0.788 (0.669–0.927)
`0.862 (0.673–1.104)
`
`0.5668
`
`14.1 (12.88–16.62)
`10.1 (9.20–11.53)
`4.4 (1.97–10.02)
`
`16.9 (14.78–18.79)
`10.7 (9.36–12.35)
`2.8 (0.92–9.82)
`
`0.768 (0.635–0.928)
`0.869 (0.71–1.063)
`0.978 (0.43–2.221)
`
`0.7231
`
`(95% CI)
`
`(95% CI)
`
`(95% CI)
`
`1226
`
`4.7 (4.21–5.36)
`
`6.9 (6.51–7.20)
`
`0.758 (0.661–0.869)
`
`853
`373
`
`699
`500
`27
`
`5.4 (4.53–5.68)
`3.9 (3.02–4.30)
`
`5.4 (4.70–5.82)
`4.1 (3.06–4.47)
`2.0 (1.51–4.76)
`
`6.9 (6.37–7.20)
`6.7 (5.75–8.21)
`
`7.2 (6.87–8.31)
`5.6 (5.06–6.90)
`2.7 (0.92–12.88)
`
`0.797 (0.679–0.936)
`0.661 (0.512–0.852)
`
`0.1958
`
`0.761 (0.633–0.913)
`0.749 (0.607–0.923)
`0.618 (0.259–1.475)
`
`0.6954
`
`Abbreviations: CI, confidence interval; ECOG PS, Eastern Cooperative Oncology Group performance status; FOLFIRI, 5-fluorouracil, leucovorin
`and irinotecan.
`
`with bevacizumab. The present analysis has established
`that the benefit of adding aflibercept to FOLFIRI in
`patients with mCRC demonstrated in the overall (ITT)
`population in the VELOUR study [4] is consistently
`observed across a range of prespecified patient sub-
`groups based on demographic variables and baseline
`clinical characteristics including stratification factors at
`randomisation (prior bevacizumab treatment and
`ECOG PS).
`For OS and PFS, there was no significant interaction
`at the 10% level between treatment arms and demo-
`graphic factors, indicating a consistent benefit of afliber-
`cept across age, gender, race and geographical region
`subgroups (Figs. 2 and 3). Similarly, there was no signif-
`icant interaction between treatment arms and stratifica-
`tion by prior exposure to bevacizumab or by ECOG PS
`(Table 1; Figs. 4 and 5). However, no conclusions can be
`drawn from the HR in patients with an ECOG PS of 2,
`because of the small sample size for this stratum (27
`patients).
`Conducting subgroup analyses based on patients’
`baseline characteristics showed a significant interaction
`between treatment arms and factors only in patients
`grouped according to liver metastases (Figs. 2B and
`3B). There was a greater aflibercept treatment effect in
`patients with liver-only metastases at baseline than in
`patients with either no liver metastases or liver metasta-
`ses plus other organ involvement. The same effect has
`been seen in another trial in the second-line setting eval-
`uating the role of adding an anti-epidermal growth fac-
`tor
`receptor
`(EGFR)
`agent,
`panitumumab,
`to
`FOLFIRI, this suggesting that this patient population
`
`may derive more benefit from a more intensive treat-
`ment [7]. No significant heterogeneity in treatment effect
`was detected in patients classified according to prior
`hypertension, the number of metastatic organs involved
`and location of the primary tumour.
`For the end-points of OS and PFS, all prespecified
`subgroup analyses showed a consistent treatment effect
`favouring aflibercept over placebo (HR, <1.0).
`Since bevacizumab in combination with chemother-
`apy is approved for first-line treatment for mCRC, it
`is important to establish whether exposure to this agent
`compromises the efficacy of subsequent therapies that
`also target the angiogenic pathway, such as aflibercept.
`Although the present study was not powered to show
`differences between treatment arms for OS within spe-
`cific subgroups, the absence of a significant interaction
`between treatment arm and prior bevacizumab treat-
`ment (yes or no) showed that the efficacy of aflibercept
`was not diminished by prior bevacizumab treatment. It
`is unknown whether these results reflect the broad
`anti-angiogenic effect of aflibercept, which targets VEG-
`FB and PIGF in addition to VEGFA. Nevertheless, this
`is an encouraging observation and potentially broadens
`the options for those patients surviving first-line regi-
`mens that include bevacizumab and who are eligible
`for further treatment. Further evidence that patients
`treated first-line with bevacizumab can benefit from sub-
`sequent therapies that target VEGF is provided by the
`results from the TML study. In this study, patients
`who had progressed on bevacizumab in combination
`with chemotherapy as first-line treatment demonstrated
`a significant prolongation of OS and PFS when
`
`8
`
`

`

`328
`
`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`A. Prior bevacizumab
`
`Symbol=Censor
`Placebo/FOLFIRI
`Aflibercept/FOLFIRI
`
`Aflibercept/
`Placebo/
`FOLFIRI
`FOLFIRI
`11.7
`12.5
`9.82–13.77
`10.78–15.51
`0.862 [0.673–1.104]
`
`Median OS
`95% CI
`HR [95% CI]
`
`1.0
`
`0.9
`
`0.8
`
`0.7
`
`0.6
`
`0.5
`
`0.4
`
`0.3
`
`0.2
`
`0.1
`
`0.0
`
`Kaplan-Meier estimate
`
`0
`
`3
`
`6
`
`9
`
`12
`
`21
`18
`15
`Time (months)
`
`24
`
`27
`
`30
`
`33
`
`36
`
`Number at risk
`170
`115
`138
`187
`Placebo
`186
`Aflibercept
`178
`121
`150
`B. No prior bevacizumab
`
`81
`89
`
`54
`59
`
`37
`36
`
`22
`22
`
`13
`13
`
`Symbol=Censor
`Placebo/FOLFIRI
`Aflibercept/FOLFIRI
`
`Aflibercept/
`Placebo/
`FOLFIRI
`FOLFIRI
`12.4
`13.9
`11.17–13.54
`12.71–15.64
`0.788 [0.669–0.927]
`
`Median OS
`95% CI
`HR [95% CI]
`
`1.0
`
`0.9
`
`0.8
`
`0.7
`
`0.6
`
`0.5
`
`0.4
`
`0.3
`
`0.2
`
`0.1
`
`0.0
`
`Kaplan-Meier estimate
`
`0
`
`3
`
`6
`
`9
`
`12
`
`15
`
`24
`21
`18
`Time (months)
`
`27
`
`30
`
`33
`
`36
`
`39
`
`Number at risk
`Placebo
`427
`426
`Aflibercept
`
`403
`388
`
`347
`348
`
`286
`295
`
`205
`222
`
`139
`157
`
`94
`112
`
`65
`82
`
`38
`62
`
`Fig. 4. Kaplan–Meier curves for overall survival stratified by prior bevacizumab. HRs are placebo/FOLFIRI versus aflibercept/FOLFIRI.
`FOLFIRI, 5-fluorouracil, leucovorin and irinotecan; OS, overall survival; HR, hazard ratio.
`
`9
`
`

`

`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`329
`
`A. Prior bevacizumab
`
`Symbol=Censor
`Placebo/FOLFIRI
`Aflibercept/FOLFIRI
`
`Aflibercept/
`Placebo/
`FOLFIRI
`FOLFIRI
`3.9
`6.7
`2.86–5.42
`4.76–8.74
`0.661 [0.399–1.095]
`
`Median PFS
`95% CI
`HR [95% CI]
`
`1.0
`
`0.9
`
`0.8
`
`0.7
`
`0.6
`
`0.5
`
`0.4
`
`0.3
`
`0.2
`
`0.1
`
`0.0
`
`Kaplan-Meier estimate
`
`0
`
`3
`
`6
`
`9
`
`12
`15
`18
`Time (months)
`
`21
`
`24
`
`27
`
`30
`
`22
`
`63
`
`87
`
`19
`23
`
`Number at risk
`96
`33
`187
`Placebo
`186
`Aflibercept
`124
`66
`B. No prior bevacizumab
`
`Symbol=Censor
`Placebo/FOLFIRI
`Aflibercept/FOLFIRI
`
`Aflibercept/
`Placebo/
`FOLFIRI
`FOLFIRI
`5.4
`6.9
`4.17–6.7
`5.82–8.15
`0.797 [0.58–1.096]
`
`Median PFS
`95% CI
`HR [95% CI]
`
`1.0
`
`0.9
`
`0.8
`
`0.7
`
`0.6
`
`0.5
`
`0.4
`
`0.3
`
`0.2
`
`0.1
`
`0.0
`
`Kaplan-Meier estimate
`
`0
`
`3
`
`6
`
`9
`
`12
`15
`Time (months)
`
`18
`
`21
`
`24
`
`75
`
`259
`296
`
`138
`181
`
`75
`76
`
`38
`36
`
`18
`14
`
`Number at risk
`Placebo
`427
`426
`Aflibercept
`
`Fig. 5. Kaplan–Meier curves for progression-free survival stratified by prior bevacizumab. HRs are placebo/FOLFIRI versus aflibercept/
`FOLFIRI. FOLFIRI, 5-fluorouracil, leucovorin and irinotecan; OS, overall survival; HR, hazard ratio.
`
`10
`
`

`

`330
`
`J. Tabernero et al. / European Journal of Cancer 50 (2014) 320–331
`
`Table 3
`Adverse events in patients stratified by prior bevacizumab status according to e-CRFs (safety population).
`Adverse eventsa, %
`patients
`
`No bevacizumab
`
`Aflibercept/FOLFIRI
`(n = 171)
`
`Placebo/FOLFIRI
`(n = 433)
`
`Prior bevacizumab
`
`Placebo/FOLFIRI (n = 172)
`
`Aflibercept/FOLFIRI
`(n = 440)
`
`47
`2
`3
`
`Serious AEs
`Fatal AE b due to PD
`Fatal AE b in other
`context than PD
`
`32
`3
`2
`
`52
`2
`2
`
`33
`2
`1
`
`Grade 3
`
`Grade 4
`
`Grade 3
`
`Grade 4
`
`Grade 3
`
`Grade 4
`
`Grade 3
`
`Grade 4
`
`Grade 3/4 AEs in >10% of patients in any treatment group
`Neutropenia (PT)
`9.3
`4.1
`Diarrhoea (PT)
`8.1
`0.6
`Infections and infestations
`6.4
`1.2
`(SOC)
`Stomatitis (PT)
`Hypertension (PT)
`Fatigue (PT)
`
`3.5
`0.6
`8.7
`
`0
`0
`0
`
`Other anti-VEGF-associated events
`Proteinuria
`0.6
`Haemorrhage
`1.2
`GI disorders (SOC)
`0.6
`Headache (PT)
`0
`Venous thromboembolic
`2.9
`event
`Pulmonary embolism
`(PT)
`Arterial thromboembolic
`event
`GI perforation
`
`0.6
`
`0
`
`0
`
`0
`0
`0
`0
`2.9
`
`2.9
`
`0
`
`0
`
`AEs leading to permanent treatment discontinuation
`Any
`9.3
`Grade 3/4 in >1% of
`Grade 3/4
`patients
`Fatigue
`Diarrhoea
`Hypertension
`Pulmonary embolism
`Asthenia
`Dehydration
`Rectal haemorrhage
`Neutropenic infection
`
`0.6
`0.6
`0
`0
`0
`0
`0
`1.2
`
`13.5
`18.1
`12.9
`
`10.5
`16.4
`14.6
`
`9.4
`3.5
`3.5
`0.6
`4.7
`
`0
`
`0.6
`
`0
`
`6.4
`0.6
`0.6
`
`0
`0
`0
`
`0
`0
`0
`0
`2.3
`
`2.3
`
`1.2
`
`0
`
`17.1
`7.4
`6.0
`
`5.1
`1.8
`7.2
`
`1.4
`1.8
`1.2
`0.5
`2.5
`
`0
`
`0.5
`
`0.2
`
`8.3
`0
`0.7
`
`0
`0
`0.2
`
`0
`0
`0
`0
`3.9
`
`3.7
`
`0
`
`0.2
`
`18.0
`19.3
`10.2
`
`13.4
`20.0
`10.9
`
`7.3
`2.5
`1.1
`2.0
`2.5
`
`0
`
`0.9
`
`0.2
`
`9.1
`0.2
`1.6
`
`0.2
`0.2
`0.9
`
`0
`0.2
`0.2
`0
`5.7
`
`5.5
`
`0.9
`
`0.5
`
`25.7
`Gra