Background
Pathogenic mutations in the
BRCA1 and
BRCA2 genes confer a high lifetime risk of breast (and ovarian) cancers [
1]. Over the 19 years since the identification of the
BRCA1 and
BRCA2 genes [
2,
3]; genetic testing requests to identify pathogenic mutations in
BRCA1 and
BRCA2 have risen steadily. NHS
BRCA testing in the UK was introduced gradually from the mid 1990’s and has traditionally been delivered through a clinical genetics model driven typically by a strong family history of breast cancer. Demand on the genetics service has increased steadily, with peaks of referral for testing sparked by greater public awareness from press reporting of high profile figures such as Angelina Jolie [
4,
5]. A growing interest by oncologists in novel targeted approaches to the management of triple negative breast tumours, together with recognition that this phenotype, (particularly at young ages), is associated with a higher frequency of
BRCA1 mutation carriers, has also increased requests for more rapid access to genetic testing [
4]. Young onset triple negative breast cancer cases are eligible for genetic testing even without a family history of breast or ovarian cancer [
6]. Faster access to genetic testing without the need for referral to genetics specialists is one of the aims of the “Mainstreaming of genetic testing” agenda [
7].
A
BRCA genetic test can yield 3 possible results: Positive (a pathogenic mutation is found), negative (no mutation detected or a variant of no clinical significance – a polymorphism) or a variant of uncertain significance (VUS). A VUS is an alteration in the gene sequence with unknown consequences on the function of the gene product or risk of causing disease. One classification proposed in 2008 by a workshop of experts convened at the International Agency for Research into Cancer, suggested a 5 point classification scheme following a parallel system to that used in reporting cytology or radiology results [
8]. The scheme classifies mutations that are irrefutably pathogenic as class 5, and class 4 variants have a probability of being pathogenic of greater than 95 % taking into account all available data about the variant which may be derived from many different sources. Class 4 and 5 variants can be used for predictive testing in unaffected relatives. A class 3 variant (VUS) has a 0.05–0.949 probability of being pathogenic; evidence may be limited or conflicting and the closer to 95 % probability a variant reaches, the more useful additional family and functional studies are to improve the classification. Therefore Class 3 variants are usually reported out by diagnostic laboratories in order to facilitate further studies. Some of these variants behave as low penetrance gene mutations and should not be managed in the same way as a highly penetrant mutation [
9]. This is a complex concept to communicate to a patient who may have undertaken genetic testing with the main aim of gaining access to new treatment options. A class 3 variant would not allow them to be entered into trials aimed at carriers of clearly pathogenic mutations. Most will not have any functional effect but some may have relevance for other family members.
The frequency of VUS reports from different laboratories varies worldwide and depends on testing prevalence and the ancestry of the population served. In African-American populations the rate can be up to 21 %, 5–6 % in individuals of European ancestry in the USA, and 15 % in European laboratories [
10,
11]. A record of over 1500 VUS results [
12] (as well as pathogenic
BRCA1 and
BRCA2 coding variants) is held in a number of publicly accessible databases. However, these databases vary in how well annotated and curated they are and no public databases currently permit an iterative process of gathering cumulative evidence to reclassify variants. NHS (National Health Service) genetic testing laboratories undergo a national quality assurance scheme to review detection rates of pathogenic mutations and ensure that the Association for Clinical Genetics Science (ACGS) guidelines for assessing the pathogenicity of variants are observed, although these are not specific for the
BRCA genes [
13]. However there is no standard template for
BRCA reporting and no nationally adopted classification scheme.
Testing recently diagnosed breast cancer patients as opposed to testing unaffected individuals brings unique challenges with complex issues that require careful consideration and a multidisciplinary approach. Common queries include the optimum timing of the genetic test in relation to the patient’s cancer treatment, selecting which patients without a family history require a genetic test and deciding the management of a patient with a very strong family history but a negative
BRCA genetic test result as well as what to do with a VUS result. In the UK, a 2011 report from the Foundation for Genomics and Population Health (PHG foundation), “Genetics and mainstream medicine” set out a new strategy whereby in the future medical specialities including oncologists will incorporate genetics into their standard practise supported by higher level regional genetics services [
7]. The Royal Marsden Hospital is currently piloting the move of
BRCA genetic testing in breast and ovarian cancer patients away from the genetic clinics into a combined oncology-genetics model. The advantages of more patients having access to genetic tests and a more streamlined process must be balanced with concerns about whether oncologists are prepared or equipped to take on not only the ‘easy’ results but issues such as what to tell the patient when the significance of the result is uncertain.
Previous work has looked at patients’ [
14], General Practitioners’ [
15,
16] and medical specialists’ knowledge of
BRCA genetic testing, [
17,
18] including oncologists’, [
19] but there is little published describing the knowledge of non-geneticist breast cancer specialists particularly with regard to a VUS result. The only study to our knowledge specifically looking at
BRCA VUS knowledge is a survey of genetic counsellors in the US [
20] where significant variation in personal interpretation and management recommendations existed. Another study of patients with a VUS, referring family physicians and genetic counsellors concluded that national VUS-related guidelines were required [
21].
In this study we aimed to explore the current knowledge among breast oncologists and breast surgeons at specialist training/registrar and consultant level UK-wide of how to use a VUS result. We explored whether clinicians made correct interpretations and appropriate management choices when presented with two anonymised patient genetics test results taken verbatim from 2 different UK genetics laboratories reporting a VUS result. To place this in context we also approached medical geneticists to gain an insight into the variations in laboratory practice in interpreting and reporting VUS results. This variation in reporting styles from individual laboratories may not be known to individual BCS, but as trainees and consultants work in different UK geographical regions during their career the issue of heterogeneity within UK genetics laboratories is an important point to illustrate.
Methods
Ethical approval was not required for this survey. Personal, sensitive personal and confidential data was not collected from study participants. Written consent was not required for these quantitative survey, however implied consent was taken by participating in the anonymous survey. Participants were informed that answers were going to be used for publication. The collection method by “survey monkey” and questions meant no identifiable data was collected and the response could not be linked by the researchers to an individual.
Breast cancer specialists survey - questionnaire 1
A questionnaire was sent electronically in September 2013 to all members of 3 large national organisations: the UK Breast Intergroup, the Association of Cancer Physicians and the Association of Breast Surgeons. Members of these mailing lists included medical and clinical oncologists specialising in breast cancer and breast surgeons, at specialist trainee or consultant level practising within the UK. The breast specialists’ questionnaire was designed by the authors with 10 questions and included questions on demographics, level of genetics training, referral practice for genetic testing, general knowledge of VUS and the interpretation and communication of two anonymised genetic test results reporting a VUS which had been issued by two different UK NHS diagnostic laboratories transcribed verbatim. Additional file
1: Report 1 summary:
“A missense mutation in exon 11 BRCA2 gene (unclassified variant) and change in exon 13 of BRCA2 (rare polymorphism)”. Additional file
1: Report 2 summary:
“Heterozygous for BRCA2 c.9098C > T, p.Thr3033lle (clinical significance unknown)”. Full reports and questions are provided in Additional file
1. The questions were close-ended with limitations on multiple responses except for the communication to patients of the two genetics reports. Free text boxes were included to capture further responses and to allow for qualitative analysis.
The term VUS was deliberately not included in the questions about communication of the report to patients, so respondents could not just intelligently guess the answer in the knowledge this was a survey about VUS. The six responses (and other-free text box) were further categorised into appropriate, inappropriate and don’t know responses to allow significance testing between specialities. Appropriate responses, as determined by experts in cancer genetics, were “Explain there may be a hereditary cause and discuss further tests” and “Refer patients to a genetics consultant”. Inappropriate responses were “Reassure the patient that there is no hereditary cause for her breast cancer”, “Explain BRCA2 mutation contributed to causing her breast cancer”, and “Explain she has a BRCA2 mutation discuss risk reducing options”. Don’t know also included blank responses. Responses from the free text were reviewed and classified into the appropriate category by BKE.
Geneticists survey - questionnaire 2
The second survey was also sent electronically from the ENIGMA group (Evidence-based network for the interpretation of germline mutant alleles) to Medical Geneticists in December 2012 and January 2013; only respondents working in the UK have been included.
The geneticists’ questionnaire included study participants demographics questions and level of clinical experience, referral patterns, laboratory workload, BRCA and VUS reporting proportions, actions to clarify clinical significance of VUS and classification systems used.
Statistical analysis
Analysis was performed in STATA ver. 11.2. Descriptive statistics were used to describe the study population characteristics. Differences between disciplines (oncologist versus surgeon) in Table
2 were tested by Pearson chi squared test. Communication of the report to patients were categorised in dichotomous variables appropriate/inappropriate and ‘don’t know.’ Fishers exact test was used to test differences in communication of report to patients by speciality excluding don't know responses. Free text comments were scored manually and results recorded thematically.
Discussion
This survey was designed to gain a snapshot of the broad understanding of VUS’s amongst non-geneticist breast cancer specialists in the UK. We tested their interpretation and anticipated patient communication using two actual reports where a VUS had been identified. Our study demonstrates that most breast cancer specialists currently directly refer their patients to genetics service for a BRCA test. Perceived genetics knowledge is not uniform over differing specialities although genetics training is mainly limited to undergraduate level or as part of preparation for postgraduate exams. Overall the majority of survey respondents felt uncertain about their ability to understand the implications of the reported variants based on the genetics reports they were presented with. There was a significant difference between specialities with surgeons feeling more confident than medical oncologists (p = 0.030). The results from report 2 highlight the potential for wide variation in interpretation of genetics reports by non-geneticists and the uncertainty around management generated particularly in a patient without a family history of breast or ovarian cancer. However this group of patients with for example young onset triple negative breast cancer but no family history are those who are most likely to be tested directly in the oncology clinic as part of the mainstreaming agenda.
The geneticists that responded to the survey reported that their laboratories reported fewer VUS as a proportion of all reports (1–10 %) than might be anticipated from the literature or was thought to be likely by the breast cancer specialists surveyed (10–20 %). This may be because with increasing volumes of tested and reported BRCA tests over many years, a previously classified VUS may be more often recognised as polymorphisms in the tested population and therefore no longer reported as a VUS.
We deliberately chose two recent genetic test reports from laboratories where differing presentation styles reflect the variation in UK Genetics laboratories report format. There is no standardised national reporting template and not all laboratories give clear clinical guidance on management (i.e. for a VUS this would be a clear statement that the VUS should NOT be used for predictive testing and that the family history should guide management) [
8,
23]. Equally there is no consensus surrounding a single classification method for VUSs within the genetics community [
24,
25] as further demonstrated by the results of our geneticists survey where multiple classification systems were used with no clear majority leader. This can lead to further confusion when non-geneticists receive reports potentially from different laboratories. The American College of Medical Genetics and Genomics and the Association for Molecular Pathology have issued a joint guideline for interpretation of sequence variants [
26].
Multiple approaches aimed at understanding the pathogenicity of a variant in any gene exist and these are set out in the ACGS guidelines [
13]. Some methods involve in-silico analysis but some require additional samples from other affected family members or tumour samples to fully assess [
12,
25,
27]. From this survey it can be seen that UK geneticists commonly adhere to guidelines and use multiple methods (median 5). These independent lines of evidence are more powerful if combined and can be used to give a posterior probability of pathogenicity of a
BRCA VUS with resulting classification that can be linked to clinical actions [
20,
25]. Thus, for a VUS, referral onwards to the genetics services is a necessary action; however, it is important for referring clinicians to convey to their patients that most VUS are not pathogenic, most remain of unclear significance and very few end up with sufficient additional evidence to formally move into IARC Class 4 (i.e. sufficiently certain to be used as a predictor of future risk).
Other studies have found varying knowledge of breast cancer genetics by physicians. A survey looking at US oncologist’s general
BRCA knowledge found only 40 % correctly answered all 4 questions [
19]. In contrast a Dutch survey reported better knowledge on hereditary breast cancer by surgeons, medical and radiation oncologists and radiologists with average score of 6.1/7 correctly answered questions [
17]. However neither study specifically looked at
BRCA VUS knowledge. Certainly there is considerable scope for physician (and patient) misunderstanding of
BRCA results leading to inappropriate patient care [
28,
29] and uncertainty about clinical management [
23]. In our study most breast cancer specialists correctly interpreted report 1 but found report 2 more challenging with a greater degree of uncertainty.
The additional burden of an uncertain genetic test result during the emotionally charged time of a cancer diagnosis may encourage patients towards risk reducing surgery that may be inappropriate. Several studies report risk reducing surgery rates amongst patients receiving a VUS result. These vary widely from similar to those tested and receiving a negative report 7–10 % RRM [
14,
30] to high rates (42 %) within one year after receiving a VUS result [
31]. Genetic testing at the time of cancer diagnosis, sometimes called “treatment focused genetic testing” is already part of the mainstreaming agenda for more personalised treatment and is being implemented routinely in some areas of the UK particularly with a view to facilitating entry into clinical trials. The GTEOC study [
32] is looking at upfront genetic testing at time of ovarian cancer diagnosis irrespective of family history and the OlympiAD trial [
33] is recruiting
BRCA gene carriers with distant metastases. Trials of targeted therapies in
BRCA mutation carriers in the adjuvant setting are now recruiting and require early identification of gene carriers.
Patients find that genetic counselling is helpful (>90 %) and reduces cancer distress [
14]. Yet genetic counselling services in the NHS have insufficient resource to manage the inevitable increase on demand for rapid testing. Mainstreaming is inevitable and this study demonstrates that there is an urgent need to ensure not only adequate education in germline genetic testing for cancer specialists but also integration of the cancer geneticist into the multidisciplinary team. Agreement of a national reporting template to provide unambiguous test reports to the receiving non-expert clinician with clear instructions for the clinical utilisation of the test result that can be audited is needed.
Limitations
As with all studies using survey methodology, data collection is limited by selection bias of respondents most interested in the subject matter. It is recognised that the answers in the interpretation of the reports were not mutually exclusive and therefore there may have been more than one “correct” answer depending on the participant’s interpretation.
Competing interests
The authors declare that they have no competing interests.
Authors contributions
BE participated in study design, collated data, analysis and drafted the manuscript. DE conceived of the study design, collated partial datum and participated in the manuscript editing. EC participated in study design and in manuscript editing. TM provided statistical analysis. JA participated in manuscript editing. All authors read and approved of the final manuscript.