BRAF testing in metastatic colorectal carcinoma and novel, chemotherapy-free therapeutic options

In the past 25 years, treatment of metastatic colorectal cancer (mCRC) has undergone profound changes. The approval of newer chemotherapeutics such as irinotecan and oxaliplatin was followed in 2005 by the first targeted therapies, for example, monoclonal antibodies directed against the epidermal growth factor receptor (EGFR), as cetuximab and panitumumab, or the angiogenesis inhibitors bevacizumab, ramucirumab, and aflibercept. With the rapidly progressing molecular characterization of mCRC in the last 10 years and the classification of the disease in four consensus subtypes, further changes are emerging, which will promote, among other things, the introduction of protein-kinase inhibitors developed for specific molecular aberrations as well as immune checkpoint inhibitors into the treatment algorithm. Thorough molecular pathologic testing is indispensable today for guideline-compliant treatment of mCRC patients. In addition to RAS testing as a precondition for the therapy decision with regard to cetuximab and panitumumab, BRAF testing is of considerable relevance to allow decision making with regard to the newly approved chemotherapy-free combination of the BRAF inhibitor encorafenib and cetuximab in cases where a BRAF-V600E mutation is detected. Additional diagnostic tests should also include genome instability (microsatellite instability). Overall, more and more molecular alterations need to be investigated simultaneously, so that the use of focused next-generation sequencing is increasingly recommended. This overview describes the prognostic relevance of BRAF testing in the context of molecular pathologic diagnostics of mCRC, presents new treatment options for BRAF-mutated mCRC patients, and explains which modern DNA analytical and immunohistochemical methods are available to detect BRAF mutations in mCRC patients.

Despite significant advances in treatment, CRC continues to be one of the cancer entities with an unfavorable prognosis in Europe with about 250,000 deaths per year and an annual incidence of more than 500,000 new cases [1]. As the molecular characterization of metastatic CRC (mCRC) and the classification of CRC into molecular subtypes progresses [2], the number of options for the use of targeted therapies is also increasing, with molecular diagnostics becoming ever more important.
The German version of this article can be found under https://doi.org/10.1007/s00292-021-00942-9 More than 95% of all BRAF mutations are BRAF-V600 mutations, where valine (V) is mostly substituted by glutamic acid (E) at position 1799 in codon 600 (in exon 15) of the BRAF gene. Apart from this most frequent mutation, BRAF V600E , there are also less common mutations in codon 600, in which valine at position 1799 is substituted by lysine (BRAF V600K ), aspartic acid (BRAF V600D ), methionine (BRAF V600M ), or arginine (BRAF V600R ) [5]. Clinically, a comparison of the BRAF-V600 mutation with the significantly less common BRAF mutations in codons 594 and 596 shows that the former is more often found in right-sided and mucinous primary tumors with peritoneal metastasis, whereas BRAF 594/596 tumors have a better prognosis [3]. Unless expressly described otherwise, all the statements made in the following sections of this paper refer to BRAF V600E mutations in mCRC.
B-Raf is a key kinase in the Ras/RAF/ MEK-mitogen-activated protein kinase (MAPK) signaling pathway, which is involved in the regulation of cell growth. The alteration of the BRAF gene due to mutation leads to the constitutive activation of this protein kinase, thus causing uncontrolled cell division and consecutively leading to (neo-)angiogenesis and metastasis [6]. Resulting from studies on the CRC transcriptome, mCRC has been classified into four consensus subtypes (consensus molecular subtypes, CMS). KRAS mutations predominantly occur in the epithelial, "metabolic" subtype CMS3, which is characterized by metabolic dysregulation and partly also by chromosomal and microsatellite instability (MSI) [2,7]. BRAF mutations, however, are often seen in the "MSI-immune" subtype CMS1, which is dominated by somatic hypermutation and MSI.
It is extremely rare that BRAF mutations occur together with a mutation of the RAS gene. According to the current German S3 guideline, molecular testing for the presence of both mutations should be carried out before initiating first-line therapy, wherever possible. In this context, BRAF testing should best be done simultaneously with the RAS test or sequentially after exclusion of the RAS mutation [3]. epidemiological characteristics that differ from those of adenomas and which develop during a "classic adenoma-carcinoma sequence" based on mutations of the APC gene [9,10]. The BRAF-driven form of sessile serrated adenomas (SSA) leads to impaired apoptosis of crypt epithelia followed by senescence with epigenetic promoter (CpG) methylation and decreased expression of various genes (e.g., hMLH1, MGMT, p16) [3,9]. SSAs as a tumor pre-stage and precursor lesion are flat polyps that barely protrude from the mucosa, predominantly occur in the right-sided colon, and are difficult to detect even endoscopically [3,9]. Patients with large SSA have a higher risk of developing colorectal cancer; women with SSA have a five-fold higher risk than men [3].
BRAF mutation: a clearly negative prognostic factor in CRC In mCRC, advanced age is a negative prognostic marker, as is tumor location proximal to the left flexure [11]. As part of an investigation on further potential prognostic markers in this indication, the influence of BRAF mutations and MSI on metastatic spread and prognosis was analyzed within a large retrospective case series: BRAF-mutant tumors, particularly those harboring a V600E mutation, are associated with a significantly poorer overall survival (OS) than BRAF wild-type tumors (median 10.4 versus 34.7 months; hazard ratio [HR] = 10.66, p < 0.001), as well as with a higher rate of peritoneal and distant lymph node metastasis [12]. The prognostically highly unfavorable impact of BRAF V600 was also repeatedly reported in randomized controlled trials [4,13,14]; a detailed discussion of the prognostic impact of BRAF mutations and their connection with microsatellite stability and instability as a further biomarker can be found in two recent reviews [6,15]. Apart from hereditary non-polyposisassociated colorectal carcinoma (HN-PCC), MSI occurs in mCRC patients with an estimated frequency of only 4-8% [4]. If BRAF mutations and MSI occur simultaneously-the frequency is reciprocally about one third each-these alterations constitute sporadic defects of mismatch repair (dMMR) [3,4]. MSI patients appear to have a better prognosis than patients with microsatellite stability (MSS) [16]; although the number of published cases is still limited, the available clinical evidence suggests that patients with a BRAF V600E mutation and MSS have a poorer outcome than those with BRAF V600E and high MSI status. In the metastatic setting, the combination of BRAF V600E and MSS seems to predominate, with the BRAF mutation determining the poor outcome [6,13,17,18,19,20].
BRAF mutation: unclear predictive value with regard to former conventional therapies The predictive relevance of BRAF mutations for the use of anti-epidermal growth factor receptor (EGFR) therapy, i.e., the two monoclonal antibodies cetuximab and panitumumab, is currently under debate due to the fact that BRAF and RAS mutations are almost mutually exclusive [21] and that RAS mutations are known negative predictive factors for the use of anti-EGFR therapy. In CRC, cetuximab and panitumumab are approved for use in RAS wild-type patients only: for their use in patients with a BRAF mutation, only limited data are available from subgroup analyses of larger confirmatory studies (. Table 1), as well as from retrospective case series derived from clinical routine data [22,23].
Two partly overlapping meta-analyses confirmed a clinical benefit for anti-EGFR antibody therapy in patients with wild-type RAS and wild-type BRAF; in RAS wild-type patients harboring a BRAF mutation, however, data showed only a limited, non-significant clinical benefit in terms of progression-free survival (PFS) as well as OS [24,25]. The current body of evidence, on the other hand, does not justify the exclusion of anti-EGFR antibodies from the therapeutic repertoire for BRAF V600E -mutant patients either. Data from the German randomized phase-II study VOLFI comparing panitumumab plus chemotherapy vs. mono-chemotherapy in first-line treatment showed that the addition of panitumumab to chemotherapy tended to increase the overall response rate (ORR) in the 14 patients with BRAF-mutated tumors (odds ratio = 14.93, 95% confidence interval [CI] 1.03-200.00) [26].
Anti-VEGF therapy, which-as with anti-EGFR therapy-is given in combination with oxaliplatin-containing (mostly in first-line) or irinotecan-containing chemotherapy (mostly in second-line), is a clinically relevant routine treatment of mCRC and can be used independent of RAS status. However, since no or only indirect comparisons are available to date, the predictive value of BRAF testing with regard to this treatment regimen is still unclear. Results of a small phase-II study [27] and subgroup analyses of two large phase-III studies [28,29] also do not allow a clear overall assessment of the intensified chemotherapy backbone (FOLFOXIRI) in BRAF-mutant patients. A metaanalysis of five randomized studies of quite differing case numbers [range: 70-679] recently found that in BRAFmutated patients-with the total case numbers still being small-intensified combination therapy does not provide an additional benefit in the first-line setting [30]. A meta-analysis performed by the ARCAD study group, published in autumn 2020 and pooling data from two studies comparing chemotherapy plusanti-EGFR with chemotherapy plus-anti-VEGFR therapy as first-line options for mCRC, could not demonstrate a significant difference in OS for the subgroup of BRAF-mutated patients (n = 138) that received either bevacizumab-based or cetuximab-based therapy (HR = 1.01 [95% CI 0.69-1.48]) [31]. The benefit of anti-VEGF therapy with bevacizumab per se and the predictive role of BRAF for initiating bevacizumab-based therapy still requires further investigation (. Table 1).
Although targeted tyrosine kinase inhibitors have been used in clinical routine since 2011 with very good outcomes in BRAF V600 -mutated melanoma, BRAFmutated mCRC proved to be less sensitive to monotherapy [32,33]. The reason behind this may be CRC-specific resistance mechanisms in the MAPK signaling cascade. In-vitro studies demonstrated suppression of the negative feed-

Abstract
In the past 25 years, treatment of metastatic colorectal cancer (mCRC) has undergone profound changes. The approval of newer chemotherapeutics such as irinotecan and oxaliplatin was followed in 2005 by the first targeted therapies, for example, monoclonal antibodies directed against the epidermal growth factor receptor (EGFR), as cetuximab and panitumumab, or the angiogenesis inhibitors bevacizumab, ramucirumab, and aflibercept. With the rapidly progressing molecular characterization of mCRC in the last 10 years and the classification of the disease in four consensus subtypes, further changes are emerging, which will promote, among other things, the introduction of protein-kinase inhibitors developed for specific molecular aberrations as well as immune checkpoint inhibitors into the treatment algorithm. Thorough molecular pathologic testing is indispensable today for guideline-compliant treatment of mCRC patients. In addition to RAS testing as a precondition for the therapy decision with regard to cetuximab and panitumumab, BRAF testing is of considerable relevance to allow decision making with regard to the newly approved chemotherapyfree combination of the BRAF inhibitor encorafenib and cetuximab in cases where a BRAF-V600E mutation is detected. Additional diagnostic tests should also include genome instability (microsatellite instability). Overall, more and more molecular alterations need to be investigated simultaneously, so that the use of focused next-generation sequencing is increasingly recommended. This overview describes the prognostic relevance of BRAF testing in the context of molecular pathologic diagnostics of mCRC, presents new treatment options for BRAF-mutated mCRC patients, and explains which modern DNA analytical and immunohistochemical methods are available to detect BRAF mutations in mCRC patients.

Schlüsselwörter
Cetuximab · Encorafenib · BRAF-Inhibitoren · Proteinkinaseinhibitoren · Protoonkogene B-Raf-Proteine back loop between extracellular signalregulated kinase (ERK)and the EGFRunder BRAF monotherapy with overall high EGFR expression in CRC and possibly even stronger activation of the receptor by its ligands (. Fig. 1; [33,34]). This results in a reactivation of the EGFR pathway, e.g., by-passing the mutated BRAF protein via CRAF. Thus, it seems important to inhibit the EGFR pathway by simultaneously administering a therapy directed against EGFR in addition to BRAF block-ade in order to block the multi-track resistance mechanisms within the MAPK signaling pathway [33,35].

Therapeutic options in BRAFmutant mCRC
Until recently, first-and second-line treatment of mCRC has generally been based on the use of combination chemotherapies, mostly including-in the case of left-sided RAS wild-type tu-mors-EGFR antibody therapy, as described in the previous section [3,4] [4]), but also the Pan-Asian Adapted ESMO Consensus Guidelines published in early 2018. Although this article refers to the somewhat older ESMO guideline, the reference always applies to both guidelines and the current study evidence shown therein. Table 1 Studies and retrospective analyses on the significance of BRAF as a predictive marker in the use of anti-epidermal growth factor receptor therapies (A) and anti-vascular endothelial growth factor therapies (B) for BRAF-mutated metastatic colorectal cancer Study/phase (or type) of study Current treatment recommendations for mCRC are drawing much attention to the general condition of the patients, which are typically of older age. Regarding patients fit enough for systemic treatment, a distinction is made between the therapeutic objectives of "cytoreduction, " i.e., reduction of the tumor mass, and "disease control, " i.e., delaying further progression.

First-line therapy in BRAF-mutant mCRC: a controversial treatment standard
The combination of an antimetabolite (5-fluorouracil, plus leucovorin as a folinic acid derivative) and a platinum compound interfering with DNA replication (oxaliplatin) together with a topoisomerase I inhibitor (irinotecan) and the angiogenesis inhibitor (anti-VEGF) bevacizumab represents the current European guideline standard for BRAF-mutant mCRC patients in good general health [4]. However, considering the evidence level in the BRAFmutated subgroup, it has to be noted that this recommendation is based on a very small number of patients (N = 28) from the TRIBE study and is therefore associated with uncertainties [28]. In the BRAF subgroup of this phase-III study, the OS under FOLFOXIRI plus bevacizumab was 19 months with an ORR of 56%; however, there was no significant difference to the comparator group consisting of FOLFIRI plus bevacizumab (. Table 1; [28]). This first-line standard  [30]. Due to their poor prognosis, the concept of an 'aggressive' first-line treatment in BRAF V600E -mutated mCRC, using almost the entire arsenal of therapy modalities, is currently of clinical relevance with regard to the current recommendations for this mCRC patient collective-especially when cytoreduction is the therapeutic objective. On the other hand, it is unclear to what extent patients in Germany are treated with this intensi-fied first-line therapy, which is associated with relevant adverse events (AE).
As mentioned above, the issue of using anti-EGFR-based therapy in BRAFmutant tumors is currently the subject of controversial debates due to the inconclusive results of two meta-analyses [24,25].

New chemotherapy-free, targeted option after systemic therapy
Due to the limited therapeutic options after completion of first-line treatment, no clear recommendations could be drawn to date for second-and third-line therapy of BRAF-mutant mCRC patients [3,4]. The German S3 guideline stated in 2019 that "Individual (presently) not approved therapeutic approaches, e.g. with a BRAF inhibitor, MEK inhibitor and anti-EGFR antibody or, if possible, treatment within a clinical study (are) to be taken into consideration"; until recently, none of the therapeutic options mentioned here was approved for this indication.
In June 2020, however, the combination of the BRAF inhibitor encorafenib and the anti-EGFR antibody cetuximab was granted European Union (EU) approval, making such a chemotherapyfree, purely targeted dual blockade available for routine care. This combination is indicated for the treatment of adult patients with mCRC and a BRAF V600E mutation who have received prior systemic treatment [43].
The phase-III BEACON CRC trial, on which the approval was based, investigated the triple blockade with en- The tolerability of the dual blockade was slightly more favorable than that of the triple blockade and of the control group (. Fig. 2). The safety profile of the combination of encorafenib and cetuximab was well manageable and showed, in terms of AE, the expected class effects. The most common AEs included: elevated creatinine level (50%), nausea (34%), diarrhea (33%), low hemoglobin level (32%), fatigue (30%), acneiform dermatitis (29%), and decreased appetite (27%) [44].
Due to the comparable efficacy results of triple versus dual blockade and the slightly more favorable tolerability of encorafenib plus cetuximab, the European Medicines Agency (EMA) approved the dual combination regimen in June 2020 [43].
A new update analysis after a median follow-up of 12.8 months confirmed the above-described results with a consistent tolerability profile [45] Fig. 3).

BRAF diagnostics
As a consequence of the above-described clinical situation, diagnostic testing for the presence of RAS and BRAF mutations in mCRC is essential: for BRAF-mutated patients, encorafenib plus cetuximab can be taken into consideration as a newly available therapeutic option. The guidelines recommend performing these tests either before initiating first-line therapy in mCRC, or already at the time of initial diagnosis of CRC, in order to exclude the presence of Lynch syndrome via additional dMMR testing [3,4]. In the same way as BRAFmutated tumors, such hereditary CRCs without polyposis (HNPCC) also constitute a biologically distinct subtype of CRC. If a BRAF mutation is present in a dMMR/MSI tumor, Lynch syndrome can be mostly excluded. Thus, the determination of BRAF mutation status is of diagnostic and therapeutic relevance and helps to differentiate somatic from genetic "mismatch" repair defects [46,47,48]. The diagnosis of a sporadic tumor, thus excluding HNPCC/ Lynch syndrome, can be supported by analyzing MLH1 promoter methylation, since the presence of such methylation additionally corroborates the diagnosis of a sporadic, high MSI (. Fig. 4). BRAF testing can be performed either simultaneously with RAS testing or stepwise after excluding a RAS mutation. Nowadays, however, the simultaneous approach is recommended applying gene panel diagnostics based on focused nextgeneration sequencing (NGS).

Sample preparation
In the course of the diagnostic workup, specimens are mostly obtained during colonoscopy or surgical removal of the primary tumor. After fixation via 10% neutral buffered formalin (4% formaldehyde) for 24-48 h and embedding in paraffin, the specimens are well suited for the tests indicated in . Fig. 4. For molecular determination of MSI status, healthy tissue samples distant from the tumor should also be stored and analyzed [4]. If tissue samples contain a low percentage of tumor cells, a macrodissection prior to DNA extraction is highly recommended for cancer cell enrichment.
In recent years, NGS methods have been increasingly used in molecular diagnostics, allowing the detection of selected genes/genomic regions relevant for diagnosis and therapy (targeted NGS) simultaneously with high sensitivity and specificity. Therefore, BRAF mutation testing in CRC is often no longer carried out as an isolated individual test but integrated into the parallel detection of other molecular alterations such as KRAS and NRAS. The sensitivity of NGS-based methods is generally very high (approximately 1% detection limit), but hampered by artifacts occurring during formalin fixation. In many laboratories, a threshold of 5% variant allele frequency (VAF) is therefore requested, which can however be undercut in specific situations [49,50]. Today, ThermoFisher Scientific and Illumina are the most prominent NGS platforms available. Both platforms enable analysis of numerous commer-cial or in-house gene panels, based on either amplicon-based (multiplex polymerase chain reaction, PCR) or hybridcapture methods for enriching the selected target regions. Numerous bioinformatic programs are available for the evaluation of NGS data. However, these should be used by scientists/physicians that have profound experience in molecular diagnostics.
In the BEACON CRC study, which was conducted in a total of 221 centers in 28 countries (111 of which were centers in Europe), evaluation of the procedures used for BRAF status determination from 510 samples revealed the following picture: in 48.8% of the analyses, single gene detection was still used for BRAF testing; protein-based methods (immunohistology) were used in 0.7% of the analyses. However, the majority of BRAF tests were performed together with the detection of other gene alterations (e.g., as focused, amplicon-based NGS) (50.5%). Discrepancies observed betweenlocal and central testing showthe relevance of standardization of diagnostic procedures, especially in view of the increasing importance of targeted therapeutic approaches: clear confirmation of the locally detected BRAF V600E mutation was found in just 90.7% of central testing. Of note, this discrepancy was largely due to insufficient neoplastic tissue in the sample, most likely resulting from the fact that BRAF-mutant tumors are generally associated with mucinous adenocarcinoma that contain fewer tumor cells. In 1.6% of the central repeat tests, the local result was clearly negated. Taking into account this possibility of a discrepancy between local and central testing, the study protocol allowed the inclusion of patients based on local BRAF V600E detection in molecular prescreening, but additionally required central confirmation, as part of the inclusion criteria, within 30 days of initial receipt of study medication. Once the study had reached the pre-specified number of discrepant test results, the assay, which was developed and then approved in the USA as a "companion diagnostic, " became a prerequisite for the inclusion of all further patients.
The EU and US regulations on invitro diagnostics (IVD) are fundamentally different: in the US, such tests are subject to central approval (Premarket Approval) by the Food and Drug Administration (FDA), while in the EU manufacturers can chose an accredited "notified body" that evaluates the conformity of their test; once confirmed, the manufacturer is allowed to label its product with the so-called CE label (CE: Confor- mité Européene) 2 . Furthermore, the FDA usually grants approval for targeted therapies only in conjunction with a defined and simultaneously approved companion diagnostic, which, at the time of drug approval, has an exclusivity of use: the usage of the respective companion diagnostic is therefore a prerequisite in the USA for the prescription of the drug by the physician [54]. The US legislation distinguishes such standardized, mostly commercially available tests from so-called laboratory-developed tests (LDT), which-like the classical test methods such as Sanger sequencing-are designed, validated, and applied by institutes for their own use. The German Accreditation Body (DAkkS) refers to such LDTs as "in-house tests." They are normally not subject to formal approval or labeling requirements; however, a few years ago, the DAkkS issued a guideline for the validation of molecular pathological examination methods [55,56]. In the US, a discussion paper was presented by the FDA in January 2017, advocating stronger prospective regulation of LDTs due to their increasing prognostic and predictive importance-this applies in particular to the role and growing importance of NGS [57,58]. With the FoundationOne ® CDx (F1CDx) test, an 2 With the "CE label, " the manufacturer, distributor, or EU authorized representative declares under EU Regulation 765/2008 "that the product meets the applicable requirements stipulated in the harmonization legislation of the Community on its affixing. " NGS method was approved as a companion diagnostic for the first time in the USA in late 2017 (. Table 3; [59]).

Review
Against this background, the diversity of competing classical and modern DNA analytical methods for BRAF mutation determination is easier to understand. Of the various commercial procedures using allele-specific PCR techniques, only the Qiagen therascreen ® test is currently recommended for BRAF-mutated mCRC in the US; in Europe, this test is CE-labeled. It is to be expected that for the tests currently approved for melanoma only, appropriate adjustments will soon be made in the US with regard to CRC. In the BEACON CRC approval study, the only methods allowed by the study protocol were PCR and NGS based on local assays [44].
A comparison between commercially available (i.e., FDA-approved) tests and LDTs for EGFR, KRAS, and BRAF mutations showed that there was no overall difference between the methods and the three tested genes in assay performance; the average analysis accuracy was 97% [60]. Since testing for KRAS, BRAF, MSI/dMMR, MLH1, and possibly other genes constitutes a prerequisite of CRC diagnostics, panel-based assays are understandably more prominent in current pathological practice-in Germany, all major university and non-university institutions are now using focused NGS for this purpose. Frequently used platforms include Illumina (MiSeq™ or NextSeq™) and Thermo Fisher (Ion Gene Studio S5™) [61,62,63]: in a multicenter validation study across Germany, a high level of consistency between different NGS platforms and gene panels was shown; apart from CRC, samples of lung and breast cancers were also tested [61].

Immunohistochemical methods
Besides DNA analytical methods, protein-based analyses using the VE1 antibody may provide an alternative to molecular pathological testing for BRAF V600E ; the latter is widely regarded as the gold standard in BRAF mutation testing [50,64]. At the same time, protein-based immunohistological detection is the only reasonably practicable method for determining the expression level of mutant BRAF protein. This method is also applicable for MSI testing. It is characterized by a specificity of 98-100%, a sensitivity of 85-100% [49,50], and an in-lab turnaround time of 1 day. Thus, the method is generally reliable, but some challenges remain, such as establishing a standardized scoring of protein expression, which is needed to avoid a substantial number of misclassifications [64]. Immunostaining is in principle a fast and cost-effective method for the determination of BRAF-mutant protein; however, as CRC meanwhile requires the determination of multiple alterations, DNA analytical methods should certainly be preferred nowadays.
Chu et al. investigated outcomes of immunohistochemical (IHC) and NGS testing in a cohort of almost 1900 CRC patients [65]. The rate of false-positive IHC tests was 17%; however, confirmatory re-testing by NGS was performed in only 43% of the IHC-tested patients. NGS-tested patients had a favorable median OS, at younger age, and a lower rate of synchronous metastases and a higher rate of therapy. The authors concluded that NGS should be considered as standard testing but that IHC might serve as an optional screening test if NGS testing is not available in a timely manner. This is underlined by the rapid availability of results via IHC and the finding that reflexive IHC testing made it possible to identify 57% more BRAF-mutated mCRC than standard NGS methods. For this article no studies with human participants or animals were performed by any of the authors. All studies performed were in accordance with the ethical standards indicated in each case.
The supplement containing this article is not sponsored by industry.
Open Access. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.