- Email: [email protected]
Clinical value of circulating endothelial cell levels in metastatic colorectal cancer patients treated with first-line chemotherapy and bevacizumab†
Annals of Oncology
original articles Annals of Oncology 23: 919–927, 2012 doi:10.1093/annonc/mdr365 Published online 8 August 2011
Clinical value of circulating endothelial cell levels in metastatic colorectal cancer patients treated with first-line chemotherapy and bevacizumab D. Malka1, V. Boige1,2, N. Jacques3, N. Vimond3, A. Adenis4, E. Boucher5, J. Y. Pierga6, T. Conroy7, B. Chauffert8, E. Franc xois9, P. Guichard10, M. P. Galais11, F. Cvitkovic12, M. Ducreux1 & F. Farace2,3*
Received 20 May 2011; revised 22 June 2011; accepted 24 June 2011
Background: We investigated whether circulating endothelial cells (CECs) predict clinical outcome of first-line chemotherapy and bevacizumab in metastatic colorectal cancer (mCRC) patients. Patients and methods: In a substudy of the randomized phase II FNCLCC ACCORD 13/0503 trial, CECs (CD452 CD31+ CD146+ 7-amino-actinomycin2 cells) were enumerated in 99 patients by four-color flow cytometry at baseline and after one cycle of treatment. We correlated CEC levels with objective response rate (ORR), 6-month progressionfree survival (PFS) rate (primary end point of the trial), PFS, and overall survival (OS). Multivariate analyses of potential prognostic factors, including CEC counts and Ko¨hne score, were carried out. Results: By multivariate analysis, high baseline CEC levels were the only independent prognostic factor for 6-month PFS rate (P < 0.01) and were independently associated with worse PFS (P = 0.02). High CEC levels after one cycle were the only independent prognostic factor for ORR (P = 0.03). High CEC levels at both time points independently predicted worse ORR (P = 0.025), 6-month PFS rate (P = 0.007), and PFS (P = 0.02). Ko¨hne score was the only variable associated with OS. Conclusion: CEC levels at baseline and after one treatment cycle may independently predict ORR and PFS in mCRC patients starting first-line bevacizumab and chemotherapy. Key words: bevacizumab, biomarker, chemotherapy, circulating endothelial cells, metastatic colorectal cancer, prognosis
introduction The addition of bevacizumab, a monoclonal antibody targeting vascular endothelial growth factor, the key mediator in angiogenesis, to first-line or second-line chemotherapy has been shown to improve objective response rate (ORR), progression-free survival (PFS), and/or overall survival (OS) compared with chemotherapy alone in metastatic colorectal cancer (mCRC) patients [1–3]. However, only a subset of patients will derive benefit from bevacizumab-based therapy. Furthermore, the assessment of antitumor efficacy of *Correspondence to: Dr F. Farace, Laboratory of Translational Research, Institut Gustave Roussy, 114 rue Edouard Vaillant, 94805 Villejuif Cedex, France. Tel: +33-142-11-51-98; Fax: +33-1-42-11-60-94; E-mail: [email protected]
Coinvestigators are listed in acknowledgements.
bevacizumab-based therapies in mCRC patients relies on RECIST [4] or perhaps more accurately on computed tomography morphological response criteria [5] that are usually obtained at least 2 months after the start of treatment. Therefore, identifying biomarkers that predict earlier the outcome of mCRC patients starting treatment with bevacizumab would allow to distinguish patients who will benefit most from treatment from those who will need to be rapidly reoriented toward another treatment approach. Additionally, such biomarkers should be markers for bevacizumab and chemotherapy altogether in combination, since bevacizumab alone is ineffective in mCRC patients [3]. Circulating endothelial cells (CECs) have emerged as a promising candidate biomarker to assess the efficacy of antiangiogenic therapies [6–8]. CECs are mature cells shed from vessel walls, and high CEC counts are observed in vascular
ª The Author 2011. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected]
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
1 Gastrointestinal Oncology Unit, Department of Medical Oncology, Institut Gustave Roussy, Villejuif; 2Institut National de la Sante et de la Recherche Medicale U981, Universite´ Paris-Sud, Villejuif; 3Laboratory of Translational Research, Institut Gustave Roussy, Villejuif; 4Department of Gastrointestinal Oncology, Center Oscar Lambret, Lille; 5Department of Medical Oncology, Center Euge`ne Marquis, Rennes; 6Department of Medical Oncology, Institut Curie, Paris; 7Department of Medical Oncology, Center Alexis Vautrin, Vandoeuvre-le`s-Nancy; 8Department of Medical Oncology, Center Georges-Francxois Leclerc, Dijon; 9Department of Medical Oncology, Center Antoine Lacassagne, Nice; 10Department of Medical Oncology, Polyclinique des 4 Pavillons, Lormont; 11Department of Medical Oncology, Center Francxois Baclesse, Caen; 12Department of Medical Oncology, Center Rene´ Huguenin, Saint-Cloud, France
original articles diseases hallmarked by endothelial insult or dysfunction including cancer [9–11]. Because of the technical difficulties and challenges associated with CEC measurement, data regarding the biological significance of CECs in cancer patients are still lacking [12, 13]. Furthermore, conflicting results concerning the actual prognostic or predictive value of CEC counts during antiangiogenic treatments have been reported [7, 13]. Further studies are needed to establish the clinical situation, i.e. which tumor type, treatment, or combined strategies, where CEC enumeration will have the most relevant significance. We recently established a reproducible and sensitive fourcolor flow cytometry assay for CEC enumeration [14]. Using this assay in a prospective substudy of FNCLCC ACCORD 13/ 0503, a randomized phase II trial assessing bevacizumab plus irinotecan and fluoropyrimidine (either FOLFIRI or XELIRI regimen) as first-line treatment of mCRC [15], we investigated the association of CEC levels measured before and shortly after the initiation of treatment with patient outcome.
patient population, study design, and treatment The randomized phase II trial FNCLCC ACCORD 13/0503 allocated patients aged 18–75 years with histologically proven, nonresectable, measurable mCRC, an Eastern Cooperative Oncology Group performance status (PS) of 0–2, and no prior palliative chemotherapy to receive either XELIRI [200 mg/m2 irinotecan on day (D)1 and 1000 mg/m2 b.i.d. capecitabine (800 mg/m2 b.i.d. in patients aged 65 years or older) on D1– 14] plus 7.5 mg/kg bevacizumab on D1, every 3 weeks for a maximum of 8 cycles, or FOLFIRI (180 mg/m2 irinotecan, 400 mg/m2 leucovorin, and 400 mg/m2 fluorouracil bolus followed by 2400 mg/m2 as a 46-h infusion) plus 5 mg/kg bevacizumab on D1, every 2 weeks for a maximum of 12 cycles [15]. In patients whose disease was controlled after 6 months of bevacizumab and chemotherapy, bevacizumab (7.5 mg/kg every 3 weeks) was continued as a single-agent maintenance therapy in both arms until disease progression. Randomization was stratified on center, PS (0–1 versus 2), age (
CD146 phycoerthyrin (PE) (clone P1H12, BD Pharmingen), and CD45 allophycocyenin (APC) (DakoCytomation, Glostrup, Denmark). Antibody batches were titrated rigorously. An immunoglobulin G (IgG) PE control was carried out in 0.5 ml of whole blood (CD45 APC/CD31 FITC/mouse IgG PE/7AAD) to measure background noise and to adjust the gates precisely. Flow-Check microbeads (Beckman Coulter, Sullerton, CA) were added to a tube containing isotypic controls to adjust forward side scatter voltage and identify cell events with a size >10 lm. Cells were analyzed on a FACSCalibur (BD Biosciences, San Jose, CA). To ensure statistical analysis, all the cells contained in the IgG PE control tube and in the CEC test tube were acquired, representing 2.5 · 106 events and 5 · 106 events, respectively. Data were analyzed using CELLQuest 3.2 software (BD Biosciences). The gating strategy used to identify CECs is illustrated in supplemental Figure S1 (available at Annals of Oncology online).
statistical analysis CEC detection was carried out whenever possible without any target statistical power. CEC levels at baseline and after one cycle of treatment and changes in CEC levels from baseline to the end of first treatment cycle (DCEC) are reported as medians with range and interquartile range (IQR). Response and survival rates are reported with 95% confidence intervals (CI). Nonparametric tests (Wilcoxon’s or Kruskal–Wallis tests as appropriate) were used to compare CEC levels at baseline, after one cycle of treatment, and DCEC according to baseline characteristics, objective tumor responses (complete or partial responses; 4), and 6-month PFS rate, the primary end point of the trial [15]; as all patients had at least 7 months of follow-up (or progressed within this period), the 6-month PFS rate is a ‘crude’ estimate. As CECs are consistently detectable in healthy subjects at levels that frequently overlap with those observed in patients with advanced cancers [14, 16], association between CEC levels at baseline, after one cycle of treatment, or at both time points, and patient outcome was assessed using the upper limit of the IQR (i.e. the 75th percentile) of CEC levels [17]. For the analysis taking into account CEC levels at both time points, no formal hypothesis was made for ‘ranking’ patients with high CEC levels at baseline but low CEC levels after one cycle of treatment or patients with low CEC levels at baseline but high CEC levels after one cycle of treatment. Patient outcome was compared between these four classes using test for trend. Univariate and multivariate analyses of potential prognostic variables were carried out using logistic model (ORR, 6-month PFS rate) or Cox
CEC enumeration The CEC substudy has been approved by local ethic Committees. A specific informed consent was obtained for all participating patients. Blood samples were drawn at baseline (D1, before start of treatment) and at the end of the first treatment cycle [i.e. D15 (FOLFIRI–bevacizumab arm) or D22 (XELIRI–bevacizumab arm)]. After discarding the first 2 ml following venupunction, whole blood was collected in Cellsave tubes (Immunicon, Huttingdon Valley, PA). Cellsave tubes contain EDTA and a cellpreservative agent that stabilizes fragile cells such as CECs and have been previously validated in our laboratory for CEC measurement by flow cytometry [14]. All blood samples were shipped at 18–25C to our laboratory and analyzed within 24 h. All CEC assays were blinded to patient baseline characteristics, allocated treatment, and outcome. CECs, identified as CD31+ CD146+ CD452 7-amino-actinomycin D (7AAD)2 viable cells, were measured in 1 ml whole blood by four-color flow cytometry according to a method we established previously [14]. Briefly, immunofluorescent staining was carried out with the following monoclonal antibodies: CD31 fluorescein isothiocyanate (FITC) (BD Pharmingen, San Diego, CA),
920 | Malka et al.
Figure 1. Study flowchart. CECs, circulating endothelial cells.
Volume 23 | No. 4 | April 2012
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
patients and methods
Annals of Oncology
original articles
Annals of Oncology
model (PFS, OS) adjusted on treatment arm. Variables included age, sex, CEC levels, and Ko¨hne prognostic score, a validated model in which mCRC patients are classified in low-risk, intermediate-risk, or high-risk groups according to four baseline parameters (PS, number of metastatic sites, white blood cell count, and alkaline phosphatase level) [18, 19]. Survival curves were calculated using the Kaplan–Meier method. PFS and OS according to CEC levels were compared using Cox model adjusted on treatment arm. Cutoff date for analysis was 15 March 2010.
Table 1. Baseline CEC levels according to patient characteristics N (%)
P
(56) 17 (44) 14
0–93 0–82
10–24 8–20
0.54a
(44) 17 (39) 15 (16) 16
0–93 0–61 2–60
8–30 8–19 12–34
0.44b
(57) 12 (43) 17
0–61 0–93
7–20 13–30
0.02a
(76) 15 (24) 17
0–61 0–93
9–22 7–43
0.54a
(49) 16 (51) 15
0–82 0–93
9–22 9–25
0.64a
(51) 17 (49) 15
0–82 0–93
8–23 10–29
0.60a
(48) 15 (43) 17 (9) 20
0–82 0–61 10–93
8–22 8–24 12–41
0.33b
a
Wilcoxon’s test. Kruskal–Wallis test. c Missing data for one patient for whom alkaline phosphatase level was not available. CEC, circulating endothelial cell; IQR, interquartile range; PS, performance status. b
results Among the 145 patients enrolled in the FNCLCC ACCORD 13/ 0503 trial in 15 centers from March 2006 to January 2008, 99 (68%) agreed to participate in the CEC study in 11 centers (Figure 1). These 99 patients were representative of the whole patient population with respect to sex, age, PS, location of the primary tumor, number of metastatic sites, prior adjuvant chemotherapy, and allocated treatment arm, except for Ko¨hne prognostic score (low risk, 48% versus 43%; intermediate risk, 42% versus 41%; high risk, 10% versus 16%; P = 0.02) (supplemental Table S1, available at Annals of Oncology online). Blood samples were available for CEC enumeration at baseline in 97 patients (98%), after one cycle of treatment in 91 patients (92%), and at both time points in 89 patients (90%) (Figure 1). Baseline CEC levels (median, 16/ml; IQR, 9–23) were significantly higher in patients with altered PS (PS 1–2, 17/ml versus PS 0, 12/ml; P = 0.02) but did not differ according to Ko¨hne prognostic score, sex, age, location of the primary tumor, number of metastatic sites, and treatment arm (Table 1). Baseline CEC levels did not significantly differ according to tumor objective response but were lower in patients with disease control at 6 months (15/ml versus 32/ml; P = 0.01) (supplemental Table S2, available at Annals of Oncology online). CEC levels after one cycle of treatment (median, 18/ml; IQR, 9–32) were lower in cases of objective response (13/ml versus 22/ml; P = 0.03) and disease control at 6 months (14/ml versus 26/ml; P = 0.03). No statistically significant association was seen between DCEC (median, 2/ml; IQR, 29 to 12) and response to treatment. When patients were dichotomized according to whether their CEC levels were below or above the 75th percentile of the IQR, patients with high baseline CEC levels had a worse 6-month PFS rate (P < 0.01), whereas patients with high CEC levels after one cycle of treatment had worse ORR (P = 0.03) and worse 6-month PFS rate (P = 0.04) (Table 2). Kaplan–Meier analyses showed that after a median follow-up of 36 months (range, 25–47), PFS was significantly shorter in patients with high
Table 2. Response to treatment in patients with high or low CEC levels at baseline and after one cycle of treatment (univariate analysis) CEC levels (/ml) Baseline (n = 97) £23 >23 After one cycle (n = 91) £32 >32
ORRa n/Nb (%)
6-Month PFS rate n/Nd (%) OR
OR
95% CI
Pc
50/71 (70) 12/23 (52)
2.2
0.8–5.7
0.11
65/72 (90) 15/25 (60)
49/67 (73) 10/21 (48)
3.0
1.1–8.2
0.03
58/68 (85) 15/23 (65)
95% CI
Pc
6.2
2.0–18.9
<0.01
3.1
1.0–9.2
0.04
a
Three patients were not assessable for tumor response (early treatment arrest, n = 2; missing data, n = 1). n, number of patients with objective response; N, number of patients assessed. c Wilcoxon’s test. d n, number of patients without disease progression; N, number of patients assessed. CEC, circulating endothelial cell; CI, confidence interval; OR, odds ratio; ORR, objective response rate; PFS, progression-free survival. b
Volume 23 | No. 4 | April 2012
doi:10.1093/annonc/mdr365 | 921
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
Sex Male 54 Female 43 Age (years) £59 43 60–69 38 ‡70 16 PS 0 55 1 or 2 42 Primary Colon 74 Rectum 23 Number of metastatic sites 1 48 >1 49 Treatment arm FOLFIRI–bevacizumab 49 XELIRI–bevacizumab 48 Ko¨hne prognostic scorec Low risk 46 Intermediate risk 41 High risk 9
Baseline CEC levels (/ml) Median Range IQR
The level of statistical significance was set at 0.05 using two-sided tests. All analyses were carried out using SAS version 9.1 (SAS Institute Inc., Cary, NC).
original articles baseline CEC levels (median, 7.6 versus 9.8 months; log-rank test; P = 0.02; Figure 2A). However, PFS did not significantly differ between patients with low and high CEC levels after one cycle of treatment (median, 8.4 versus 9.4 months; P = 0.75). Also, OS did not significantly differ according to CEC levels at baseline (median, 19.8 versus 26.1 months; P = 0.13; Figure 2B)
Annals of Oncology
or after one cycle of treatment (median, 22.0 versus 25.6 months; P = 0.84). In multivariate analysis, baseline CEC level was the only independent prognostic factor for 6-month PFS rate (P < 0.01) and was an independent prognostic factor for PFS (P = 0.02) with Ko¨hne prognostic score (Table 3). None of the tested baseline variables were significantly associated with
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
Figure 2. Kaplan–Meier survival curves according to circulating endothelial cell levels (above or below the 75th percentile of the interquartile range of values) at baseline. (A) progression-free survival. (B) Overall survival. Log-rank test, P = 0.13.
922 | Malka et al.
Volume 23 | No. 4 | April 2012
original articles
Annals of Oncology
Table 3. Multivariate analysis of prognostic factors in patients classified by CEC levels at baseline or after one cycle of treatment
95% CI
P
6-Month PFS rate OR 95% CI
P
PFS HR
95% CI
2.0
0.7–5.5
0.17
5.8
1.8–18.0
<0.01
1.8
1.1–2.9
0.5
0.1–2.6
0.82
0.5
0.1–2.7
0.50
0.4
0.4 1.0 2.3 3.3
0.1–1.8 1.0–1.1 0.9–5.9 1.1–9.7
0.18 0.33 0.09 0.03
0.5 1.0 1.2 3.0
0.1–2.9 0.9–1.1 0.4–4.0 1.0–9.5
0.63 0.91 0.74 0.06
0.4
0.1–2.2
0.52
0.3
0.1–1.4
0.4 1.0 2.7
0.1–1.8 1.0–1.1 1.0–7.5
0.30 0.38 0.05
0.2 1.0 1.4
0.1–1.2 1.0–1.1 0.5–4.3
OS HR
95% CI
P
0.02
1.5
0.9–2.6
0.12
0.2–0.9
0.02
0.4
0.2–1.0
0.05
0.4 1.0 1.0 1.1
0.2–0.9 1.0–1.0 0.7–1.6 0.7–1.9
0.02 0.29 0.90 0.61
0.4 1.0 1.0 1.1
0.2–0.9 1.0–1.1 0.6–1.6 0.6–2.0
0.02 0.06 0.92 0.76
0.38
0.3
0.2–0.7
<0.01
0.4
0.2–0.9
0.02
0.16 0.89 0.56
0.3 1.0 1.0
0.2–0.7 1.0–1.0 0.7–1.6
<0.01 0.40 0.91
0.4 1.0 1.0
0.2–0.8 1.0–1.1 0.6–1.6
0.02 0.18 0.92
P
a Three patients were not assessable for tumor response (early treatment arrest, n = 2; missing data, n = 1). CEC, circulating endothelial cell; CI, confidence interval; HR, hazards ratio; OR, odds ratio; ORR, objective response rate; OS, overall survival; PFS, progression-free survival.
ORR. CEC level after one cycle of treatment was the only independent prognostic factor for ORR (P = 0.03)—sex being of borderline significance—and tended to be associated with 6-month PFS rate (P = 0.06), but not with PFS. Ko¨hne prognostic score was the only variable associated with OS. Finally, we classified patients into four groups according to whether their CEC levels were below or above the 75th percentile of the IQR at both time points (baseline and after one cycle of treatment) (Table 4). Patients with high CEC levels at both time points had significantly worse ORR [36% versus 74%; OR, 5.0 (95% CI 1.3–19.7); P = 0.05] and 6-month PFS rate [42% versus 89%; OR, 11.2 (95% CI 2.7–46.7); P < 0.01] than patients with low CEC levels at both time points. Patients with only one high CEC level (either at baseline or after one cycle of treatment) harbored intermediate outcomes. Test for trend between the four classes was significant for ORR (P = 0.02) and 6-month PFS rate (P = 0.004). Kaplan–Meier analyses showed that PFS (log-rank test; P = 0.04; Figure 3A), but not OS (P = 0.68; Figure 3B), significantly differ between the four patient subgroups. In multivariate analysis, CEC classification was the only independent prognostic factor for ORR (test for trend; P = 0.025) and 6-month PFS rate (test for trend; P = 0.007) (Table 5). Patients with high CEC levels at both baseline and after one cycle of treatment had a worse PFS than patients with low CEC levels at both time points (P = 0.02); however, the test for trend between the four patient subgroups was not significant (P = 0.23). Ko¨hne prognostic score was the only variable associated with OS.
discussion In this prospective substudy of the randomized phase II FNCLCC ACCORD 13/0503 trial [15], we found an association of CEC levels with outcome of mCRC patients starting first-line
Volume 23 | No. 4 | April 2012
Table 4. Response to treatment in the four patient subgroups defined by CEC levels at both baseline and after one cycle of treatment (univariate analysis) Patient groups N
ORRa % OR 95% CI
b
CEC Low–low High–low Low–high High–high
6-Month PFS rate % OR 95% CI
P c
54 12 11 12
74 73 60 36
0.02 1 1.1 0.3–4.6 0.36 1.9 0.5–7.8 0.91 5.0 1.3–19.7 0.05
P
0.004c 89 1 75 2.7 0.6–12.7 0.75 91 0.8 0.1–7.4 0.22 42 11.2 2.7–46.7 <0.01
a
Three patients were not assessable for tumor response (early treatment arrest, n = 2; missing data, n = 1). b Where ‘low’ and ‘high’ denote CEC levels below and above, respectively, the 75th percentile of the interquartile range of the median value (baseline, 23/ml; after one cycle of treatment, 32/ml). The first term denotes CEC levels at baseline and the second one CEC levels after one cycle of treatment. c Test for trend. CEC, circulating endothelial cell. CI, confidence interval. OR, odds ratio. ORR, objective response rate. PFS, progression-free survival.
bevacizumab and chemotherapy. Baseline CEC levels were the only independent prognostic factor for 6-month PFS rate, the primary end point of the trial [15], and were independently associated with PFS. High CEC levels after one cycle of treatment were the only independent prognostic factor for ORR. Furthermore, patients with high CEC levels at both time points (baseline and after one cycle of treatment) had significantly worse ORR, 6-month PFS rate, and PFS than patients with low CEC levels at both time points, whereas patients with high CEC levels either at baseline or at the end of first cycle—who may correspond to distinct (and perhaps opposite) biological situations—harbored intermediate outcomes. In multivariate analysis, CEC classification was the
doi:10.1093/annonc/mdr365 | 923
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
CEC at baseline (>23 versus £23) Ko¨hne prognostic score Intermediate versus high risk Low versus high risk Age Sex (male versus female) CEC after one cycle (>32 versus £32) Ko¨hne prognostic score Intermediate versus high risk Low versus high risk Age Sex (male versus female)
ORRa OR
original articles
Annals of Oncology
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
Figure 3. Kaplan–Meier survival curves for the four patient subgroups defined by CEC levels at baseline (above or below the 75th percentile of the interquartile range of values) and at the end of first treatment cycle. (A) progression-free survival. Medians: ‘low–low’, 9.6 (9.1–10.6) months; ‘low–high’, 10.7 (9.3–26.0) months; ‘high–low’, 8.1 (7.8–10.9) months; and ‘high–high’, 7.0 (5.7–8.4) months. Log-rank test, P = 0.04. (B) Overall survival. Medians: low–low, 32 (21.9 to not evaluable) months; high–low, 23 (19.8 to not evaluable) months; low–high, 23 (14.0 to not evaluable) months; and high–high, 19 (9.7 to not evaluable) months. Log-rank test, P = 0.68.
924 | Malka et al.
Volume 23 | No. 4 | April 2012
original articles
Annals of Oncology
Table 5. Multivariate analysis of prognostic factors in patients classified by CEC levels at both time points ORRa OR CEC Low–low High–low Low–high High–high Ko¨hne prognostic score Intermediate versus high risk Low versus high risk Age Sex (male versus female)
95% CI
6-Month PFS rate OR 95% CI
P b
0.007
0.025 1 1.2 2.4 4.8
0.41 0.69 0.12
1 2.6 0.8 10.9
0.3–5.4 0.6–10.6 1.1–20.7
0.5–12.6 0.1–7.9 2.3–51.2
0.6
0.1–3.7
0.72
0.6
0.6 1.0 2.4
0.1–3.5 1.0–1.1 0.9–6.9
0.60 0.36 0.09
0.7 1.0 1.0
PFS HR
P
95% CI
b
OS HR
P
95% CI
b
0.26b
0.23
0.78 0.24 0.01
1 1.3 0.7 2.5
0.6–2.5 0.4–1.5 1.2–5.3
0.1–4.4
0.74
0.5
0.1–4.4 1.0–1.1 0.3–3.4
0.75 0.74 0.94
0.5 1.0 0.9
P
0.51 0.39 0.02
1 1.3 1.2 1.7
0.5–3.0 0.4–3.2 0.7–4.0
0.61 0.83 0.27
0.2–1.1
0.07
0.3
0.1–0.8
0.01
0.2–1.0 1.0–1.0 0.5–1.4
0.06 0.14 0.52
0.3 1.0 1.0
0.1–0.8 1.0–1.1 0.5–2.0
0.01 0.37 0.98
a
Three patients were not assessable for tumor response (early treatment arrest, n = 2; missing data, n = 1). Test for trend. CEC, circulating endothelial cell; CI, confidence interval; HR, hazards ratio; OR, odds ratio; ORR, objective response rate; OS, overall survival; PFS, progression-free survival.
b
Volume 23 | No. 4 | April 2012
one identified by Ronzoni et al. (CD34+ CD146+ CD452 CD1332) [17] and ours (CD31+ CD146+ CD452 7AAD2). In accordance with recommendations for the identification of extremely rare events, we developed a flow cytometry assay to detect CECs, which characteristics include [14, 30] (i) staining of a large volume of whole blood (1 ml), (ii) use of a viability marker (7AAD), (iii) use of a whole-blood IgPE control to measure accurately background noise, and (iv) accumulation of a large number of events (5 · 106) to ensure statistical analysis. Additionally, we previously demonstrated that cells identified as CECs were of endothelial origin by co-staining of CD31, CD146, and CD144 (vascular endothelial–cadherin) [14]. Finally, our study demonstrates that CEC enumeration by flow cytometry is feasible in a prospective, multicenter study setting despite the requirement for rapid shipping of blood samples—which must be analyzed within 24 h. In fact, CEC levels could be measured at both time points in ‡90% of patients in our study. Several results of our study deserve comments. First, baseline CEC levels were higher in patients with PS 1–2 compared with patients with PS 0. High levels of viable CECs at baseline in PS 1–2 patients—in median three-fold higher than in healthy individuals [14]—without evidence of vascular disease or damage (as mandatory for patient eligibility in the trial; 15), may reflect a high rate of remodeling and turnover of the tumor endothelium which has been previously associated with worse prognosis in various types of cancer, including mCRC [31–33]. PS is a strong prognostic factor for PFS and OS in mCRC [34]. Of note, no significant association was observed between baseline CEC levels and any of the other baseline characteristics, including the number of metastatic sites and Ko¨hne prognostic score—which includes PS and the number of metastatic sites. This suggests that baseline CEC levels are a marker of biological aggressiveness of mCRC rather than a mere reflect of disease burden. Secondly, Ko¨hne score was the only variable associated with OS in multivariate analysis, thus validating this prognostic model not only in mCRC patients treated with fluorouracilbased chemotherapy [18] but also in those receiving bevacizumab plus irinotecan-based chemotherapy, as suggested
doi:10.1093/annonc/mdr365 | 925
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
only independent prognostic factor for ORR and 6-month PFS rate, and patients in the ‘high–high’ CEC group had a significantly worse PFS than patients in the ‘low–low’ CEC group (P = 0.02), although the test for trend between the four patient subgroups was not significant (P = 0.23). These results suggest that CEC levels measured at baseline and as early as after one cycle of treatment might help to identify patients with the highest likelihood of tumor control or tumor shrinkage when using bevacizumab and chemotherapy. Conflicting results concerning the clinical value of CEC levels during antiangiogenic therapy have been reported, due in part to the different clinical situations explored, but also to the different techniques used for CEC detection [14, 16, 17, 20–26]. In fact, the clinical evaluation of CECs has been hampered by the extreme rarity of these cells and the lack of consensus on CEC surface marker phenotype and on the techniques used for CEC enumeration [12, 27–29]—most commonly flow cytometry and immunomagnetic separation followed by microscopic detection. Two studies have recently reported conflicting results on the clinical value of CECs in bevacizumab–chemotherapy-treated mCRC patients [17, 26]. Using flow cytometry, Ronzoni et al. [17] reported in a study of 40 mCRC patients treated with firstline bevacizumab and chemotherapy that baseline CEC levels above the 75th percentile were associated with a significantly shorter PFS. By contrast, using the semiautomated immunomagnetic enrichment-based CellSearch technique, Simkens et al. [26] found no correlation between CEC levels and patient outcome (PFS and OS) in 473 patients with advanced colorectal cancer treated in the phase III CAIRO-2 trial with firstline chemotherapy and targeted therapies including bevacizumab. Taken together, these results and ours suggest that the lack of correlation in the study by Simkens et al. [26] may be due to the technique used for CEC enumeration. Immunomagnetic separation-based techniques tend to select preferentially the cells with the highest levels of capture antigen (i.e. CD146), thus potentially resulting in a significant loss of information. Moreover, the CEC phenotype identified in the study by Simkens et al. (CD105+ CD146+ CD452 DAPI+) [26] differed from the
original articles
acknowledgements We are indebted to the patients and their families. We thank F. Billiot for technical assistance; M. Taylor for editorial assistance; J. Gene`ve, M. Torres-Macque, S. Leveˆque, F. NaitAtmane, A. C. Le Gall (FNCLCC, BECT, Paris); M. Abbas (data manager, Institut de Cance´rologie Gustave Roussy, Villejuif); and the Ligue Contre le Cancer. Co-investigators: Rahmane Ouabdesselam, Anne-Laure Villing, Pierre Zimmermann (Institut Gustave Roussy, Villejuif); Bruno Coudert (Center Georges-Francxois Leclerc, Dijon); Philippe Follana (Center Antoine Lacassagne, Nice); Barbara Dieumegard, Fawzia Mefti, Nicole Renody (Center Rene´ Huguenin, Saint-Cloud); Sophie Dominguez, Marie Vanhuyse (Center Oscar Lambret, Lille); Christine Kaminsky, Lionel Uwer (Center Alexis Vautrin, Vandoeuvre-Les-Nancy); Ve´ronique Girre, Ste´phanie Henry, Sophie Piperno-Neumann (Institut Curie, Paris); Jean-Marc Guilloit, Jacques-Henri Jacob, Jean-Michel Ollivier (Center Francxois Baclesse, Caen); Sylvain Manfredi, Jean-Luc Raoul, Florence Trivin (Center Euge`ne Marquis, Rennes), France.
funding Roche; Pfizer; Chugai.
926 | Malka et al.
disclosure Authors’ disclosure of potential conflicts of interest: Antoine Adenis, Chugai (honoraria); Michel Ducreux, Roche (consulting, honoraria, and research funding), Pfizer (consulting, honoraria, and research funding), and Chugai (research funding); Eric Francxois, Roche (research funding); David Malka, Roche (consulting, honoraria, and research funding), Pfizer (research funding), and Chugai (research funding); and Jean-Yves Pierga, Roche (consulting, honoraria, and research funding) and Pfizer (consulting). The other authors have declared no conflicts of interest.
references 1. Hurwitz H, Fehrenbacher L, Novotny W et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004; 350: 2335–2342. 2. Saltz LB, Clarke S, Dı´az-Rubio E et al. Bevacizumab in combination with oxaliplatin-based chemotherapy as first-line therapy in metastatic colorectal cancer: a randomized phase III study. J Clin Oncol 2008; 26: 2013–2019. 3. Giantonio BJ, Catalano PJ, Meropol NJ et al. Bevacizumab in combination with oxaliplatin, fluorouracil, and leucovorin (FOLFOX4) for previously treated metastatic colorectal cancer: results from the Eastern Cooperative Oncology Group Study E3200. J Clin Oncol 2007; 25: 1539–1544. 4. Therasse P, Arbuck SG, Eisenhauer EA et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 2000; 92: 205–216. 5. Chun YS, Vauthey JN, Boonsirikamchai P et al. Association of computed tomography morphologic criteria with pathologic response and survival in patients treated with bevacizumab for colorectal liver metastases. JAMA 2009; 302: 2338–2344. 6. Jubb AM, Oates AJ, Holden S et al. Predicting benefit from anti-angiogenic agents in malignancy. Nat Rev Cancer 2006; 6: 626–635. 7. Bertolini F, Shaked Y, Mancuso P et al. The multifaceted circulating endothelial cell in cancer: towards marker and target identification. Nat Rev Cancer 2006; 6: 835–845. 8. Bhatt RS, Seth P, Sukhatme VP. Biomarkers for monitoring antiangiogenic therapy. Clin Cancer Res 2007; 13: 777s–780s. 9. Dignat-George F, Sampol J, Lip G et al. Circulating endothelial cells: realities and promises in vascular disorders. Pathophysiol Haemost Thromb 2003–2004; 33: 495–499. 10. Dignat-George F, Sampol J. Circulating endothelial cells in vascular disorders: new insights into an old concept. Eur J Haematol 2000; 65: 215–220. 11. Blann AD, Woywodt A, Bertolini F et al. Circulating endothelial cells. Biomarker of vascular disease. Thromb Haemost 2005; 93: 228–235. 12. Dignat-George F, Sabatier F, Blann A et al. Detection of circulating endothelial cells: CD146-based magnetic separation enrichment or flow cytometric assay? J Clin Oncol 2007; 25: e1–e2. 13. Strijbos MH, Gratama JW, Kraan J et al. Circulating endothelial cells in oncology: pitfalls and promises. Br J Cancer 2008; 98(11): 1731–1735. 14. Jacques N, Vimond N, Conforti R et al. Quantification of circulating mature endothelial cells using a whole blood four-color flow cytometric assay. J Immunol Methods 2008; 337: 132–143. 15. Ducreux M, Adenis A, Mendiboure J et al. Efficacy and safety of bevacizumab (BEV)based combination regimens in patients with metastatic colorectal cancer (mCRC): Randomized phase II study of BEV + FOLFIRI versus BEV + XELIRI, (FNCLCC ACCORD 13/0503 study). J Clin Oncol. 2009; 27(Suppl): 15s (abstr 4086). 16. Rowand JL, Martin G, Doyle GV et al. Endothelial cells in peripheral blood of healthy subjects and patients with metastatic carcinomas. Cytometry A 2007; 71: 105–113. 17. Ronzoni M, Manzoni M, Mariucci S et al. Circulating endothelial cells and endothelial progenitors as predictive markers of clinical response to
Volume 23 | No. 4 | April 2012
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
in a recent retrospective study [19]. In fact, although median OS was of 19 months in the high–high CEC patient subgroup as compared with 32 months in the low–low group, the difference did not reach statistical significance. Besides an insufficient statistical power of our study, one may hypothesize that second-line and further therapies may have significantly impacted on OS. Larger studies are warranted to assess the prognostic value of CEC counts for OS, perhaps also by implementing repeated CEC measurements throughout patients’ course. Finally, one may hypothesize that baseline CEC levels, which correlated with 6-month PFS rate and PFS, but not with ORR, are mostly prognostic, whereas CEC levels after one cycle of treatment, which correlated with ORR, are mostly predictive. However, the design of our study did not allow to assess whether CEC levels were prognostic, predictive, or both. Furthermore, it did not allow assessing whether CEC levels correlated with the efficacy of bevacizumab, chemotherapy, or their combination—although biomarkers that correlate with the activity of bevacizumab and chemotherapy altogether in combination might be clinically meaningful, since bevacizumab alone is ineffective in mCRC patients [3]. Prospective studies in mCRC patients receiving chemotherapy with or without bevacizumab are ongoing. In conclusion, our study suggests that CEC counts before and early after the start of first-line bevacizumab-based therapy may help to distinguish patients who will benefit most from treatment from those who will need to be rapidly reoriented toward another treatment approach. It also suggests that the phenotypic characterization of CECs and the technique used for their enumeration may be of importance when measuring their role as a marker for prognosis and/or prediction of response in patients undergoing antiangiogenic treatments. These promising data warrant further validation studies.
Annals of Oncology
Annals of Oncology
18.
19.
20.
21.
22.
23.
25.
26. Simkens LH, Tol J, Terstappen LW et al. The predictive and prognostic value of circulating endothelial cells in advanced colorectal cancer patients receiving first-line chemotherapy and bevacizumab. Ann Oncol 2010; 21: 2447–2448. 27. Widemann A, Sabatier F, Arnaud L et al. CD146-based immunomagnetic enrichment followed by multiparameter flow cytometry: a new approach to counting circulating endothelial cells. J Thromb Haemost 2008; 6(5): 869–876. 28. Strijbos MH, Kraan J, den Bakker MA et al. Cells meeting our immunophenotypic criteria of endothelial cells are large platelets. Cytometry B Clin Cytom 2007; 72: 86–93. 29. Strijbos MH, Kraan J, Lamers CH et al. Quantification of circulating endothelial cells by flow cytometry. Clin Cancer Res 2009; 15: 3640. 30. Taylor M, Ro¨ssler J, Geoerger B et al. High levels of circulating VEGFR2+ bone marrow-derived progenitor cells correlate with metastatic disease in patients with pediatric solid malignancies. Clin Cancer Res 2009; 15: 4561. 31. Frank RE, Saclarides TJ, Leurgans S et al. Tumor angiogenesis as a predictor of recurrence and survival in patients with node-negative colon cancer. Ann Surg 1995; 222: 695–699. 32. Galizia G, Lieto E, Ferraraccio F et al. Determination of molecular marker expression can predict clinical outcome in colon carcinomas. Clin Cancer Res 2004; 10: 3490–3499. 33. Des Guetz G, Uzzan B, Nicolas P et al. Microvessel density and VEGF expression are prognostic factors in colorectal cancer. Meta-analysis of the literature. Br J Cancer 2006; 94: 1823–1832. 34. Sargent DJ, Ko¨hne CH, Sanoff HK et al. Pooled safety and efficacy analysis examining the effect of performance status on outcomes in nine first-line treatment trials using individual data from patients with metastatic colorectal cancer. J Clin Oncol 2009; 27: 1948–1955.
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
24.
bevacizumab-based first-line treatment in advanced colorectal cancer patients. Ann Oncol 2010; 21: 2382–2389. Ko¨hne CH, Cunningham D, Di CF et al. Clinical determinants of survival in patients with 5-fluorouracil-based treatment for metastatic colorectal cancer: results of a multivariate analysis of 3825 patients. Ann Oncol 2002; 13: 308–317. Kabbinavar F, Irl C, Zurlo A, Hurwitz H. Bevacizumab improves the overall and progression-free survival of patients with metastatic colorectal cancer treated with 5-fluorouracil-based regimens irrespective of baseline risk. Oncology 2008; 75: 215–223. Batchelor TT, Sorensen AG, di Tomaso E et al. AZD2171, a pan-VEGF receptor tyrosine kinase inhibitor, normalizes tumor vasculature and alleviates edema in glioblastoma patients. Cancer Cell 2007; 11: 83–95. Mancuso P, Burlini A, Pruneri G et al. Resting and activated endothelial cells are increased in the peripheral blood of cancer patients. Blood 2001; 97: 3658–3661. Willett CG, Boucher Y, di Tomaso E et al. Direct evidence that the VEGF-specific antibody bevacizumab has antivascular effects in human rectal cancer. Nat Med 2004; 10: 145–147. Calleri A, Bono A, Bagnardi V et al. Predictive potential of angiogenic growth factors and circulating endothelial cells in breast cancer patients receiving metronomic chemotherapy plus bevacizumab. Clin Cancer Res 2009; 15: 7652–7657. Matsusaka S, Suenaga M, Mishima Y et al. Circulating endothelial cells predict for response to bevacizumab-based chemotherapy in metastatic colorectal cancer. Cancer Chemother Pharmacol 2010; doi:10.1007/s00280-010-1543-2. Bidard FC, Mathiot C, Degeorges A et al. Clinical value of circulating endothelial cells and circulating tumor cells in metastatic breast cancer patients treated first line with bevacizumab and chemotherapy. Ann Oncol 2010; 21: 1765–1771.
original articles
original articles Annals of Oncology 23: 919–927, 2012 doi:10.1093/annonc/mdr365 Published online 8 August 2011
Clinical value of circulating endothelial cell levels in metastatic colorectal cancer patients treated with first-line chemotherapy and bevacizumab D. Malka1, V. Boige1,2, N. Jacques3, N. Vimond3, A. Adenis4, E. Boucher5, J. Y. Pierga6, T. Conroy7, B. Chauffert8, E. Franc xois9, P. Guichard10, M. P. Galais11, F. Cvitkovic12, M. Ducreux1 & F. Farace2,3*
Received 20 May 2011; revised 22 June 2011; accepted 24 June 2011
Background: We investigated whether circulating endothelial cells (CECs) predict clinical outcome of first-line chemotherapy and bevacizumab in metastatic colorectal cancer (mCRC) patients. Patients and methods: In a substudy of the randomized phase II FNCLCC ACCORD 13/0503 trial, CECs (CD452 CD31+ CD146+ 7-amino-actinomycin2 cells) were enumerated in 99 patients by four-color flow cytometry at baseline and after one cycle of treatment. We correlated CEC levels with objective response rate (ORR), 6-month progressionfree survival (PFS) rate (primary end point of the trial), PFS, and overall survival (OS). Multivariate analyses of potential prognostic factors, including CEC counts and Ko¨hne score, were carried out. Results: By multivariate analysis, high baseline CEC levels were the only independent prognostic factor for 6-month PFS rate (P < 0.01) and were independently associated with worse PFS (P = 0.02). High CEC levels after one cycle were the only independent prognostic factor for ORR (P = 0.03). High CEC levels at both time points independently predicted worse ORR (P = 0.025), 6-month PFS rate (P = 0.007), and PFS (P = 0.02). Ko¨hne score was the only variable associated with OS. Conclusion: CEC levels at baseline and after one treatment cycle may independently predict ORR and PFS in mCRC patients starting first-line bevacizumab and chemotherapy. Key words: bevacizumab, biomarker, chemotherapy, circulating endothelial cells, metastatic colorectal cancer, prognosis
introduction The addition of bevacizumab, a monoclonal antibody targeting vascular endothelial growth factor, the key mediator in angiogenesis, to first-line or second-line chemotherapy has been shown to improve objective response rate (ORR), progression-free survival (PFS), and/or overall survival (OS) compared with chemotherapy alone in metastatic colorectal cancer (mCRC) patients [1–3]. However, only a subset of patients will derive benefit from bevacizumab-based therapy. Furthermore, the assessment of antitumor efficacy of *Correspondence to: Dr F. Farace, Laboratory of Translational Research, Institut Gustave Roussy, 114 rue Edouard Vaillant, 94805 Villejuif Cedex, France. Tel: +33-142-11-51-98; Fax: +33-1-42-11-60-94; E-mail: [email protected]
Coinvestigators are listed in acknowledgements.
bevacizumab-based therapies in mCRC patients relies on RECIST [4] or perhaps more accurately on computed tomography morphological response criteria [5] that are usually obtained at least 2 months after the start of treatment. Therefore, identifying biomarkers that predict earlier the outcome of mCRC patients starting treatment with bevacizumab would allow to distinguish patients who will benefit most from treatment from those who will need to be rapidly reoriented toward another treatment approach. Additionally, such biomarkers should be markers for bevacizumab and chemotherapy altogether in combination, since bevacizumab alone is ineffective in mCRC patients [3]. Circulating endothelial cells (CECs) have emerged as a promising candidate biomarker to assess the efficacy of antiangiogenic therapies [6–8]. CECs are mature cells shed from vessel walls, and high CEC counts are observed in vascular
ª The Author 2011. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected]
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
1 Gastrointestinal Oncology Unit, Department of Medical Oncology, Institut Gustave Roussy, Villejuif; 2Institut National de la Sante et de la Recherche Medicale U981, Universite´ Paris-Sud, Villejuif; 3Laboratory of Translational Research, Institut Gustave Roussy, Villejuif; 4Department of Gastrointestinal Oncology, Center Oscar Lambret, Lille; 5Department of Medical Oncology, Center Euge`ne Marquis, Rennes; 6Department of Medical Oncology, Institut Curie, Paris; 7Department of Medical Oncology, Center Alexis Vautrin, Vandoeuvre-le`s-Nancy; 8Department of Medical Oncology, Center Georges-Francxois Leclerc, Dijon; 9Department of Medical Oncology, Center Antoine Lacassagne, Nice; 10Department of Medical Oncology, Polyclinique des 4 Pavillons, Lormont; 11Department of Medical Oncology, Center Francxois Baclesse, Caen; 12Department of Medical Oncology, Center Rene´ Huguenin, Saint-Cloud, France
original articles diseases hallmarked by endothelial insult or dysfunction including cancer [9–11]. Because of the technical difficulties and challenges associated with CEC measurement, data regarding the biological significance of CECs in cancer patients are still lacking [12, 13]. Furthermore, conflicting results concerning the actual prognostic or predictive value of CEC counts during antiangiogenic treatments have been reported [7, 13]. Further studies are needed to establish the clinical situation, i.e. which tumor type, treatment, or combined strategies, where CEC enumeration will have the most relevant significance. We recently established a reproducible and sensitive fourcolor flow cytometry assay for CEC enumeration [14]. Using this assay in a prospective substudy of FNCLCC ACCORD 13/ 0503, a randomized phase II trial assessing bevacizumab plus irinotecan and fluoropyrimidine (either FOLFIRI or XELIRI regimen) as first-line treatment of mCRC [15], we investigated the association of CEC levels measured before and shortly after the initiation of treatment with patient outcome.
patient population, study design, and treatment The randomized phase II trial FNCLCC ACCORD 13/0503 allocated patients aged 18–75 years with histologically proven, nonresectable, measurable mCRC, an Eastern Cooperative Oncology Group performance status (PS) of 0–2, and no prior palliative chemotherapy to receive either XELIRI [200 mg/m2 irinotecan on day (D)1 and 1000 mg/m2 b.i.d. capecitabine (800 mg/m2 b.i.d. in patients aged 65 years or older) on D1– 14] plus 7.5 mg/kg bevacizumab on D1, every 3 weeks for a maximum of 8 cycles, or FOLFIRI (180 mg/m2 irinotecan, 400 mg/m2 leucovorin, and 400 mg/m2 fluorouracil bolus followed by 2400 mg/m2 as a 46-h infusion) plus 5 mg/kg bevacizumab on D1, every 2 weeks for a maximum of 12 cycles [15]. In patients whose disease was controlled after 6 months of bevacizumab and chemotherapy, bevacizumab (7.5 mg/kg every 3 weeks) was continued as a single-agent maintenance therapy in both arms until disease progression. Randomization was stratified on center, PS (0–1 versus 2), age (
CD146 phycoerthyrin (PE) (clone P1H12, BD Pharmingen), and CD45 allophycocyenin (APC) (DakoCytomation, Glostrup, Denmark). Antibody batches were titrated rigorously. An immunoglobulin G (IgG) PE control was carried out in 0.5 ml of whole blood (CD45 APC/CD31 FITC/mouse IgG PE/7AAD) to measure background noise and to adjust the gates precisely. Flow-Check microbeads (Beckman Coulter, Sullerton, CA) were added to a tube containing isotypic controls to adjust forward side scatter voltage and identify cell events with a size >10 lm. Cells were analyzed on a FACSCalibur (BD Biosciences, San Jose, CA). To ensure statistical analysis, all the cells contained in the IgG PE control tube and in the CEC test tube were acquired, representing 2.5 · 106 events and 5 · 106 events, respectively. Data were analyzed using CELLQuest 3.2 software (BD Biosciences). The gating strategy used to identify CECs is illustrated in supplemental Figure S1 (available at Annals of Oncology online).
statistical analysis CEC detection was carried out whenever possible without any target statistical power. CEC levels at baseline and after one cycle of treatment and changes in CEC levels from baseline to the end of first treatment cycle (DCEC) are reported as medians with range and interquartile range (IQR). Response and survival rates are reported with 95% confidence intervals (CI). Nonparametric tests (Wilcoxon’s or Kruskal–Wallis tests as appropriate) were used to compare CEC levels at baseline, after one cycle of treatment, and DCEC according to baseline characteristics, objective tumor responses (complete or partial responses; 4), and 6-month PFS rate, the primary end point of the trial [15]; as all patients had at least 7 months of follow-up (or progressed within this period), the 6-month PFS rate is a ‘crude’ estimate. As CECs are consistently detectable in healthy subjects at levels that frequently overlap with those observed in patients with advanced cancers [14, 16], association between CEC levels at baseline, after one cycle of treatment, or at both time points, and patient outcome was assessed using the upper limit of the IQR (i.e. the 75th percentile) of CEC levels [17]. For the analysis taking into account CEC levels at both time points, no formal hypothesis was made for ‘ranking’ patients with high CEC levels at baseline but low CEC levels after one cycle of treatment or patients with low CEC levels at baseline but high CEC levels after one cycle of treatment. Patient outcome was compared between these four classes using test for trend. Univariate and multivariate analyses of potential prognostic variables were carried out using logistic model (ORR, 6-month PFS rate) or Cox
CEC enumeration The CEC substudy has been approved by local ethic Committees. A specific informed consent was obtained for all participating patients. Blood samples were drawn at baseline (D1, before start of treatment) and at the end of the first treatment cycle [i.e. D15 (FOLFIRI–bevacizumab arm) or D22 (XELIRI–bevacizumab arm)]. After discarding the first 2 ml following venupunction, whole blood was collected in Cellsave tubes (Immunicon, Huttingdon Valley, PA). Cellsave tubes contain EDTA and a cellpreservative agent that stabilizes fragile cells such as CECs and have been previously validated in our laboratory for CEC measurement by flow cytometry [14]. All blood samples were shipped at 18–25C to our laboratory and analyzed within 24 h. All CEC assays were blinded to patient baseline characteristics, allocated treatment, and outcome. CECs, identified as CD31+ CD146+ CD452 7-amino-actinomycin D (7AAD)2 viable cells, were measured in 1 ml whole blood by four-color flow cytometry according to a method we established previously [14]. Briefly, immunofluorescent staining was carried out with the following monoclonal antibodies: CD31 fluorescein isothiocyanate (FITC) (BD Pharmingen, San Diego, CA),
920 | Malka et al.
Figure 1. Study flowchart. CECs, circulating endothelial cells.
Volume 23 | No. 4 | April 2012
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
patients and methods
Annals of Oncology
original articles
Annals of Oncology
model (PFS, OS) adjusted on treatment arm. Variables included age, sex, CEC levels, and Ko¨hne prognostic score, a validated model in which mCRC patients are classified in low-risk, intermediate-risk, or high-risk groups according to four baseline parameters (PS, number of metastatic sites, white blood cell count, and alkaline phosphatase level) [18, 19]. Survival curves were calculated using the Kaplan–Meier method. PFS and OS according to CEC levels were compared using Cox model adjusted on treatment arm. Cutoff date for analysis was 15 March 2010.
Table 1. Baseline CEC levels according to patient characteristics N (%)
P
(56) 17 (44) 14
0–93 0–82
10–24 8–20
0.54a
(44) 17 (39) 15 (16) 16
0–93 0–61 2–60
8–30 8–19 12–34
0.44b
(57) 12 (43) 17
0–61 0–93
7–20 13–30
0.02a
(76) 15 (24) 17
0–61 0–93
9–22 7–43
0.54a
(49) 16 (51) 15
0–82 0–93
9–22 9–25
0.64a
(51) 17 (49) 15
0–82 0–93
8–23 10–29
0.60a
(48) 15 (43) 17 (9) 20
0–82 0–61 10–93
8–22 8–24 12–41
0.33b
a
Wilcoxon’s test. Kruskal–Wallis test. c Missing data for one patient for whom alkaline phosphatase level was not available. CEC, circulating endothelial cell; IQR, interquartile range; PS, performance status. b
results Among the 145 patients enrolled in the FNCLCC ACCORD 13/ 0503 trial in 15 centers from March 2006 to January 2008, 99 (68%) agreed to participate in the CEC study in 11 centers (Figure 1). These 99 patients were representative of the whole patient population with respect to sex, age, PS, location of the primary tumor, number of metastatic sites, prior adjuvant chemotherapy, and allocated treatment arm, except for Ko¨hne prognostic score (low risk, 48% versus 43%; intermediate risk, 42% versus 41%; high risk, 10% versus 16%; P = 0.02) (supplemental Table S1, available at Annals of Oncology online). Blood samples were available for CEC enumeration at baseline in 97 patients (98%), after one cycle of treatment in 91 patients (92%), and at both time points in 89 patients (90%) (Figure 1). Baseline CEC levels (median, 16/ml; IQR, 9–23) were significantly higher in patients with altered PS (PS 1–2, 17/ml versus PS 0, 12/ml; P = 0.02) but did not differ according to Ko¨hne prognostic score, sex, age, location of the primary tumor, number of metastatic sites, and treatment arm (Table 1). Baseline CEC levels did not significantly differ according to tumor objective response but were lower in patients with disease control at 6 months (15/ml versus 32/ml; P = 0.01) (supplemental Table S2, available at Annals of Oncology online). CEC levels after one cycle of treatment (median, 18/ml; IQR, 9–32) were lower in cases of objective response (13/ml versus 22/ml; P = 0.03) and disease control at 6 months (14/ml versus 26/ml; P = 0.03). No statistically significant association was seen between DCEC (median, 2/ml; IQR, 29 to 12) and response to treatment. When patients were dichotomized according to whether their CEC levels were below or above the 75th percentile of the IQR, patients with high baseline CEC levels had a worse 6-month PFS rate (P < 0.01), whereas patients with high CEC levels after one cycle of treatment had worse ORR (P = 0.03) and worse 6-month PFS rate (P = 0.04) (Table 2). Kaplan–Meier analyses showed that after a median follow-up of 36 months (range, 25–47), PFS was significantly shorter in patients with high
Table 2. Response to treatment in patients with high or low CEC levels at baseline and after one cycle of treatment (univariate analysis) CEC levels (/ml) Baseline (n = 97) £23 >23 After one cycle (n = 91) £32 >32
ORRa n/Nb (%)
6-Month PFS rate n/Nd (%) OR
OR
95% CI
Pc
50/71 (70) 12/23 (52)
2.2
0.8–5.7
0.11
65/72 (90) 15/25 (60)
49/67 (73) 10/21 (48)
3.0
1.1–8.2
0.03
58/68 (85) 15/23 (65)
95% CI
Pc
6.2
2.0–18.9
<0.01
3.1
1.0–9.2
0.04
a
Three patients were not assessable for tumor response (early treatment arrest, n = 2; missing data, n = 1). n, number of patients with objective response; N, number of patients assessed. c Wilcoxon’s test. d n, number of patients without disease progression; N, number of patients assessed. CEC, circulating endothelial cell; CI, confidence interval; OR, odds ratio; ORR, objective response rate; PFS, progression-free survival. b
Volume 23 | No. 4 | April 2012
doi:10.1093/annonc/mdr365 | 921
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
Sex Male 54 Female 43 Age (years) £59 43 60–69 38 ‡70 16 PS 0 55 1 or 2 42 Primary Colon 74 Rectum 23 Number of metastatic sites 1 48 >1 49 Treatment arm FOLFIRI–bevacizumab 49 XELIRI–bevacizumab 48 Ko¨hne prognostic scorec Low risk 46 Intermediate risk 41 High risk 9
Baseline CEC levels (/ml) Median Range IQR
The level of statistical significance was set at 0.05 using two-sided tests. All analyses were carried out using SAS version 9.1 (SAS Institute Inc., Cary, NC).
original articles baseline CEC levels (median, 7.6 versus 9.8 months; log-rank test; P = 0.02; Figure 2A). However, PFS did not significantly differ between patients with low and high CEC levels after one cycle of treatment (median, 8.4 versus 9.4 months; P = 0.75). Also, OS did not significantly differ according to CEC levels at baseline (median, 19.8 versus 26.1 months; P = 0.13; Figure 2B)
Annals of Oncology
or after one cycle of treatment (median, 22.0 versus 25.6 months; P = 0.84). In multivariate analysis, baseline CEC level was the only independent prognostic factor for 6-month PFS rate (P < 0.01) and was an independent prognostic factor for PFS (P = 0.02) with Ko¨hne prognostic score (Table 3). None of the tested baseline variables were significantly associated with
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
Figure 2. Kaplan–Meier survival curves according to circulating endothelial cell levels (above or below the 75th percentile of the interquartile range of values) at baseline. (A) progression-free survival. (B) Overall survival. Log-rank test, P = 0.13.
922 | Malka et al.
Volume 23 | No. 4 | April 2012
original articles
Annals of Oncology
Table 3. Multivariate analysis of prognostic factors in patients classified by CEC levels at baseline or after one cycle of treatment
95% CI
P
6-Month PFS rate OR 95% CI
P
PFS HR
95% CI
2.0
0.7–5.5
0.17
5.8
1.8–18.0
<0.01
1.8
1.1–2.9
0.5
0.1–2.6
0.82
0.5
0.1–2.7
0.50
0.4
0.4 1.0 2.3 3.3
0.1–1.8 1.0–1.1 0.9–5.9 1.1–9.7
0.18 0.33 0.09 0.03
0.5 1.0 1.2 3.0
0.1–2.9 0.9–1.1 0.4–4.0 1.0–9.5
0.63 0.91 0.74 0.06
0.4
0.1–2.2
0.52
0.3
0.1–1.4
0.4 1.0 2.7
0.1–1.8 1.0–1.1 1.0–7.5
0.30 0.38 0.05
0.2 1.0 1.4
0.1–1.2 1.0–1.1 0.5–4.3
OS HR
95% CI
P
0.02
1.5
0.9–2.6
0.12
0.2–0.9
0.02
0.4
0.2–1.0
0.05
0.4 1.0 1.0 1.1
0.2–0.9 1.0–1.0 0.7–1.6 0.7–1.9
0.02 0.29 0.90 0.61
0.4 1.0 1.0 1.1
0.2–0.9 1.0–1.1 0.6–1.6 0.6–2.0
0.02 0.06 0.92 0.76
0.38
0.3
0.2–0.7
<0.01
0.4
0.2–0.9
0.02
0.16 0.89 0.56
0.3 1.0 1.0
0.2–0.7 1.0–1.0 0.7–1.6
<0.01 0.40 0.91
0.4 1.0 1.0
0.2–0.8 1.0–1.1 0.6–1.6
0.02 0.18 0.92
P
a Three patients were not assessable for tumor response (early treatment arrest, n = 2; missing data, n = 1). CEC, circulating endothelial cell; CI, confidence interval; HR, hazards ratio; OR, odds ratio; ORR, objective response rate; OS, overall survival; PFS, progression-free survival.
ORR. CEC level after one cycle of treatment was the only independent prognostic factor for ORR (P = 0.03)—sex being of borderline significance—and tended to be associated with 6-month PFS rate (P = 0.06), but not with PFS. Ko¨hne prognostic score was the only variable associated with OS. Finally, we classified patients into four groups according to whether their CEC levels were below or above the 75th percentile of the IQR at both time points (baseline and after one cycle of treatment) (Table 4). Patients with high CEC levels at both time points had significantly worse ORR [36% versus 74%; OR, 5.0 (95% CI 1.3–19.7); P = 0.05] and 6-month PFS rate [42% versus 89%; OR, 11.2 (95% CI 2.7–46.7); P < 0.01] than patients with low CEC levels at both time points. Patients with only one high CEC level (either at baseline or after one cycle of treatment) harbored intermediate outcomes. Test for trend between the four classes was significant for ORR (P = 0.02) and 6-month PFS rate (P = 0.004). Kaplan–Meier analyses showed that PFS (log-rank test; P = 0.04; Figure 3A), but not OS (P = 0.68; Figure 3B), significantly differ between the four patient subgroups. In multivariate analysis, CEC classification was the only independent prognostic factor for ORR (test for trend; P = 0.025) and 6-month PFS rate (test for trend; P = 0.007) (Table 5). Patients with high CEC levels at both baseline and after one cycle of treatment had a worse PFS than patients with low CEC levels at both time points (P = 0.02); however, the test for trend between the four patient subgroups was not significant (P = 0.23). Ko¨hne prognostic score was the only variable associated with OS.
discussion In this prospective substudy of the randomized phase II FNCLCC ACCORD 13/0503 trial [15], we found an association of CEC levels with outcome of mCRC patients starting first-line
Volume 23 | No. 4 | April 2012
Table 4. Response to treatment in the four patient subgroups defined by CEC levels at both baseline and after one cycle of treatment (univariate analysis) Patient groups N
ORRa % OR 95% CI
b
CEC Low–low High–low Low–high High–high
6-Month PFS rate % OR 95% CI
P c
54 12 11 12
74 73 60 36
0.02 1 1.1 0.3–4.6 0.36 1.9 0.5–7.8 0.91 5.0 1.3–19.7 0.05
P
0.004c 89 1 75 2.7 0.6–12.7 0.75 91 0.8 0.1–7.4 0.22 42 11.2 2.7–46.7 <0.01
a
Three patients were not assessable for tumor response (early treatment arrest, n = 2; missing data, n = 1). b Where ‘low’ and ‘high’ denote CEC levels below and above, respectively, the 75th percentile of the interquartile range of the median value (baseline, 23/ml; after one cycle of treatment, 32/ml). The first term denotes CEC levels at baseline and the second one CEC levels after one cycle of treatment. c Test for trend. CEC, circulating endothelial cell. CI, confidence interval. OR, odds ratio. ORR, objective response rate. PFS, progression-free survival.
bevacizumab and chemotherapy. Baseline CEC levels were the only independent prognostic factor for 6-month PFS rate, the primary end point of the trial [15], and were independently associated with PFS. High CEC levels after one cycle of treatment were the only independent prognostic factor for ORR. Furthermore, patients with high CEC levels at both time points (baseline and after one cycle of treatment) had significantly worse ORR, 6-month PFS rate, and PFS than patients with low CEC levels at both time points, whereas patients with high CEC levels either at baseline or at the end of first cycle—who may correspond to distinct (and perhaps opposite) biological situations—harbored intermediate outcomes. In multivariate analysis, CEC classification was the
doi:10.1093/annonc/mdr365 | 923
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
CEC at baseline (>23 versus £23) Ko¨hne prognostic score Intermediate versus high risk Low versus high risk Age Sex (male versus female) CEC after one cycle (>32 versus £32) Ko¨hne prognostic score Intermediate versus high risk Low versus high risk Age Sex (male versus female)
ORRa OR
original articles
Annals of Oncology
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
Figure 3. Kaplan–Meier survival curves for the four patient subgroups defined by CEC levels at baseline (above or below the 75th percentile of the interquartile range of values) and at the end of first treatment cycle. (A) progression-free survival. Medians: ‘low–low’, 9.6 (9.1–10.6) months; ‘low–high’, 10.7 (9.3–26.0) months; ‘high–low’, 8.1 (7.8–10.9) months; and ‘high–high’, 7.0 (5.7–8.4) months. Log-rank test, P = 0.04. (B) Overall survival. Medians: low–low, 32 (21.9 to not evaluable) months; high–low, 23 (19.8 to not evaluable) months; low–high, 23 (14.0 to not evaluable) months; and high–high, 19 (9.7 to not evaluable) months. Log-rank test, P = 0.68.
924 | Malka et al.
Volume 23 | No. 4 | April 2012
original articles
Annals of Oncology
Table 5. Multivariate analysis of prognostic factors in patients classified by CEC levels at both time points ORRa OR CEC Low–low High–low Low–high High–high Ko¨hne prognostic score Intermediate versus high risk Low versus high risk Age Sex (male versus female)
95% CI
6-Month PFS rate OR 95% CI
P b
0.007
0.025 1 1.2 2.4 4.8
0.41 0.69 0.12
1 2.6 0.8 10.9
0.3–5.4 0.6–10.6 1.1–20.7
0.5–12.6 0.1–7.9 2.3–51.2
0.6
0.1–3.7
0.72
0.6
0.6 1.0 2.4
0.1–3.5 1.0–1.1 0.9–6.9
0.60 0.36 0.09
0.7 1.0 1.0
PFS HR
P
95% CI
b
OS HR
P
95% CI
b
0.26b
0.23
0.78 0.24 0.01
1 1.3 0.7 2.5
0.6–2.5 0.4–1.5 1.2–5.3
0.1–4.4
0.74
0.5
0.1–4.4 1.0–1.1 0.3–3.4
0.75 0.74 0.94
0.5 1.0 0.9
P
0.51 0.39 0.02
1 1.3 1.2 1.7
0.5–3.0 0.4–3.2 0.7–4.0
0.61 0.83 0.27
0.2–1.1
0.07
0.3
0.1–0.8
0.01
0.2–1.0 1.0–1.0 0.5–1.4
0.06 0.14 0.52
0.3 1.0 1.0
0.1–0.8 1.0–1.1 0.5–2.0
0.01 0.37 0.98
a
Three patients were not assessable for tumor response (early treatment arrest, n = 2; missing data, n = 1). Test for trend. CEC, circulating endothelial cell; CI, confidence interval; HR, hazards ratio; OR, odds ratio; ORR, objective response rate; OS, overall survival; PFS, progression-free survival.
b
Volume 23 | No. 4 | April 2012
one identified by Ronzoni et al. (CD34+ CD146+ CD452 CD1332) [17] and ours (CD31+ CD146+ CD452 7AAD2). In accordance with recommendations for the identification of extremely rare events, we developed a flow cytometry assay to detect CECs, which characteristics include [14, 30] (i) staining of a large volume of whole blood (1 ml), (ii) use of a viability marker (7AAD), (iii) use of a whole-blood IgPE control to measure accurately background noise, and (iv) accumulation of a large number of events (5 · 106) to ensure statistical analysis. Additionally, we previously demonstrated that cells identified as CECs were of endothelial origin by co-staining of CD31, CD146, and CD144 (vascular endothelial–cadherin) [14]. Finally, our study demonstrates that CEC enumeration by flow cytometry is feasible in a prospective, multicenter study setting despite the requirement for rapid shipping of blood samples—which must be analyzed within 24 h. In fact, CEC levels could be measured at both time points in ‡90% of patients in our study. Several results of our study deserve comments. First, baseline CEC levels were higher in patients with PS 1–2 compared with patients with PS 0. High levels of viable CECs at baseline in PS 1–2 patients—in median three-fold higher than in healthy individuals [14]—without evidence of vascular disease or damage (as mandatory for patient eligibility in the trial; 15), may reflect a high rate of remodeling and turnover of the tumor endothelium which has been previously associated with worse prognosis in various types of cancer, including mCRC [31–33]. PS is a strong prognostic factor for PFS and OS in mCRC [34]. Of note, no significant association was observed between baseline CEC levels and any of the other baseline characteristics, including the number of metastatic sites and Ko¨hne prognostic score—which includes PS and the number of metastatic sites. This suggests that baseline CEC levels are a marker of biological aggressiveness of mCRC rather than a mere reflect of disease burden. Secondly, Ko¨hne score was the only variable associated with OS in multivariate analysis, thus validating this prognostic model not only in mCRC patients treated with fluorouracilbased chemotherapy [18] but also in those receiving bevacizumab plus irinotecan-based chemotherapy, as suggested
doi:10.1093/annonc/mdr365 | 925
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
only independent prognostic factor for ORR and 6-month PFS rate, and patients in the ‘high–high’ CEC group had a significantly worse PFS than patients in the ‘low–low’ CEC group (P = 0.02), although the test for trend between the four patient subgroups was not significant (P = 0.23). These results suggest that CEC levels measured at baseline and as early as after one cycle of treatment might help to identify patients with the highest likelihood of tumor control or tumor shrinkage when using bevacizumab and chemotherapy. Conflicting results concerning the clinical value of CEC levels during antiangiogenic therapy have been reported, due in part to the different clinical situations explored, but also to the different techniques used for CEC detection [14, 16, 17, 20–26]. In fact, the clinical evaluation of CECs has been hampered by the extreme rarity of these cells and the lack of consensus on CEC surface marker phenotype and on the techniques used for CEC enumeration [12, 27–29]—most commonly flow cytometry and immunomagnetic separation followed by microscopic detection. Two studies have recently reported conflicting results on the clinical value of CECs in bevacizumab–chemotherapy-treated mCRC patients [17, 26]. Using flow cytometry, Ronzoni et al. [17] reported in a study of 40 mCRC patients treated with firstline bevacizumab and chemotherapy that baseline CEC levels above the 75th percentile were associated with a significantly shorter PFS. By contrast, using the semiautomated immunomagnetic enrichment-based CellSearch technique, Simkens et al. [26] found no correlation between CEC levels and patient outcome (PFS and OS) in 473 patients with advanced colorectal cancer treated in the phase III CAIRO-2 trial with firstline chemotherapy and targeted therapies including bevacizumab. Taken together, these results and ours suggest that the lack of correlation in the study by Simkens et al. [26] may be due to the technique used for CEC enumeration. Immunomagnetic separation-based techniques tend to select preferentially the cells with the highest levels of capture antigen (i.e. CD146), thus potentially resulting in a significant loss of information. Moreover, the CEC phenotype identified in the study by Simkens et al. (CD105+ CD146+ CD452 DAPI+) [26] differed from the
original articles
acknowledgements We are indebted to the patients and their families. We thank F. Billiot for technical assistance; M. Taylor for editorial assistance; J. Gene`ve, M. Torres-Macque, S. Leveˆque, F. NaitAtmane, A. C. Le Gall (FNCLCC, BECT, Paris); M. Abbas (data manager, Institut de Cance´rologie Gustave Roussy, Villejuif); and the Ligue Contre le Cancer. Co-investigators: Rahmane Ouabdesselam, Anne-Laure Villing, Pierre Zimmermann (Institut Gustave Roussy, Villejuif); Bruno Coudert (Center Georges-Francxois Leclerc, Dijon); Philippe Follana (Center Antoine Lacassagne, Nice); Barbara Dieumegard, Fawzia Mefti, Nicole Renody (Center Rene´ Huguenin, Saint-Cloud); Sophie Dominguez, Marie Vanhuyse (Center Oscar Lambret, Lille); Christine Kaminsky, Lionel Uwer (Center Alexis Vautrin, Vandoeuvre-Les-Nancy); Ve´ronique Girre, Ste´phanie Henry, Sophie Piperno-Neumann (Institut Curie, Paris); Jean-Marc Guilloit, Jacques-Henri Jacob, Jean-Michel Ollivier (Center Francxois Baclesse, Caen); Sylvain Manfredi, Jean-Luc Raoul, Florence Trivin (Center Euge`ne Marquis, Rennes), France.
funding Roche; Pfizer; Chugai.
926 | Malka et al.
disclosure Authors’ disclosure of potential conflicts of interest: Antoine Adenis, Chugai (honoraria); Michel Ducreux, Roche (consulting, honoraria, and research funding), Pfizer (consulting, honoraria, and research funding), and Chugai (research funding); Eric Francxois, Roche (research funding); David Malka, Roche (consulting, honoraria, and research funding), Pfizer (research funding), and Chugai (research funding); and Jean-Yves Pierga, Roche (consulting, honoraria, and research funding) and Pfizer (consulting). The other authors have declared no conflicts of interest.
references 1. Hurwitz H, Fehrenbacher L, Novotny W et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004; 350: 2335–2342. 2. Saltz LB, Clarke S, Dı´az-Rubio E et al. Bevacizumab in combination with oxaliplatin-based chemotherapy as first-line therapy in metastatic colorectal cancer: a randomized phase III study. J Clin Oncol 2008; 26: 2013–2019. 3. Giantonio BJ, Catalano PJ, Meropol NJ et al. Bevacizumab in combination with oxaliplatin, fluorouracil, and leucovorin (FOLFOX4) for previously treated metastatic colorectal cancer: results from the Eastern Cooperative Oncology Group Study E3200. J Clin Oncol 2007; 25: 1539–1544. 4. Therasse P, Arbuck SG, Eisenhauer EA et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 2000; 92: 205–216. 5. Chun YS, Vauthey JN, Boonsirikamchai P et al. Association of computed tomography morphologic criteria with pathologic response and survival in patients treated with bevacizumab for colorectal liver metastases. JAMA 2009; 302: 2338–2344. 6. Jubb AM, Oates AJ, Holden S et al. Predicting benefit from anti-angiogenic agents in malignancy. Nat Rev Cancer 2006; 6: 626–635. 7. Bertolini F, Shaked Y, Mancuso P et al. The multifaceted circulating endothelial cell in cancer: towards marker and target identification. Nat Rev Cancer 2006; 6: 835–845. 8. Bhatt RS, Seth P, Sukhatme VP. Biomarkers for monitoring antiangiogenic therapy. Clin Cancer Res 2007; 13: 777s–780s. 9. Dignat-George F, Sampol J, Lip G et al. Circulating endothelial cells: realities and promises in vascular disorders. Pathophysiol Haemost Thromb 2003–2004; 33: 495–499. 10. Dignat-George F, Sampol J. Circulating endothelial cells in vascular disorders: new insights into an old concept. Eur J Haematol 2000; 65: 215–220. 11. Blann AD, Woywodt A, Bertolini F et al. Circulating endothelial cells. Biomarker of vascular disease. Thromb Haemost 2005; 93: 228–235. 12. Dignat-George F, Sabatier F, Blann A et al. Detection of circulating endothelial cells: CD146-based magnetic separation enrichment or flow cytometric assay? J Clin Oncol 2007; 25: e1–e2. 13. Strijbos MH, Gratama JW, Kraan J et al. Circulating endothelial cells in oncology: pitfalls and promises. Br J Cancer 2008; 98(11): 1731–1735. 14. Jacques N, Vimond N, Conforti R et al. Quantification of circulating mature endothelial cells using a whole blood four-color flow cytometric assay. J Immunol Methods 2008; 337: 132–143. 15. Ducreux M, Adenis A, Mendiboure J et al. Efficacy and safety of bevacizumab (BEV)based combination regimens in patients with metastatic colorectal cancer (mCRC): Randomized phase II study of BEV + FOLFIRI versus BEV + XELIRI, (FNCLCC ACCORD 13/0503 study). J Clin Oncol. 2009; 27(Suppl): 15s (abstr 4086). 16. Rowand JL, Martin G, Doyle GV et al. Endothelial cells in peripheral blood of healthy subjects and patients with metastatic carcinomas. Cytometry A 2007; 71: 105–113. 17. Ronzoni M, Manzoni M, Mariucci S et al. Circulating endothelial cells and endothelial progenitors as predictive markers of clinical response to
Volume 23 | No. 4 | April 2012
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
in a recent retrospective study [19]. In fact, although median OS was of 19 months in the high–high CEC patient subgroup as compared with 32 months in the low–low group, the difference did not reach statistical significance. Besides an insufficient statistical power of our study, one may hypothesize that second-line and further therapies may have significantly impacted on OS. Larger studies are warranted to assess the prognostic value of CEC counts for OS, perhaps also by implementing repeated CEC measurements throughout patients’ course. Finally, one may hypothesize that baseline CEC levels, which correlated with 6-month PFS rate and PFS, but not with ORR, are mostly prognostic, whereas CEC levels after one cycle of treatment, which correlated with ORR, are mostly predictive. However, the design of our study did not allow to assess whether CEC levels were prognostic, predictive, or both. Furthermore, it did not allow assessing whether CEC levels correlated with the efficacy of bevacizumab, chemotherapy, or their combination—although biomarkers that correlate with the activity of bevacizumab and chemotherapy altogether in combination might be clinically meaningful, since bevacizumab alone is ineffective in mCRC patients [3]. Prospective studies in mCRC patients receiving chemotherapy with or without bevacizumab are ongoing. In conclusion, our study suggests that CEC counts before and early after the start of first-line bevacizumab-based therapy may help to distinguish patients who will benefit most from treatment from those who will need to be rapidly reoriented toward another treatment approach. It also suggests that the phenotypic characterization of CECs and the technique used for their enumeration may be of importance when measuring their role as a marker for prognosis and/or prediction of response in patients undergoing antiangiogenic treatments. These promising data warrant further validation studies.
Annals of Oncology
Annals of Oncology
18.
19.
20.
21.
22.
23.
25.
26. Simkens LH, Tol J, Terstappen LW et al. The predictive and prognostic value of circulating endothelial cells in advanced colorectal cancer patients receiving first-line chemotherapy and bevacizumab. Ann Oncol 2010; 21: 2447–2448. 27. Widemann A, Sabatier F, Arnaud L et al. CD146-based immunomagnetic enrichment followed by multiparameter flow cytometry: a new approach to counting circulating endothelial cells. J Thromb Haemost 2008; 6(5): 869–876. 28. Strijbos MH, Kraan J, den Bakker MA et al. Cells meeting our immunophenotypic criteria of endothelial cells are large platelets. Cytometry B Clin Cytom 2007; 72: 86–93. 29. Strijbos MH, Kraan J, Lamers CH et al. Quantification of circulating endothelial cells by flow cytometry. Clin Cancer Res 2009; 15: 3640. 30. Taylor M, Ro¨ssler J, Geoerger B et al. High levels of circulating VEGFR2+ bone marrow-derived progenitor cells correlate with metastatic disease in patients with pediatric solid malignancies. Clin Cancer Res 2009; 15: 4561. 31. Frank RE, Saclarides TJ, Leurgans S et al. Tumor angiogenesis as a predictor of recurrence and survival in patients with node-negative colon cancer. Ann Surg 1995; 222: 695–699. 32. Galizia G, Lieto E, Ferraraccio F et al. Determination of molecular marker expression can predict clinical outcome in colon carcinomas. Clin Cancer Res 2004; 10: 3490–3499. 33. Des Guetz G, Uzzan B, Nicolas P et al. Microvessel density and VEGF expression are prognostic factors in colorectal cancer. Meta-analysis of the literature. Br J Cancer 2006; 94: 1823–1832. 34. Sargent DJ, Ko¨hne CH, Sanoff HK et al. Pooled safety and efficacy analysis examining the effect of performance status on outcomes in nine first-line treatment trials using individual data from patients with metastatic colorectal cancer. J Clin Oncol 2009; 27: 1948–1955.
Downloaded from http://annonc.oxfordjournals.org/ by guest on May 9, 2015
24.
bevacizumab-based first-line treatment in advanced colorectal cancer patients. Ann Oncol 2010; 21: 2382–2389. Ko¨hne CH, Cunningham D, Di CF et al. Clinical determinants of survival in patients with 5-fluorouracil-based treatment for metastatic colorectal cancer: results of a multivariate analysis of 3825 patients. Ann Oncol 2002; 13: 308–317. Kabbinavar F, Irl C, Zurlo A, Hurwitz H. Bevacizumab improves the overall and progression-free survival of patients with metastatic colorectal cancer treated with 5-fluorouracil-based regimens irrespective of baseline risk. Oncology 2008; 75: 215–223. Batchelor TT, Sorensen AG, di Tomaso E et al. AZD2171, a pan-VEGF receptor tyrosine kinase inhibitor, normalizes tumor vasculature and alleviates edema in glioblastoma patients. Cancer Cell 2007; 11: 83–95. Mancuso P, Burlini A, Pruneri G et al. Resting and activated endothelial cells are increased in the peripheral blood of cancer patients. Blood 2001; 97: 3658–3661. Willett CG, Boucher Y, di Tomaso E et al. Direct evidence that the VEGF-specific antibody bevacizumab has antivascular effects in human rectal cancer. Nat Med 2004; 10: 145–147. Calleri A, Bono A, Bagnardi V et al. Predictive potential of angiogenic growth factors and circulating endothelial cells in breast cancer patients receiving metronomic chemotherapy plus bevacizumab. Clin Cancer Res 2009; 15: 7652–7657. Matsusaka S, Suenaga M, Mishima Y et al. Circulating endothelial cells predict for response to bevacizumab-based chemotherapy in metastatic colorectal cancer. Cancer Chemother Pharmacol 2010; doi:10.1007/s00280-010-1543-2. Bidard FC, Mathiot C, Degeorges A et al. Clinical value of circulating endothelial cells and circulating tumor cells in metastatic breast cancer patients treated first line with bevacizumab and chemotherapy. Ann Oncol 2010; 21: 1765–1771.
original articles