Characteristics of BRAF^V600E Mutant, Deficient Mismatch Repair/Proficient Mismatch Repair, Metastatic Colorectal Cancer: A Multicenter Series of 287 Patients

This article describes the real‐life management of BRAF^V600E‐mutant metastatic colorectal cancer in a large cohort of patients treated in 16 French centers from 2006 to 2017. Potential prognostic factors to steer treatment decisions are identified.

Abstract

Background.

BRAF^V600E mutations occurring in about 10% of metastatic colorectal cancers (mCRCs) are usually associated with a poor outcome. However, their prognostic factors are unknown.

Materials and Methods.

We built a multicenter clinico‐biological database gathering data from patients with BRAF^V600E‐mutant mCRC treated in one of the 16 French centers from 2006 to 2017. The primary endpoint was to identify prognostic factors using a Cox model.

Results.

We included 287 patients (median age, 67 years [28–95]; female, 57%). Their median overall survival was 20.8 months (95% confidence interval [CI], 17.97–27.04), and median progression‐free survival in the first‐line setting was 4.34 months (95% CI, 3.81–5.03). Chemotherapy regimen and biological agents (antiangiogenic or anti‐epidermal growth factor receptor) were not associated with overall and progression‐free survival. Stage IV disease (synchronous metastases) and absence of curative‐intent surgery were statistically associated with poor overall survival. Among the 194 patients with mismatch repair (MMR) status available, overall survival was significantly longer in patients with deficient MMR tumors compared with those with proficient MMR tumors (adjusted hazard ratio = 0.56; p = .009).

Conclusion.

Despite that BRAF^V600E‐mutant mCRCs are associated with poor overall and progression‐free‐survival, patients with deficient MMR tumors and/or resectable disease experienced a longer survival. These results highlight the importance of MMR testing and resectability discussion in patients with BRAF^V600E mCRC in day‐to‐day practice.

Implications for Practice.

Mismatch repair (MMR) testing and resectability discussion in patients with BRAF^V600E metastatic colorectal cancer (mCRC) should be performed in day‐to‐day practice to steer treatment decision making in patients with BRAF^V600E‐mutant mCRC.

Introduction

The BRAF protein is a member of the serine/threonine kinase family that transduces signals downstream of RAS via a cascade of phosphorylation from MAPK/ERK kinase 1/2 (MEK1/2) to extracellular signal‐regulated kinase 1/2 (ERK1/2). Incidence of BRAF mutations, identified in human cancers in 2002, varies according to tumor types. The highest BRAF mutation rates are observed in melanoma and papillary thyroid cancers and are generally involved the V600 codon [1]. BRAF mutations are currently classified into three categories, according to their kinase activity, RAS dependency, and dimerization status underlying the differential tumor response to epidermal growth factor receptor (EGFR), BRAF, and MEK inhibitors [2], [3]. Indeed, BRAF inhibitor monotherapies are not effective in BRAF^V600E‐mutant metastatic colorectal cancer (mCRC) [4], [5], [6].

Colorectal cancer (CRC) is the third most common cancer type with more than 1,800,000 new cases and about 883,000 deaths worldwide in 2018 according to GLOBOCAN estimates [7]. RAS mutations are identified in about half of mCRC. They are used as predictive biomarker for resistance to anti‐EGFR therapies [8]. In addition, BRAF^V600E mutations observed in about 10% of mCRC [9] are often associated with MLH1 promoter hypermethylation and CpG island methylator phenotype inducing a microsatellite instability (MSI) phenotype [10], [11]. A rate of mismatch repair (MMR) deficiency (dMMR) from 20% to 40% was reported in BRAF^V600E‐mutant mCRC [12]. In metastatic setting, BRAF^V600E mutations have a significant negative prognostic impact leading to a reduced median overall survival (OS) of about 12 months [13], [14], [15]. On the other hand, its predictive role in anti‐EGFR resistance remains controversial [16], [17], [18]. Indeed, the management of patients with BRAF^V600E‐mutant mCRC is heterogeneous, based on the administration of tri‐ or bi‐chemotherapy regimen combined or not with bevacizumab, according to patient's age and performance status [19], [20]. Despite that recent studies demonstrated the benefit of immune checkpoint inhibitors in patients with dMMR tumors, no other biomarker allows to guide therapeutic choices [21], [22].

Our study describes the real‐life management of patients with BRAF^V600E‐mutant mCRC in a large retrospective multicenter French cohort. We aimed to identify potential prognostic factors to steer treatment decision making in patients with BRAF^V600E‐mutant mCRC.

Materials and Methods

Study Design

We retrospectively enrolled data from consecutive patients with BRAF^V600E‐mutant mCRC identified from January 2006 to September 2017 in 16 centers. Eligible patients were older than 18 years with BRAF^V600E‐mutation identified by polymerase chain reaction (PCR) or next‐generation sequencing analysis on tumor sample. Patients with mCRC who had experienced a curative resection of their metastases were also included. Noninclusion criteria were patients without metastases or for whom treatment and follow‐up information were not available.

Patient and Tumor Characteristics

Initial CRC stage was categorized according to the 7th Tumor Node Metastasis (TNM) classification system [23]. Somatic mutations in RAS genes and MMR status were collected, when available. MMR status was determined by MSI testing (pentaplex PCR) and/or analysis of immunohistochemistry [24]. dMMR status was defined as the presence of an instability for more than 20% of the microsatellites or a loss of MLH1, MSH2, PMS2, and/or MSH6 expression. Routine follow‐up consisted of physical examination, biological tests, and computed tomography scan every 2–3 months to evaluate treatment response and toxicity.

Statistical Analysis

Descriptive statistics were used to summarize patients’ characteristics. Median OS was defined as the time between the date of metastase(s) diagnosis and date of death (from any cause) or censored at the date of last follow‐up (September 1, 2017). Progression‐free survival (PFS) of the first‐line treatment (PFS1) was defined as the time between the initiation date of first‐line treatment and date of first disease progression, or the initiation date of second‐line treatment, or death, or censored at the date of last follow‐up. Patients with curative surgery for metastase(s) and primary tumor were also censored at the surgery date. Patients with curative surgery first and then “adjuvant” chemotherapy were excluded for PFS1 analysis. PFS of chemotherapy lines 2, 3, and 4 were calculated with the same definition. Survival curves for OS and PFS with associated log‐rank tests were generated using the Kaplan‐Meier method. Median follow‐up was calculated using reverse Kaplan‐Meier estimation.

A Cox proportional hazards model was used to investigate prespecified factors for OS and PFS. The variables considered were age at metastases diagnosis, gender, stage IV disease (synchronous or metachronous metastase(s)), primary tumor site, metastatic site (liver, lung, bone, and brain), number of metastatic sites, surgery of primary tumor and/or of metastase(s), and type of palliative treatment. For OS and PFS1, the sufficiently informed variables (less than 10% of missing data) and significant at a 0.20 level were included in a backward selection procedure to keep factors significant at 5% level in the final multivariate Cox model. SAS version 9.4 was used for all statistical analyses (SAS Institute Inc., Cary, NC).

Results

Clinical Characteristics

A total of 287 patients from 16 French centers (8 comprehensive cancer care centers, 7 university hospitals, and 1 private hospital) were included (Table 1). The median age was 67 years, and 57.1% of patients were women. Two (0.69%) patients had a known germinal mutation in one MMR gene. Nearly two thirds of patients (65.9 %) had synchronous metastases, and the primary tumor localization was mainly ascendant colon (65.4%). The most frequent metastatic sites were liver (51.9%), followed by peritoneum (37.3%), lymph nodes (31.0%), and lung (25.8%). More than half of patients (55.4%) had one metastatic site, including 82 patients (28.6%) with liver‐only metastases.

Table 1. Demographic and clinical characteristics.

Global population of BRAF^V600E‐mutant patients. Data are median (minimum–maximum) or n (%).

^aAccording to International Union Against Cancer.

Abbreviations: dMMR, deficient mismatch repair; M, metastasis; MMR, mismatch repair; N, node; pMMR, proficient mismatch repair; T, tumor.

Pathological and Molecular Characteristics

Most patients had advanced T and N stage at diagnosis (T3–4: 74.6%; N1–2: 57.9%). A total of 98 (34.1%) patients had a specific pathological component, including 79 (27.5%) with mucinous component in >50% of the tumor. Of note, other rare histological CRC subtypes such as signet ring cells carcinoma (n = 7) or sarcomatoid carcinoma (n = 2) were identified. The MMR status was determined for 194 patients and identified 85 (43.8%) patients with dMMR and 109 patients with proficient MMR (pMMR), among whom were 42 with MLH1 promoter methylation. Among the patients with KRAS status determined (n = 279), only two tumors harbored a RAS mutation (n = 2).

Treatments

Most patients (75.6%) had experienced a resection of primary tumor. Among the 96 patients with a stage II–III tumor at diagnosis, 57.3% received adjuvant chemotherapy (CT), mainly oxaliplatin‐based regimen (65.5%). Forty‐four patients (15.3%) had a curative metastasectomy, including liver metastases (n = 41) and complete cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (n = 5). Of the 44 patients, 19 experienced upfront metastasectomy without neoadjuvant or perioperative chemotherapy. Most patients (n = 256, 89.2%) had a first‐line CT (supplemental online Table 1). The 31 patients (10.8%) who did not begin CT were either too old (80.5 [49–95]) or had poor performance status. The median duration of the first‐line CT was 3.2 months (0.1–38.9). Only 25 patients (9.7%) received tri‐CT (FOLFOXIRI or FOLFIRINOX), in combination with either antiangiogenic or anti‐EGFR therapy for 16 of them. Doublet with oxaliplatin‐based CT was the most used protocol (49.2%), followed by irinotecan‐based CT (28.9%). Cytotoxic drugs were combined with bevacizumab in 42.6% or anti‐EGFR therapies in 17.6% of the patients. The main reasons for first‐line CT discontinuation were disease progression (51.5%), treatment holiday (24.3%), and toxicity (11.1%).

Of the 256 patients who received a first‐line CT, 158 (61.7%) initiated a second‐line CT for a median duration of 2.3 (0.03–47.0) months. In the second‐line setting, oxaliplatin‐based bi‐CT and irinotecan‐based CT were used in 18.6% and 44.2% of the patients, respectively. Of note, in the 109 patients treated with an antiangiogenic drug in the first‐line setting, 29 (26.6%) continued the antiangiogenic drug in the second‐line setting while switching chemotherapy regimen. Tumor progression was the main reason for second‐line CT discontinuation (70.8%). A total of 81 (28.2%) patients subsequently initiated a third line and 34 (11.8%) a fourth line of chemotherapy. Moreover, 7 patients received regorafenib, 8 patients received an anti‐BRAF, and 10 patients have been treated with an immune checkpoint inhibitors in clinical trials.

OS and PFS

With a median follow‐up of 47.5 months (3.3–210.7), the median OS was 20.8 months (95% CI, 18.0–27.0; Fig. 1). Overall, 68.9% of the patients were alive at 12 months (95% CI, 63.1–74) and 47.7% at 24 months (95% CI, 41.5–53.6). In the multivariate analysis, metachronous metastases (hazard ratio [HR] = 0.697; 95% CI, 0.499–0.973; p = .0341), metastasectomy (HR = 0.469; 95% CI, 0.288–0.765; p = .0024), and primary tumor surgery (HR = 0.306; 95% CI, 0.214–0.438; p < .0001) were statistically associated with better OS (Table 2). Regimen of first‐line treatment had no impact on OS (oxaliplatin‐ vs. +irinotecan‐based CT, antiangiogenic vs. anti‐EGFR therapies). The median OS for BRAF‐mutant patients who had a surgery for their metastases was 47.4 months (28.5–nonevaluable) versus 19.5 months (15.9–22.0) for those who had no metastasectomy (p < .001; Fig. 2). The median OS for patients with no surgery (neither primary tumor nor metastases) and at least one first‐line chemotherapy was 9.3 (95% CI, 6.9–11.1) months.

Figure 1.

OS.

Abbreviations: CI, confidence interval; OS, overall survival.

Table 2. Cox model for overall survival.

^aVariable eliminated during the backward selection procedure.

^bOrder of elimination during the backward selection procedure.

Abbreviations: CI, confidence interval; CT, chemotherapy; EGFR, epidermal growth factor receptor; FP, fluoropyrimidine; HR, hazard ratio; L1, first chemotherapy line; M, metastasis; NS, not significant.

Figure 2.

OS and metastasectomy.

Abbreviations: IQR, interquartile range; NE, nonevaluable; OS, overall survival

The median PFS1 was 4.3 months (95% CI, 3.8–5.0; Fig. 3). The PFS1 rates were 22.8% (17.8–28.2) and 12.6% (8.7–17.2) at 12 and 24 months, respectively. The chemotherapy regimen and antiangiogenic/anti‐EGFR therapy did not influence the PFS1 (supplemental online Table 2). In multivariate analysis, the number of metastatic sites (HR = 1.328; 95% CI, 1.012–1.742; p = .04), primary tumor surgery (HR = 0.627; 95% CI, 0.460–0.855; p = .0031), and the metastasectomy (HR = 0.586; 95% CI, 0.395–0.868; p = .0077) were significantly associated with PFS1. In the second‐line setting, the median progression‐free survival was 3.0 months (95% CI, 2.6–3.9). The PFS rates were 12.4% (95% CI, 7.5–18.5) and 2.3% (95% CI, 0.6–6.1) at 12 and 24 months.

Figure 3.

PFS1.

Abbreviations: CI, confidence interval; PFS, progression‐free survival; PFS1, progression‐free survival of the first‐line treatment.

Focus on dMMR Tumors

The Cox model did not include the MMR status because too many statuses were not specified (n = 93, 32%). However, we performed an exploratory analysis in the population of 194 patients with MMR status available. Of note, this subgroup was slightly different from the entire population (Table 1). Indeed, patients were younger (65.5 vs. 67.0 years), and fewer had an advanced disease at diagnostic (62.9% vs. 65.9%). Mucinous carcinoma tended to be more frequent (82.2% vs. 80.6%), and the ratio liver/peritoneum metastasis was inferior (1.4 vs. 1.0). The OS of the BRAF^V600E‐mutant dMMR population was longer than that of patients with pMMR tumors (30.5 [95% CI, 24.6–42.2] vs. 16.5 [95% CI, 12.5–22.5] months; p = .0087; Fig. 4). The impact of MMR status on OS was still significant even after adjustment on primitive tumor surgery, use of antiangiogenic drug in first‐line metastatic setting, or use of immune checkpoint inhibitors (HR = 0.557; 95% CI, 0.358–0.865; p = .0091). Of note, metastasectomy was the last variable to be eliminated in the multivariate model. The PFS of the first‐line chemotherapy was 4.0 months (95% CI, 3.1–6.4) in dMMR vs. 4.2 months (95% CI, 3.3–5.1) in pMMR tumors.

Figure 4.

OS according to MMR status in BRAF^V600E‐mutant patients.

Abbreviations: CI, confidence interval; dMMR, deficient mismatch repair; MMR, mismatch repair; pMMR, proficient mismatch repair; OS, overall survival.

Discussion

To our knowledge, the present study is the largest real‐life cohort of BRAF^V600‐mutant mCRC. Most analyses published so far are retrospective studies using uni‐ or pauci‐center databases or subgroups analyses from clinical trials. Only few prospective clinical trials dedicated to patients with BRAF mutant mCRC with limited sample size (n = 19–43) have been reported [25], [26], [27]. Our study confirmed the hallmarks of BRAF^V600 mutant mCRC tumors and their association with poor prognosis and chemoresistance. The median OS of 20.8 months and the median PFS in the first‐line setting of 4.3 months are consistent with results from clinical trials, even though no selection was prospectively performed. Stage IV disease (metachronous metastases), primary tumor surgery, and metastasectomy were factors significantly associated with OS, confirming prior results observed in mCRC [28]. Furthermore, we pointed out that dMMR and tumor resection are associated with a better prognosis. In this series, the median OS (20.8 months) was longer than the 15 months’ survival duration previously published [12], [29], [30], [31]. One possible explanation is the high rate of patients treated with metastasectomy in our cohort (15.3%). Indeed, these patients are usually excluded from therapeutic clinical trials administrating new drugs and chemotherapy regimens. Excluding patients who have had complete resection of their metastases and median OS of 8.9 (95% CI, 6.57–10.77) months is consistent with published results. When we also left out patients without any chemotherapy (treated with best supportive care only), the median OS was 9.3 months (95% CI, 6.93–11.10). The median PFS in the first‐line setting of 4.3 months was in accordance with literature results [32]. No obvious advantage was observed for irinotecan‐ versus oxaliplatin‐based CT in the first‐line setting, even focusing on patients who received bi‐chemotherapy (data not shown). Moreover, no clinical benefit of antiangiogenic versus anti‐EGFR therapy was observed. It must nevertheless be interpreted with caution because less than 20% of the patients had received anti‐EGFR therapies in the first‐line setting. Contrary to what was previously published, patients with dMMR tumors did not derive any statistically significant survival benefit from the addition of bevacizumab [33], [34]. Furthermore, we included patients in the present cohort before the publication of the TRIBE study, which explains the low percentage of patients treated with tri‐CT (<10%) [20]. However, these 25 patients treated with tri‐CT did not appear to profit from intensification in terms of neither PFS nor OS, whereas 28 patients in the TRIBE study experienced a better overall survival with FOLFOXIRI + bevacizumab (19.0 months) versus FOLFIRI + bevacizumab (10.7 months) [35]. The continuous administration of bevacizumab beyond progression after first‐line chemotherapy did not benefit the few patients treated (n = 29). Although more than half of the patients received a second line of chemotherapy (55.1%), only few of them received a third line (28.2%) in our series, as previously reported in the literature. Indeed, the pooled analysis of the population of 2,530 patients issued from three randomized clinical trials—FOCUS, COIN, and PICCOLO—showed that BRAF‐mutant patients tend to progress rapidly during or following the first‐line chemotherapy, with only 33% of the patients with BRAF‐mutant mCRC having received second‐line treatment versus 51% of the patients with BRAF wild‐type tumors (p < .001) [32]. This point may be relevant and useful for the management and used to design future clinical trials to help in the prioritization of investigational options for early treatment phases in these patients with poor prognosis. We identified two subsets of patients with BRAF‐mutant mCRC with a better outcome, the first being patients with dMMR tumors who experienced a longer OS than those with pMMR tumors. In mCRC, several studies highlighted the poor prognosis of patients with dMMR CRC versus those with pMMR CRC in the metastatic setting, but the prognostic impact of the MMR status in the subgroup of BRAF^V600E mutant mCRC has been poorly studied [36], [37]. Indeed, dMMR tumors are rare (≈5%), and dMMR‐BRAF‐mutant tumors consequently constitute a very small subgroup accounting for about 1%–2% of mCRCs [38]. To date, Venderbosch et al. published the largest study evaluating the effect of BRAF^V600E mutation and MMR status in patients with mCRC as a pooled analysis of four clinical trials including 3,063 tumors [12]: 153 (5%) were dMMR, 250 (8.2%) harbored a BRAF^V600E mutation, and only 53 (1.7%) were dMMR CRC with BRAF^V600E‐mutation. Median PFS and OS significantly decreased in patients with dMMR tumors compared with those with pMMR tumors, and in patients with BRAF^V600E‐mutant tumors compared with patients harboring BRAF‐wild type tumors. However, the poor prognosis of BRAF^V600E‐mutation was exclusively observed in patients with pMMR tumors, in contrast to previous publications that did not show any significant survival difference between patients with dMMR and those with pMMR mCRC. Similarly, Tougeron et al. reported no prognostic impact of BRAF mutation, as well as Lynch syndrome status, in a multivariate analysis of prognostic factors from a large cohort of 284 patients with dMMR mCRC [34]. Our exploratory analysis in the population of patients with a specified MMR status reported that dMMR was significantly associated with better OS, independently of immune checkpoint inhibitor treatment. Median overall survival of patients with mCRC with BRAF^V600E and dMMR tumors was 30.5 months compared with 16.5 months in those with pMMR BRAF‐mutant mCRC. This benefit persisted after adjusting on metastasectomy and antiangiogenic treatment. Of note, MMR status was not statistically significantly associated with PFS1. Finally, our results showed that patients for whom resection of liver or peritoneal metastases could have been performed had a better OS. Recent results showed worse prognosis and increased risk of recurrence in patients with BRAF‐mutated tumors who had experienced hepatic resection compared with patients with BRAF‐wild‐type tumors [39]. Gagnière et al. recently identified node‐negative primary tumors, carcinoembryonic antigen (CEA) ≤200 μg/L, and the Clinical Risk Score < 4 as favorable prognostic factors following hepatectomy in BRAF‐mutant patients, allowing them to be proposed aggressive strategies in the metastatic setting [40]. According to the few results reported in patients with peritoneal carcinomatosis, peritoneal resection does not seem to be contraindicated for BRAF‐mutant patients [41]. The surgical resection of liver metastasis contributes to improve survival in patients with BRAF‐mutant mCRC, and surgical resection of peritoneal carcinomatosis for these patients should also be considered as an option and discussed within a multidisciplinary team, especially for young patients with disease control with chemotherapy and low peritoneal carcinomatosis index [42]. Only few patients received either immune checkpoint inhibitors or anti‐BRAF therapies in this cohort. Pembrolizumab, as well as nivolumab with or without ipilimumab, has been approved by the U.S. Food and Drug Administration for the treatment of metastatic refractory dMMR tumors, whatever their BRAF mutational status, and therefore is an option for dMMR BRAF‐mutant metastatic CRC [43], [44]. For BRAF‐mutant tumors, the efficacy of the combination of the anti‐BRAF monoclonal antibody (like vemurafenib) with irinotecan and anti‐EGFR monoclonal antibody (like cetuximab) has been demonstrated in a randomized phase II study, leading to the implementation in the National Comprehensive Cancer Network guidelines of this regimen in BRAF^V600E‐mutant mCRC [45]. Furthermore, the results of the BEACON trial, evaluating the combination of a MEK inhibitor, a BRAF inhibitor, and an anti‐EGFR, are greatly awaited because of the magnitude of the effect observed in the first 30 patients [46]. The best therapeutic option for dMMR BRAF‐mutant tumors is unknown at present, but may include immune checkpoint inhibitors or MEK inhibitor plus BRAF inhibitor and an anti‐EGFR. The present cohort is the largest series of BRAF‐mutant mCRC, with very few missing data and patients unselected in contrast to clinical trials proceedings. Therefore, our study contributes to a better knowledge of the population of patients with BRAF‐mutant mCRC known to be associated with poor prognosis [29], [47], [48], [49]. We hypothesize that the poor prognosis in these patients with BRAF‐mutant mCRC may result from the development of chemoresistance mechanisms.

Conclusion

We identify stage IV disease (metachronous metastases), metastasectomy, and primary tumor surgery as prognostic factors that could help treatment decision making. Indeed, when metastases can be resected, metastasectomy should be performed to prolong survival. Chemotherapy regimen and biological agent (antiangiogenic or anti‐EGFR) were not associated with OS or PFS, suggesting that a cytotoxic doublet plus an antiangiogenic or an anti‐EGFR should be a therapeutic option for frail patients not eligible for tri‐CT. Finally, patients with dMMR BRAF‐mutant tumors seem to have a better prognosis and are good candidates for immunotherapy.

See http://www.TheOncologist.com for supplemental material available online.

Author Contributions

Conception/design: Christelle de la Fouchardière

Provision of study material or patients: Christelle de la Fouchardière, Romain Cohen, David Malka, Rosine Guimbaud, Héloïse Bourien, Astrid Lièvre, Wulfran Cacheux, Pascal Artru, Eric François, Marine Gilabert, Emmanuelle Samalin‐Scalzi, Aziz Zaanan, Vincent Hautefeuille, Benoit Rousseau, Hélène Senellart, Romain Coriat, Ronan Flippot, Françoise Desseigne, David Tougeron

Collection and/or assembly of data: Christelle de la Fouchardière, Romain Cohen, David Malka, Rosine Guimbaud, Héloïse Bourien, Astrid Lièvre, Wulfran Cacheux, Pascal Artru, Eric François, Marine Gilabert, Emmanuelle Samalin‐Scalzi, Aziz Zaanan, Vincent Hautefeuille, Benoit Rousseau, Hélène Senellart, Romain Coriat, Ronan Flippot, Françoise Desseigne, David Tougeron

Data analysis and interpretation: Christelle de la Fouchardière, Audrey Lardy‐Cleaud, David Tougeron

Manuscript writing: Christelle de la Fouchardière, Romain Cohen, David Malka, Astrid Lièvre, David Tougeron

Final approval of manuscript: Christelle de la Fouchardière, Romain Cohen, David Malka, Rosine Guimbaud, Héloïse Bourien, Astrid Lièvre, Wulfran Cacheux, Pascal Artru, Eric François, Marine Gilabert, Emmanuelle Samalin‐Scalzi, Aziz Zaanan, Vincent Hautefeuille, Benoit Rousseau, Hélène Senellart, Romain Coriat, Ronan Flippot, Françoise Desseigne, Audrey Lardy‐Cleaud, David Tougeron

Disclosures

Pascal Artru: Roche, Servier, Amgen, Merck, Pierre Fabre, Bayer (C/A); Eric François: Roche, Sanofi, Servier (C/A), Novartis, Amgen, Servier (H); Aziz Zaanan: Baxter, Roche, Merck Serono, Merck Sharp & Dohme, Amgen, Servier, Sanofi, Eli Lilly and Company (C/A); Benoit Rousseau: Bayer, Servier, Roche (C/A). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board