Download Development of Testing Criteria for Familial Breast/Ovarian Cancer

UK Genetic Testing Network
Developing testing criteria for familial breast and ovarian cancer:
incorporating NICE guidelines
15th July 2014
Chandos House, London
Meeting report
Executive summary
The revised NICE guidelines1 on the classification and care of people at risk of familial breast
cancer and management of breast cancer and related risks in people with a family history of
breast cancer were published in 2013 with one of the aims being to promote equity in
access to BRCA1/2 germline mutation testing for this group of individuals. A number of key
changes with respect to the offer of genetic testing are presented in the revised guidelines,
most significantly:
 A lowering of the threshold at which testing is offered. Testing is now offered when
the likelihood of a pathogenic mutation is 10%, as opposed to the previous threshold
of 20%.
 Testing of unaffected relatives of breast/ovarian cancer patients, and provision of
testing where an affected relative is unavailable for testing is now suggested.
Given this change in testing guidelines, the UK Genetic Testing Network (UKGTN) invited
professionals working in the field of breast cancer to a meeting to draw up consensus
testing criteria for BRCA1/2 gene testing in breast and ovarian cancer patients and their
relatives that would complement the revised NICE guidelines. The participants included
clinical geneticists, genetic counsellors, oncologists, patient representatives and clinical
scientists. The workshop was chaired by Dr Fiona Lalloo, chair of the Cancer Genetics Group.
Presentations described current statistics on detection rates; the impact of the revised NICE
guidelines on patient experience and healthcare services; the current practice for genetic
testing and the interface with oncology practice; strategies to rationalise testing approaches
across regions; and current diagnostic technologies. The presentations were followed by a
group discussion on consensus testing criteria which are shown below.
1
NICE Familial breast cancer: Classification and care of people at risk of familial breast cancer and
management of breast cancer and related risks in people with a family history of breast cancer. Available at:
http://www.nice.org.uk/guidance/cg164
1
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Based on the group’s discussions and statistical evidence showing the likelihood of
identifying a pathological BRCA1/2 mutation in different groups of patients with breast and
ovarian cancer and their relatives, the following testing criteria were agreed by consensus,
which aim to include all patients and relatives who have a greater than 10% likelihood of
having a BRCA1/2 mutation. Although the NICE guidelines do not include patients with
ovarian cancer, this group of patients are included in the UKGTN testing criteria as they have
a significant likelihood of having a BRCA1/2 mutation. It should be noted that the threshold
of 10% is not absolute, and testing may be offered to individuals whose ‘risk’ is below this
level, according to clinical judgement.
Recommended UKGTN consensus testing criteria
2
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
UKGTN Testing Criteria
Approved name and symbol of disorder/condition(s):
Hereditary Breast and Ovarian Cancer
OMIM number(s):
604370 & 612555
Approved name and symbol of gene(s):
BRCA1 and BRCA2
OMIM number(s):
113705 & 600185
Patient name:
Date of birth:
Patient postcode:
NHS number:
Name of referrer:
Title/Position:
Lab ID:
Referrals will only be accepted from one of the following:
Referrer
Tick if this refers to
you.
Consultant Clinical Geneticist/Registered Genetic Counsellor
OR named Multi-Disciplinary Team clinician:
Consultant Oncologist
Consultant Gynaeoncologist
Consultant Breast Surgeon
Minimum criteria required for testing to be appropriate as stated in the Gene Dossier:
Criteria
Tick if this patient
meets criteria
Woman with breast cancer who has ONE of the following:
1. Bilateral invasive ductal breast cancer and both cancers diagnosed <40
years
2. Grade 3 triple negative breast cancer diagnosed <40 years or <50 if family
history unclear or unknown
3. Non- mucinous epithelial ovarian cancer
4. A first-degree relative* with breast cancer and both diagnosed <40 years
5. A first-degree relative* with a histologically confirmed non-mucinous
epithelial ovarian cancer
6. A family history with a pathology adjusted Manchester score greater than or
equal to 15
OR Woman with ovarian cancer who has:
1. Histology consistent with a high- grade serous epithelial carcinoma OR
2. A family history with a pathology adjusted Manchester score greater than or
equal to 15
3
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
OR Man with a BRCA-related (prostate, breast or pancreas) cancer who has:
1. A family history with a Manchester score greater than or equal to 15
OR Affected individual with Ashkenazi Jewish/ Polish ancestry who has:
1. Female breast cancer diagnosed <50 or a male BRCA-related cancer
(founder mutation screen)
OR Unaffected individuals:
Referrals only accepted from Consultant Clinical Geneticist or Registered
Genetic Counsellor
1. Unaffected individual who has a family history with a Manchester score
greater than or equal to 20 AND a first-degree relative with breast/ ovarian/
prostate/ pancreatic cancer where there are no affected relatives available
for testing (ovarian cancer and cancer in first-degree relative should be
confirmed)
2. Unaffected individual with Ashkenazi Jewish/Polish ancestry who has a firstdegree relative with female breast cancer diagnosed <50 or a male BRCArelated cancer and a Manchester score greater than or equal to 10 (founder
mutation screen$)
*Or a second-degree relative via a father
$
This should only be done if testing cannot be performed in an affected relative
OR At risk family members where familial mutation is known.
Additional Information:
If the sample does not fulfil the clinical criteria or you are not one of the specified types of referrer
and you still feel that testing should be performed please contact the laboratory to discuss testing of
the sample.
First-degree relative: parent, sibling, child
Second-degree relative: uncle, aunt, nephew, niece, grandparent, grandchild, half-sibling
4
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Introduction
Understanding of genetic factors in the aetiology of breast and ovarian cancer has advanced
rapidly in the past twenty years. Around 5-10% of breast cancer is hereditary. In the 1990s,
the first genes to be identified in familial cases of breast cancer were BRCA1 and BRCA2.
Mutations in these two tumour-suppressor genes account for around 2-3% of all breast
cancer cases, and testing has been available in the UK for the past 15 years. Hundreds of
mutations in these genes have been reported which result in an increased susceptibility to
breast and ovarian cancer. BRCA2 also increases the risk of other malignancies, such as
prostate and pancreatic cancer. Genetic testing may target known familial mutations or
examine the entire coding region of the genes. Many other genes have subsequently been
identified which impact on the risk of breast and ovarian cancer, for example ATM and
PALB2, although at present mutation testing in the UK is usually limited to analysis of the
BRCA1/2 genes unless there is a clinical indication for targeted testing of other genes. There
is a move towards panel testing using Next Generation Sequencing (NGS) technology in
research and some areas of clinical practice.
Influence of mutation status in the care of unaffected and affected individuals
Knowledge of BRCA1/2 mutation status in unaffected individuals may influence decisions
regarding risk reduction strategies including risk reducing mastectomy and salpingooophorectomy or tamoxifen usage. For this reason, the NICE guidelines emphasise the
testing of individuals with a significant family history to allow appropriate preventative
choices to be made. A large body of evidence has now shown that BRCA1/2 mutation
carriers differ from non-carrier breast cancer patients in terms of their response to certain
therapies and disease recurrence rates. In breast cancer patients’ knowledge of mutation
status may impact on adjuvant and later treatments, contralateral risk-reduction options
and eligibility for clinical trials.
Testing algorithms
Data continues to be collated on ever larger numbers of individuals who have undergone
BRCA1/2 testing, which helps to refine algorithms which estimate the likelihood of
identification of a pathogenic mutation. Numerous predictive algorithms have arisen to aid
decision making including BOADICEA, the Manchester scoring system and the Tyrer-Cuzick
model. Each varies in terms of input, but all take into account age at diagnosis and the
number of relatives with breast or ovarian cancer. BRCA1/2 mutations are also seen at
different frequencies in patients with histologically distinct tumours. For example BRCA1/2
mutations are rarely seen in HER2 positive breast cancer, and therefore histology is
considered in some predictive algorithms. The revised NICE guidelines recommend the use
5
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
of one of these algorithms to calculate this likelihood and recommend the offer of BRCA1/2
testing when the likelihood of identifying a mutation is greater than or equal to 10%.
Consistency in testing
Rates of genetic testing in familial breast and ovarian cancer vary across clinical specialties
in secondary care, between NHS trusts and across regions. The revised NICE guidelines aim
to promote consistency by recommending that testing is offered to all individuals and their
relatives who have a greater than 10% likelihood of having a BRCA1/BRCA2 mutation. The
publication of the revised NICE guidelines with implications for genetic testing, means that
UKGTN testing criteria for BRCA1/2 should be concordant, and reflect consistency in testing.
Therefore the UKGTN is committed to facilitating the development of consensus testing
criteria for BRCA1/2 testing which is compliant with NICE guidelines. These will support
clinicians and laboratories in implementing the guidelines and increase consistency in the
provision and access to BRCA1 and BRCA2 testing for hereditary breast and ovarian cancer
in the UK.
6
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Workshop introduction, objectives and scope
Dr Shehla Mohammed introduced the workshop and explained the scope of the meeting to
consider BRCA1/2 testing in breast and ovarian cancer patients and their unaffected
relatives, and the objective to produce consensus testing criteria (compliant with the 2013
revised NICE guidelines) based on professional opinion from those working with patients in
the field. The testing criteria should therefore promote testing which maximises benefit and
minimises harm to patients, whilst representing appropriate use of NHS resources.
Dr Fiona Lalloo chaired the workshop and is the chair of the Cancer Genetics Group. As a
member of the Medical Genetics Clinical Reference Group (CRG), Dr Lalloo has also been
asked to be involved with developing guidelines on the wider breast cancer care pathway
for NHS England. It was hoped that the agreed testing criteria would also inform and
complement this process.
Dr Mohammed began by providing some background on the UKGTN process to evaluate
new genetic tests for NHS service. UKGTN reviews the analytical and clinical validity and
clinical utility of new tests. Testing criteria are a part of the appraisal system and provide
key clinical features in a test target population, and specify from whom referrals for testing
should be made. Testing for genes associated with breast cancer has existed for some time
prior to the inauguration of the UKGTN and BRCA1/2 testing is therefore referred to as a
‘grandfather test’. Consequently no formal testing criteria currently exist for BRCA1/2
testing. In light of the revised NICE guidelines it was therefore felt timely to propose
consensus testing criteria for these tests. This decision was supported by the presentation of
data from the UKGTN’s 2012 audit of genetic testing showing the rates of BRCA1/2 testing
and variation between different UK regions. Limitations of the data were highlighted, in that
they reflect activity from UKGTN member laboratories only, but nevertheless, the 2012
rates of testing varied by a factor of six within England, and a factor of 15 within the UK as a
whole.
7
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Presentation summaries
1 Clinical background
Professor Gareth Evans, Manchester Centre for Genomic Medicine, St Mary’s Hospital,
Manchester
Introduction
The 2013 NICE guidelines replace the 2004 guidelines for care and management of
individuals with familial breast cancer. Some key features include recommending the offer
of a test to women and men who meet a 10% risk threshold, regardless of whether they are
affected themselves; provision for risk calculation in women who are already affected by
breast cancer, and discussion of risk assessment in both primary and secondary care
settings. The guidelines are specific to patients with breast cancer or breast cancer family
history. Whilst links between this disease and ovarian cancer are well known, this could not
be covered within the guidelines.
It is important to consider how the recommended thresholds were decided upon and to
acknowledge that the guidelines recommend the offer of a test at or above a given
threshold, but this does not mean that this offer cannot be considered below the threshold.
A cost effectiveness study was carried out in creating the guidelines and, based on the
current price of tests, (test prices are falling), it was shown to be cost effective to test
affected and unaffected individuals with a family history with a likelihood of a BRCA1/2
mutation of greater than 10%. There was uncertainty around the 5-10% risk level, and this
group were originally included, but removed after further consideration of comments during
the consultation process from clinical geneticists.
The guidelines make clear the need to use tools for risk assessment, without recommending
one specifically from the various options including BRCAPRO, BOADICEA and the
Manchester scoring system.
New recommendations in the 2013 guidelines suggest referral to specialist genetics clinics
for genetic testing during initial management or any point thereafter. Discussion with the
specialist genetics team should take place as part of the multidisciplinary approach to care.
Fast-track genetic testing (within 4 weeks of breast cancer diagnosis) should be offered only
as part of a clinical trial. This will normally only arise where the results of genetic testing
would inform decisions on pre-surgical treatment neoadjuvant chemotherapy. Any such
requests should be discussed with a consultant in cancer genetics, and if appropriate, with
the laboratory.
8
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Manchester scoring system and risk thresholds
The Manchester scoring system scores each cancer in a direct lineage based on age at
diagnosis of breast or ovarian cancer. For BRCA2, pancreas and prostate cancer are also
scored. The combined Manchester score gives an assessment of risk of both BRCA1 and
BRCA2 mutations. Given the relationship between histological findings and underlying
genetic mutations, a modified Manchester scoring system was subsequently created, which
adjusts the overall score according to the tumour histology. HER2 positive cancer is rarely
seen in BRCA mutation carriers and therefore an adjustment of -4 is made to the combined
score, whereas grade 3 triple negative breast cancer (HER2, ER, PR negative) is commonly
seen in BRCA mutation carriers, conferring an adjustment of +4.
Data from a paper by Evans et al.2 from 2005 showed that a combined Manchester score of
20-24 met the 2004 NICE guideline threshold of 20%, whilst a score of 15-19 meant
individuals were above the current 10% threshold. Updated 2014 figures on a larger cohort
of over 3000 women show that in the 20-24 category, where one might be offering testing
to an unaffected relative of that affected proband, it comfortably meets the 20% risk in the
proband and 10% threshold in the unaffected relative. At the 15-19 level, it no longer easily
meets the 10% risk threshold in the unaffected relative. Significantly, below 15 points the
likelihood plummets to 4-5%.
Ovarian cancer
A substantial amount of testing for ovarian cancer is based on evidence from a paper by
Alsop et al.3 which stated that around 16% of patients with non-mucinous ovarian cancer
were found to have a BRCA mutation, as were around 23% of patients with high-grade
serous ovarian cancer. However, the detection rate in women with high grade serous
ovarian cancer over the age of 60 without a family history of breast/ovarian cancer was only
6.4%. Unselected cases (which therefore includes those with a family history) of clear cell
and endometroid ovarian cancers show a mutation detection rate of less than 10% at any
age. Therefore it is important to note that data from published studies should be
interpreted with caution in terms of overall detection rates, and it is worth considering that,
within a clinical group, there will be subgroups with greater and lower risk.
Triple negative breast cancer
Grade 3 triple negative tumours are the most commonly observed breast cancer subtype
observed in patients with BRCA1 mutations. A study by Robertson et al.4 recommended
2
Evans et al. J Med Genet 2005; 42:e39
Alsop et al. J Clin Oncol 2012; 30(21):2654-63.
4
Robertson et al. Br J Cancer 2012; 106(6):1234-8.
3
9
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
testing all individuals under 50 years of age for BRCA1 with triple negative breast cancer on
the grounds that there is a pick up rate of greater than 10% in unselected cases. Over the
age of 50, the detection rate is relatively low, at around 4% of unselected cases. Further
work is needed to clarify the detection rates in sporadic cases, but in cases without a family
history, 5 out of 44 cases (11%) were found to have a BRCA1 mutation diagnosed in their
thirties from the combined Manchester POSH studies5. Although 4/38 (11%) were found to
have a mutation in their forties in the Robertson study4 with a Manchester score below 15,
these were not pathology adjusted.
Further data has shown that grade 3 triple negative breast cancer with a histology adjusted
Manchester score of 10-11 points comfortably meets the 10% threshold but at 7-9 points is
below the threshold for BRCA1. However, looking at the BRCA1/BRCA2 combined score, 1519 points does come in above the 10% threshold for BRCA1, but under 15 points is below
the 10% threshold for most cases (although the confidence intervals would probably include
10%).
Looking at age at diagnosis, data from a number of studies on grade 3 triple negative breast
cancer (unselected series or selected on the basis of family history) was presented and this
showed that patients diagnosed with a tumour at 40 years or younger did meet the 10%
threshold. After age 50 the detection rate for BRCA1 plummets in grade 3 triple negative
breast cancer. The important question is at what age between 40-50 this effect is seen.
Looking at the detection rate for an unadjusted Manchester score of less than 15 and then
adding 4 points for triple negative breast cancer, the 10% threshold will not be met.
Therefore it is not proven that sporadic breast cancer between the ages of 40-49 with triple
negative breast cancer will meet the 10% threshold.
Other proposed criteria
Professor Evans presented data from over 3000 tested cases with cancer which showed that
those fulfilling the category of patients with bilateral breast cancer and a relative diagnosed
with breast cancer under 60 years of age who had a Manchester score of <15, had a
detection rate of only 5% (4/78).
The 10% threshold and UKGTN testing criteria
It was felt that some categories should be included without question including: women with
bilateral breast cancer under 40 years of age; individuals with breast cancer and a firstdegree relative with breast cancer both diagnosed before 40 years of age; individuals with
breast or ovarian cancer and a family history with a Manchester score of greater than 15 or
unaffected with a first-degree affected relative and a Manchester score of 20 or more and
5Evans
et al J Med Genet 2011; 48(8):520-2
10
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
any grade 3 triple negative breast cancer under 40 years of age. Professor Evans showed
that, assuming a threshold of 10%, testing of sporadic ovarian cancer cases should be
limited to high grade serous ovarian cancer with a cutoff of 60 years of age.
Data presented showed that the following criteria should not be used as many would be
well below the 10% threshold. These were: patients with bilateral breast cancer and a
relative diagnosed with breast cancer under 60 years of age; individuals with a first-degree
relative with a non-mucinous epithelial ovarian cancer.
11
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
2 View and impact of recent NICE guidelines
Dr Julian Adlard, Yorkshire Regional Genetics Service, Leeds
New NICE guidelines: key changes
Changes in the new NICE guidelines on Familial Breast Cancer were summarised. These now
include breast mammography/MRI screening recommendations for women who have family
history and have been affected with breast cancer themselves. Increased frequency of
breast mammography is recommended than in the previous guidelines, particularly for
women at moderately increased risk. Recommendations regarding MRI screening in the
previous guidelines incorporated factors such as age-related risk and breast density. The
latter has been de-emphasised in the new guidelines, and the recommendations are more
focused on the presence, or suspected presence, of a BRCA/TP53 gene mutation.
Chemoprevention with tamoxifen or raloxifene is now included and should be offered to
women with moderate or high risk of breast cancer. Finally, the guidelines incorporate
reduced thresholds for genetic testing and the inclusion of cases for testing where an
affected relative is unavailable.
Impact on laboratory services
The lowered risk threshold for testing has led to increased workload for laboratories and an
audit of referrals showed a marked increase following publication of the new NICE
guidelines, and publicity surrounding the Angelina Jolie case. A significant amount of this
workload constitutes testing unaffected patients (new in the 2013 guidelines), in addition to
lowering the threshold for affected women.
Difficulties with new guidelines
The length and complexity of the guidelines was felt to be a drawback e.g. taking into
account age, risk, whether or not already affected, and type of mutation there are about 30
categories which women could fall into regarding mammographic/MRI screening. The use of
the terms ‘offer’ and ‘consider’ screening or chemoprevention was felt to be unclear, and
open to interpretation with factors such as patient preference, individual clinical opinion,
and commissioning viewpoints potentially coming into the equation. Affected high risk
women can now have MRI screening irrespective of BRCA testing, but the guidelines do not
define how to classify an affected woman as remaining at high risk. This is very complex in
practice, as it includes not only family history and genetic testing results, but also histology,
treatment received and prognosis from the first cancer diagnosis. The definition of an
affected relative being ‘unavailable’ for testing was also left unclear, potentially leading to
differences in interpretation.
12
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
With regard to chemoprevention with tamoxifen, the guidelines do not recommend the age
at which to start treatment. Furthermore, there remain concerns about GP’s willingness to
prescribe tamoxifen, which is currently unlicensed as a preventative therapy. As most
clinical geneticists do not prescribe drugs, this could lead to difficulties for patients in
accessing treatment. The new NICE guidelines do not come with specific patient information
documents. This has provided NHS services, specialist groups and cancer charities with
challenges. It is likely to lead to inefficiency and variation, with multiple groups attempting
to develop their own information leaflets/booklets.
There is no directly associated funding for implementation of the new guidelines, which may
lead to inconsistencies in implementation depending on local/regional commissioning,
particularly regarding screening/tamoxifen in the moderate risk/‘consider’ categories. Extra
gene testing must be met by laboratory budgets, with no clear source of increased funding
and this inevitably leads to competing pressures.
Uniform implementation
With the overall aim of unified implementation of the 2013 guidelines in the Yorkshire &
Humber Strategic Clinical Network (SCN), a scoping exercise was carried out to identify
stakeholders, to describe the current service configuration, and to assess the funding
implications of implementation. To give some idea of the complexities of the exercise, the
Yorkshire/Humber region has a population of 6.3 million; one strategic clinical network; two
regional genetic services; 12 NHS trusts; 20 CCGs and around 800 GP practices. The NICE
guidelines therefore impact on a large number of professionals and in different clinical
specialities in a large number of hospitals and GP practices. In this region, the Strategic
Clinical Network decided to use the 2013 guidelines as an example of how to implement a
uniform guideline across the network. The aims of the exercise were identified as working
towards full compliance, including the implementation of genetic testing for those above
the risk threshold of 10%, and production of guidance and information on clinical pathways
for the network.
It was clear from the scoping exercise that there were already variations within the SCN
region, for example in the screening services that were being offered, and how and where
risk assessment was taking place. Resource gaps were evident, such as for mammography,
where most radiology services felt that there was no current capacity, without additional
funding, to meet the extra screening suggested in the new guidelines.
In terms of genetic testing, it was clear that there had already been some move downwards
from the 20% risk threshold, so it was felt that the new 10% threshold might not be as
challenging as first thought. It was decided to meet the 10% threshold using the Manchester
scoring system with a cutoff of 15 for an affected person and 17 for an unaffected individual
with a first-degree affected relative. In partnership with the SCN, some guidance was
13
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
produced detailing how to do risk assessment and which patients would be referred to
clinical genetics. This was mainly aimed at secondary care services, but also for use in GP
practices, and was accompanied by education sessions to raise awareness amongst health
professionals in the region.
Testing
Details on how to determine breast cancer risk or current eligibility for genetic testing in the
NICE guidelines may not be specific enough. For some families the decision to test is
relatively straightforward, as the phenotype and family history clearly indicates a high risk
scenario. However, other family histories are less clear. Risk assessment tools are important,
but the NICE guidelines do not specify a single tool, nor advise on action if different tools
produce different results. The guidelines also do not advise on risk assessment of
patients/families who return a BRCA negative screening result (around 85% of tests). Some
risk algorithms such as Tyrer-Cuzick and BOADICEA allow input of BRCA-negative tests to
reassess the risk to family members, but the NICE guidelines give no specific indication of if
and how such reassessment should be made.
There is currently a national debate about testing all ovarian cancers (of specific histology)
even where there is no other family history. However, this scenario is not included in the
NICE guidelines. In Leeds, this testing is not currently commissioned to be offered, despite
oncologists being eager for it to be performed for treatment purposes. The molecular
genetics laboratory is currently at capacity, so in order to implement this, further funding
would be required and this needs to be considered with commissioners. Realistically, the
implementation of testing in ovarian cancer may ultimately be achieved through the
mainstreaming model, where oncologists will make the offer of testing and clinical genetics
services will only see people with a family history or positive result.
Future direction
New technology has impacted on testing possibilities and it has become clear that a number
of other genes impact on risk status in cancer (such as E-cadherin, RAD51 genes), raising the
possibility of panel testing. In future, therefore, risk assessments may be better served by
testing a wider range of genes and this may be particularly relevant in the significant
proportion of families who present not just with breast cancer, but with other malignancies
as well.
Panel tests of genes associated with cancer risk are now potentially available. These include
some genes which have only recently been verified, or confer moderate rather than high
risk. Generally, laboratories are not routinely testing for them and they have not yet been
approved by the UKGTN.
14
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Clinicians may request testing for selected genes on panel tests which may plausibly explain
the family history. In this case, the results reported will be limited to the genes requested.
For example in a mixed breast and bowel cancer family, data on around 10 genes might be
reported. The referring clinician must consider the need to appropriately counsel the
patient and consider the implications of cascade testing and management options in the
event of a positive test.
Looking forward, there is a trajectory from the current situation of rigid testing criteria for a
few genes, to potentially testing multiple genes in wider groups of patients at initial
presentation. The benefits of this are a more equitable approach with reduction in
‘rationing’, which satisfies patient demand and avoids the need for complex eligibility
criteria. However, it is constrained by resource issues and leads to the greater likelihood of
generating results of uncertain significance.
The cost of a standard test (e.g. for BRCA1 and 2) is around £530, whilst the available panel
test is about £860. Panel tests require additional interpretation time for laboratory scientists
as well as equipment/reagents.
Summary
The Yorkshire & Humber region is not yet fully compliant with the updated NICE guidelines.
The reasons for this include gaps in screening and secondary care risk assessment, and
resources needed for clinics to implement this. Negotiations between local commissioners
and secondary care services are continuing. The need for patient information is being
addressed by various groups including the UK Cancer Genetics Group and cancer charities.
The benefits of the NICE guidelines were felt to be the fact that they are based on a robust
review of the evidence; they function to raise awareness; they set benchmarks and apply
pressure for service improvements, promoting additional genetic testing and breast
screening. The drawbacks include the complexity of the guidelines; the difficulties
associated with a circumscribed remit; the slow production and consequent ‘shelf life’ of the
guidelines; the lack of associated funding for implementation and consequent variability
with which they are implemented.
15
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
3 Review of current practice for genetic testing
Dr Marc Tischkowitz, Clinical Genetics Service, Cambridge
Background
A summary of current mutation prediction and breast cancer risk algorithms was presented
showing that all take into account family history, but they differ with respect to other
inputs, for example histology or lifestyle and reproductive history. Therefore clinical
judgement is very important in interpreting risk relating to the 10% threshold.
The 2013 NICE guidelines offer examples of acceptable algorithms such as BOADICEA and
the Manchester scoring system, but do not recommend any one system. Other systems
include Tyrer-Cuzick, Claus, BRCAPRO, and for non BRCA1/BRCA2 risk calculation, the PTEN
mutation probability calculator. Each system has strengths and weaknesses, for example the
BOADICEA system is not easy to use in clinic, whereas the Tyrer-Cuzick system is userfriendly and therefore quite popular in secondary care.
Such algorithms are vital in assessing an individual’s position relative to the testing
threshold, but there are questions as to whether thresholds will continue to be necessary. In
an era of falling test prices, one might ask whether thresholds are in fact necessary. For
example, the cost of fluorescence in situ hybridisation (FISH) testing for HER2 in breast
cancer is in the region of £100-200, compared to BRCA1/BRCA2 mutation testing at around
£400. If test prices for BRCA1/BRCA2 testing can be reduced further this may preclude the
need for thresholds.
Some yields were outlined, with around 10-15% of women with non-mucinous epithelial
ovarian cancer testing positive for BRCA1/BRCA2, 5-10% of women with grade 3 triple
negative breast cancer; 5-6% of women with breast cancer under the age of 50 and around
2% of women with Ashkenazi Jewish heritage.
Other genes and panel tests
Moving on from testing restricted to BRCA1/BRCA2, recent work on the PALB2 gene was
presented which highlights the potential for inclusion of other genes in risk calculations. The
PALB2 gene confers a lifetime risk that overlaps with BRCA1/BRCA2, and is more penetrant
than previously thought. There are panel tests available in the US and elsewhere such as the
Myriad or BROCA tests, which cost in the region of $1500-3000 and typically include testing
for 10-30 genes. However, the evidence for including some of these genes is questionable
and in order to implement such technology in the UK, the cost would need to decrease with
scale. At the 2014 Hereditary Breast and Ovarian Cancer Symposium in Montreal
(http://www.BRCAsymposium.ca) delegates were asked to define which genes on these
panels they would recommend for testing in BRCA1/2 negative breast cancer patients and
16
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
their relatives. There was consistency for most genes which would be offered both to
affected patients and for predictive testing, whereas others were much less likely to be
recommended for predictive testing.
Some known founder mutations are associated with increased breast cancer risk and
targeted analysis can be used to help these families. However, most mutations are familyspecific and this presents a challenge to identify the family-specific mutation involved.
Genetic testing in epithelial ovarian cancer (GTEOC)
Information was presented from an ongoing study in the East Anglia region which aims to
recruit 390 ovarian cancer patients over 24 months. Women with newly diagnosed ovarian
cancer are offered genetic testing following test information and consent procedures given
over the telephone. Testing has so far been conducted on 68 samples. Three BRCA1 and
three BRCA2 mutations have been found along with eight variants of uncertain significance
(VUS). In cases where a mutation was found, referral for genetic counselling and cascade
testing was initiated. The benefits for clinical genetic services were highlighted as they
maintain overall control of testing, and so can monitor and engage with families in need of
care, without shouldering all the burden of an increased workload. We estimate that
coordinating a testing service this way requires 0.5 FTE of a genetic counsellor for a
population of 2.5 million but this model does not require an overall increase in manpower
as referrals for isolated ovarian cancers are no longer seen by clinical genetics. Such a model
sits between current counselling-based, resource-intensive models of clinical genetic service
provision and other models where the whole process is devolved to other specialities.
Service implications
The NICE guidelines have significant service implications in terms of increased volume of
tests. The technology available is changing, and technical barriers present less of a problem,
whilst service delivery, in terms of counselling is more problematic. The future may involve
greater use of NGS; higher throughput leading to lower costs; rapid testing services e.g. for
BRCA1/BRCA2, and better integration of clinical genetics with oncology services.
17
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
4 A view from the oncology clinic
Professor Andrew Tutt, Kings College and Institute of Cancer Research, London
As part of oncology service delivery, engagement with clinical genetics is very important, not
just from a patient perspective, but also in terms of clinical case management. Patients have
become much more aware of their disease, familial risk and treatment options, in part due
to coverage of these issues in the media. However, it is important to note that these
discussions come at a time in their care when patients are having to consider many other
aspects such as therapeutic options. Those therapeutic options may well be influenced by
results of genetic testing but further evidence is needed, and collaboration between
oncology and clinical genetic services is therefore critical.
Genetic testing in the care pathway
From a clinical perspective there are a number of reasons for wanting to know a patient’s
BRCA1/ BRCA2 mutation status at different points in the care pathway.
Early in the care pathway, these include:
Choice of imaging/therapeutic/prophylactic options. The mutation status may influence
choice of imaging modality; the choice of neoadjuvant therapy in treatment (early
adopters), and in clinical trials (as mutation carriers are known to have differing response
e.g. to platinum based therapies). Knowledge of BRCA mutation status can inform a
patient’s decision with regard to contralateral mastectomy, as some plastic surgical
techniques cannot be performed sequentially. Knowledge of mutation status may prompt a
decision to undergo prophylactic mastectomy and therefore radiotherapy treatment (with
its associated cardiac risk) would not have been recommended.
Later in the care pathway, reasons include:
Follow up and long term surveillance/prevention. BRCA mutation status can inform
decisions regarding long term follow up with MRI and prophylaxis e.g. risk reducing bilateral
salpingo-oophorectomy (RRBSO).
Communication within families. Knowledge of a mutation is essential to enable informed
discussion within families regarding risk, screening, prevention and reproductive options.
Therapeutic options in metastatic disease. Previously, less emphasis has been placed on
genetic testing in patients with advanced disease, but evidence from research trials suggests
that mutation status is relevant to treatment options in this group, in terms of response.
However, due to the clinical situation, such decisions need to be made rapidly and this can
present a challenge for clinical genetics services.
18
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Current clinical trial data
Neoadjuvant therapies
Informing therapeutic options with genetic testing is a burgeoning field and many clinical
trials are ongoing in this area. Eligibility for clinical trials can be influenced by knowledge of
BRCA1/BRCA2 mutation status. Current clinical trials are testing PARP-inhibitors in early
breast cancer (many women at this stage have not had their mutation status assessed) and
are therefore increasing knowledge about the relative effect of these drugs in mutation and
non-mutation carriers. Data presented at a recent ASCO meeting by Telli et al6 described the
impact of neoadjuvant therapies in patients with triple negative breast cancer and showed
that this was particularly successful in patients with BRCA1/BRCA2 mutations, with 56%
patients in this group (9 out of 16) achieving pathological complete response (pCR).
A further study by von Minckwitz et al.7 was described, using neoadjuvant therapies in triple
negative breast cancer patients, with a family history of breast cancer or ovarian cancer (not
all were BRCA mutation carriers). 54.5% of BRCA1/BRCA2 and RAD51C mutation carriers
achieved pCR but a significant improvement was also seen in non-carrier patients (pCR=
41.6%). The results suggest that other genes may also be significant in non-BRCA/RAD51C
mutation carriers, and raises the possibility of wider testing, perhaps panel testing, which
may inform treatment choices which would benefit patients beyond those with BRCA1/2
mutations.
Other studies are examining the association between mutation status and DNA repair
capacity. In a paper by Telli et al.8, homologous recombination deficiency (HRD) assays have
been used to predict response to treatment in triple negative breast cancer and it was
found that defects in homologous recombination in tumour cells is predictive of a good
response in BRCA1/2 mutation carriers and non-carriers. This raises the possibility of
discovering germline mutations in other homologous recombination DNA repair genes
which may influence response to therapies which target this area of cell function.
Unanswered questions remain about the relative utility of identifying this type of
therapeutic target through HRD assays (which measure the consequence) or identifying the
underlying mutation (the cause).
6
Telli et al. Abstract available at: http://meetinglibrary.asco.org/content/114316-132
Von Minckwitz et al. Abstract available at: http://meetinglibrary.asco.org/content/131770-144
8 Telli et al. Info at: http://www.ascopost.com/ViewNews.aspx?nid=8471
7
19
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Palliative chemotherapy
Knowledge of BRCA1/ BRCA2 mutation status may also influence palliative chemotherapy
choices. A study by Isakoff et al.9 described differing response to single-agent platinum
chemotherapy in metastatic triple negative breast cancer. BRCA1/ BRCA2 mutation carriers
were seen to have greater response to such therapies than patients with the wild type
allele.
Further trials
Trials are ongoing to examine the correlations between treatment effectiveness and a
number of somatic and germline genetic features. For example, BRCA mutation status is
being used to define eligibility for further trials looking at treatment efficacy in advanced
HER2 negative breast cancer.
Summary
There are numerous points in the care pathway at which information regarding mutation
status for BRCA1/BRCA2 and other alleles can influence treatment decisions and therefore
clinical genetics services must expect referrals from a wide group of patients and for
differing reasons. In addition, a large group of referrals come from unaffected individuals
with a family history to inform screening and prevention decisions. Clinical genetic services
must be configured to deal with this workload effectively and consider the optimal method
for informing and counselling patients, whether through genetic counsellors, clinicians or
nurses.
9
Isakoff et al. Abstract available at: http://meetinglibrary.asco.org/content/130285-144
20
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
5 Perspective from the Pan London approach
Dr Adam Shaw, Guy’s and St Thomas’ Hospital, London
The Pan-Thames Consortium of Cancer Genetics centres comprises the four nationally
commissioned services covering the four quadrants of London and surrounding counties,
plus the Cancer Genetics service of the Royal Marsden Hospital, covering a total population
of around 17 million. The BRCA1 and BRCA2 genes were discovered in the mid 1990s and
testing for these genes became part of clinical practice in the UK in the early 2000s.
However due to cost and capacity, the availability of testing within NHS regional genetics
services laboratories was limited. NICE guidance first published in 2004, triggered
widespread expansion in BRCA1/2 testing with the aim of offering screening of the whole
coding sequences of both genes, to those patients with a 20% or greater likelihood of
harbouring a mutation. Clinical genetics services across the UK used a variety of family
history assessment algorithms to identify such patients. However all such tools have
limitations, with relatively poor sensitivity and specificity, and different algorithms given the
same input data will give different outputs.
By 2011, cost, quality, and speed of genetic testing had improved significantly. However
variation in family history assessment practices created anomalies whereby individuals from
the same family, but living in different parts of South-East England, would receive discordant
views on eligibility for testing. The Pan-Thames Cancer Genetics consortium agreed to
harmonise testing criteria, and audited outcomes over a six month period. These novel
criteria used a combination of simple situation-based triggers, and the Manchester scoring
system for more complex scenarios.
Compliance with the new criteria was excellent (99%) and overall, mutations were detected
in 19% of patients tested. In 2013, NICE guidance 164 reduced the threshold to a detection
probability of 10%, and for the first time recommended testing individuals unaffected by
cancer and without a known familial mutation. Testing criteria were reviewed in the light of
this and other evidence on mutation rates in specific clinical groups. For example studies of
cohorts of women with ER/PR/HER-2 negative breast tumours Robertson et al. (2012)10, and
epithelial ovarian cancer (Arnold et al. (2011)11, Alsop et al. (2011)12). The most recent
agreed criteria across the consortium are shown in Figure 1.
10
Robertson et al. Br J Cancer 2012; 106(6):1234-8
Arnold et al. Info available at: http://meetinglibrary.asco.org/content/59051?format=posterImg
12
Alsop et al. Abstract available at: http://meeting.ascopubs.org/cgi/content/abstract/29/15_suppl/5026ASCO
11
21
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Figure 1. BRCA1/BRCA2 testing criteria for Pan-Thames Consortium 2014
B Woman with
breast cancer who
has
O Woman with
ovarian cancer who
has
1) bilateral BC and both cancers diagnosed <50 yrs
2) triple negative BC diagnosed <50 yrs
3) OC
1) histology consistent with epithelial carcinoma
M Man with BRCA
related cancer who
has
U Unaffected man/
woman who has
1) a family history with a MS ≥15
1) a MS ≥17 and a FDR with breast/ ovarian/ prostate/ pancreatic
cancer where there are no affected relatives available for testing (OC
and cancer in FDR confirmed)
A recent audit of test results at the Guy's service showed that 309 tests were performed in
the 12 months to September 2013. 34 pathogenic mutations were identified (11%) and 38
genetic variants of uncertain significance (12%).
Agreeing criteria across services covering a large population has brought several advantages
including bringing consistency and clarity to patients, and aiding the management of
families that are spread across regions. Difficulties may arise when trying to apply fixed
criteria to different models of service configuration, and different populations. It is also
important to maintain the ability to revise criteria efficiently in response to new evidence
and changes in clinical and laboratory practice.
22
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
6 Patient perspective
Dr Emma Pennery CBE, Breast Cancer Care, London
Background
Breast Cancer Care is the only UK-wide charity providing specialist support and tailored
information for anyone affected by breast cancer. Their clinical expertise and emotional
support network help many thousands of people find a way to live with, through and
beyond breast cancer. They have more than 100 publications and receive nearly 12,000 calls
to their helpline and 1,600 ask-the nurse email enquiries every year. Their moderated online
discussion forums have more than 500,000 registered users. Around 3-4% of helpline calls
are from individuals concerned about family history but who have not had a diagnosis of
breast cancer themselves. Their booklet on family history has been updated in response to
the 2013 NICE guidelines. The challenge of presenting some of the technical information at a
level which could be easily understood was highlighted.
Impact of media coverage
The media coverage following publication of the revised NICE guidelines focused almost
exclusively on chemoprevention drug treatment and largely failed to mention other
important aspects such as the lower threshold for testing; the testing of affected women
and testing when affected relatives are unavailable. There are inherent difficulties in trying
to achieve a balanced, but perhaps less ‘headline- grabbing’ press release for the media.
The media coverage resulted in some confusion and on the basis of enquiries, this included
misunderstandings about the relative difference between drugs versus surgery on the
magnitude of risk reduction and concerns about or lack of awareness of side effects
associated with tamoxifen. There was concern about the fact that GPs may be unwilling to
prescribe the tamoxifen because of its license indications. Concerns were expressed about
tamoxifen, but not raloxifene which is also mentioned in the NICE guidelines. There was also
some confusion about the length of time it should be taken (five years or longer) because of
concurrent publicity about extended adjuvant hormone therapy.
Online forums: key questions from patients
A snapshot of online forum topics was presented and key discussion areas outlined: Am I at
risk? How do I get assessed? I have no living relatives so does this preclude me from testing?
Along with questions about appropriate screening follow-up by gene carriers already
affected by breast cancer. Other issues of concern included delays in accessing treatment
and the complexities of family dynamics in communication of risk.
23
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Triple negative breast cancer (HER2, ER and PR negative breast cancer)
The impression from discussions with patients is that genetic testing is not offered as much
as expected for this group of women. There were reports that some women had not been
offered testing on the basis of their age, and testing occurring after treatment rather than
before.
Chemoprevention
There is some confusion about whether the low uptake of chemoprevention in high risk
women in the year since the NICE guidelines is due to it not being offered or not being
accepted. Of note, in the US, where FDA approval for chemoprevention was granted 10
years ago, around 60% of women stop taking the drugs before the full five years. In the UK,
despite it being offered to most eligible women, centres report only a handful of women
taking up chemoprevention. Therefore relatively low rates of high risk women taking
tamoxifen or raloxifene may not reflect access/availability issues, but rather unwillingness
from patients to take or continue to take this preventive treatment. Unresolved issues
remain, which the NICE guidelines did not cover, including the appropriate age to start
chemoprevention and who is responsible for monitoring and investigating side effects of
treatment as care of these women can fall between clinical genetics teams and GPs.
24
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
7 Impact of emerging technology platforms
Dr Andrew Wallace, St Mary’s Hospital, Manchester
Laboratories have seen a large increase in workload in recent years in parallel with
improved understanding of genetic influences in cancer and rapid technological
developments. There is therefore a need to consider carefully the workload trajectory, for
example with regard to testing of ovarian cancer patients.
Next generation sequencing (NGS)
The technological revolution has come in the transition from Sanger sequencing to next
generation sequencing (NGS) methods. Whilst the benefits of NGS are clear in terms of
capacity, lower costs and automated analysis, there are some important limitations which
must be considered.
Platform heterogeneity
There has been rapid evolution of platforms and there is considerable variation in
performance characteristics, in terms of throughput and cost. The choice of instrument, run
mode, library preparation methods, input sample quality and sample multiplexing can all
influence this. This results for example in differences in read length and run time. Each
platform varies in the rate and types of error. For example, the Illumina sequencing
chemistry has an error rate of around 0.1% and the primary errors are base substitutions,
the Ion Torrent sequencing chemistry has an error rate of 0.46-2.4% and the primary errors
encountered are indels.
Assay validation
There are many different platforms, and the technology is constantly evolving. Therefore
switching to a new platform or assay may seem tempting, but there are significant
overheads associated with validation of a new assay.
Assay performance
In contrast to Sanger sequencing, NGS technologies are less well able to detect some classes
of mutation, namely indels > 20 bp, which do comprise a proportion of BRCA mutations.
Some GC-rich or repetitive regions of genes are difficult to cover and variant calling can be
problematic at low read depth. This is one reason why laboratories currently confirm all
abnormal results with Sanger sequencing.
Sample preparation and enrichment
There are choices to be made with respect to DNA preparation and sample enrichment. PCR
and hybridisation methods are employed for sample enrichment, with the latter being
25
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
better suited to panel based testing. Each has pros and cons, for example with respect to
incidental findings. In essence, laboratory-developed enrichment methods are cheaper to
use but validation costs are higher.
Bioinformatic analysis
In contrast to Sanger sequencing, data from NGS assays must undergo in depth analysis to
allow accurate interpretation. Various options are available in terms of commercial platform
specific or custom pipeline tools. The commercially available platform specific tools benefit
from being easy to use, but lack the discriminatory power of custom pipeline analysis. The
latter provides highly processed data which filters out SNPs and takes into account read
depth, therefore reducing analysis time. However this requires capital investment with the
need for at least one bioinformatician.
Future directions
Copy number variants (CNVs) in the form of indels, (which are seen in a proportion of BRCA
mutations), still present a problem and enrichment with PCR methods is particularly
unsuited to this. Hybridisation methods are more appropriate for providing CNV data, but
this still relies on complex bioinformatic analysis and the robustness for clinical diagnostic
testing remains to be demonstrated.
There is increasing demand for analysis of tumour samples to identify mutations that may
influence response to therapy. There is evidence to suggest that some tumours with BRCA
mutations have somatic inactivation only. A recent feasibility study showed that even in
poor quality breast and ovarian cancer samples, 90% coverage of BRCA1/BRCA2 genes was
achievable in 89% of samples, and this was confirmed by Sanger sequencing. This raises the
possibility of generating somatic sequencing data from tumour samples.
26
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Discussion
Ovarian cancer
There was discussion surrounding the exclusion of ovarian cancer from the NICE guidelines.
It was explained that this was the case because it fell outside the specific remit of the
guidelines.
Triple negative breast cancer
The issue of women with grade 3 triple negative breast cancer and the significance of family
history in considering genetic testing was raised. A recent unpublished study was cited
which looked at a large number of women with triple negative breast cancer and aimed to
question the validity of US NCCN (National Comprehensive Cancer Network) guidance that
recommends testing anyone under the age of 60 with triple negative breast cancer (without
information on family history). The study stated that below age 50 there is a detection rate
of 25%, and between 50-60 it was over 10%. However it was noted that the study
participants were tested for a panel of around 14 genes associated with breast cancer and
the pick up rate for those over 40 with no family history, only went above 10% as a result of
adding in these additional mutations.
It was felt that questions remain amongst the UK clinical community about when to seek
specialist advice from genetics colleagues in relation to grade 3 triple negative breast cancer
patients, and the reliability/availability of family history. It was felt that patients who cannot
provide a family history (for example, those who were adopted) should be considered for
testing providing they meet the criteria for testing.
Data sharing
Ownership and access to data was discussed, and the barriers which lack of sharing creates.
The 100,000 genomes project was highlighted in relation to NHS laboratories sharing data,
and contrasted with the commercial sector. It was suggested that lack of sharing may result
from the absence of a single consistent database. With regard to patients, it was noted that,
even with no consent issues, most are not aware of the underlying reasons for sharing data.
Exome sequencing
The utility of exome sequencing in relation to breast cancer testing was explored, although
it was felt unlikely to be implemented within the next five years due to test development
requirements and the costs of providing such a clinical diagnostic service for breast cancer.
27
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Testing costs
In relation to costs, it was suggested that BRCA mutation testing is unlikely to come down
below the £200-300 mark, due to factors such as laboratory personnel costs.
28
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Development of consensus referral and testing guidelines
The group identified the list of professionals considered suitable to make genetic testing
referrals for BRCA1/2. There was some discussion on the difference between who provides
consent and counselling, who is qualified to assess a patient’s suitability for testing and who
is responsible for post-testing counselling and management. It was felt that registered
genetic counsellors should be included alongside consultant clinical geneticists. With
regards to non-genetics specialties, it was felt that the list of referrers should be restricted
to consultant level professionals. Following discussions about suitability of referrals and
mainstreaming it was felt that the best approach was to include professionals in surgical
disciplines as referrers as a named multi-disciplinary team (MDT) clinician, under the
umbrella of the MDT which would allow for input from clinical genetics colleagues.
The group discussed the minimum criteria for recommending testing.
Women with breast cancer
There was agreement on the inclusion of several categories in this group due to clear
evidence that these individuals would be above the 10% threshold, and these were:

Women with invasive bilateral breast cancer under 40 years of age.
However, it was agreed that this should only apply to patients with ductal and not
lobular breast cancer. The difficulties of classifying tumours as multifocal versus
multicentric were noted with respect to this criterion. The group agreed that the
criterion should limit inclusion to women with invasive breast cancer as opposed to
ductal carcinoma in situ (DCIS), due to the rarity of the latter in BRCA mutation
carriers.

Women with grade 3 triple negative breast cancer diagnosed <40 years or <50 if
family history unclear or unknown.
The strong association between BRCA1/2 and grade 3 triple negative breast cancer
was reflected in the fact that testing is recommended in any patient with grade 3
triple negative breast cancer under 40 years of age. There was some discussion
surrounding the age limit in this testing criteria, but based on the earlier evidence
presented regarding pick up rate in different age groups, it was felt that the limit of
40 years of age would satisfy the 10% threshold. For patients where family history is
unknown it was felt that this should be extended to patients under 50 years of age.
29
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London

Women with breast cancer and non-mucinous epithelial ovarian cancer
This was considered to comfortably meet the 10% threshold, and therefore a
suitable testing criterion.

Women with breast cancer and a first-degree relative* both of whom were
diagnosed below 40 years of age.
This was considered to comfortably meet the 10% threshold, and therefore a
suitable testing criterion.

Women with breast cancer and a first-degree relative* with a histologically
confirmed non-mucinous epithelial ovarian cancer.
Within this group there will be groups with lower pick up rates, and there are
difficulties in confirming the histology of the ovarian tumour. Therefore it was felt
important that the criterion should include the phrase histologically confirmed nonmucinous epithelial ovarian cancer.

Women with breast cancer and a family history with a pathology adjusted
Manchester score greater than or equal to 15.
The Manchester scoring system was considered useful in discerning the likelihood of
mutation, and the utility of including pathology in the calculations was emphasised
with the inclusion of pathology adjusted Manchester score of greater than or equal
to 15.
*Or a second-degree relative via a father
Women with ovarian cancer

Women with ovarian cancer with a histology consistent with a high grade serous
epithelial carcinoma OR A family history with a pathology adjusted Manchester score
greater than or equal to 15
Ovarian cancer patients were included in the testing criteria, and there was
again discussion about the most appropriate age cut-off. It was agreed that,
aside from patients with a high grade serous histology, the issue of age at
diagnosis would be encompassed within the Manchester score.
30
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Men with BRCA-related cancer

Man with a BRCA-related (prostate, breast or pancreas) cancer who has a family
history with a Manchester score greater than or equal to 15
Affected individuals with Ashkenazi Jewish or Polish ancestry

Female breast cancer diagnosed <50 or a male BRCA-related cancer (founder
mutation screen)
These patients would usually undergo a founder mutation screen initially.
Unaffected individuals
It was felt important to state that referrals for testing of unaffected individuals should only
be made by consultant clinical geneticists or registered genetic counsellors.
The criteria for testing in this group were agreed as follows:

Unaffected individual who has a family history with a Manchester score greater
than or equal to 20 AND a first-degree relative with breast/ ovarian/ prostate/
pancreatic cancer where there are no affected relatives available for testing
(ovarian cancer and cancer in first-degree relative should be confirmed)

Unaffected individual with Ashkenazi Jewish/ Polish ancestry who has a first-degree
relative with female breast cancer diagnosed <50 or a male BRCA-related cancer
and a Manchester score greater than or equal to 10 (founder mutation screen$)
$
This should only be done if testing cannot be performed in an affected relative.
The group agreed to limit testing in this group to those with a Manchester score of
above 10. There was some discussion surrounding the testing procedure and the
possible need to screen for founder mutations, depending on the availability of
affected relatives, possibly followed by wider testing.
At risk family members where familial mutation is known.
Some groups were not considered suitable for inclusion in the testing criteria as the
evidence suggests that they would most likely fall below the 10% threshold based on testing
algorithm calculations.
The group recommended that the consensus testing criteria are adopted for the testing of
BRCA1/2 in the UK. The testing criteria are shown in Appendix 1.
31
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Authors
Louise Cameron, PHG Foundation
Shehla Mohammed, UKGTN
Mark Kroese, UKGTN and PHG Foundation
The PHG Foundation was commissioned by the UKGTN to write this
workshop report.
32
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Appendix 1
UKGTN Testing Criteria
Approved name and symbol of disorder/condition(s):
Hereditary Breast and Ovarian Cancer
OMIM number(s):
604370 & 612555
Approved name and symbol of gene(s):
BRCA1 and BRCA2
OMIM number(s):
113705 & 600185
Patient name:
Date of birth:
Patient postcode:
NHS number:
Name of referrer:
Title/Position:
Lab ID:
Referrals will only be accepted from one of the following:
Referrer
Tick if this refers to
you.
Consultant Clinical Geneticist/Registered Genetic Counsellor
OR named Multi-Disciplinary Team clinician:
Consultant Oncologist
Consultant Gynaeoncologist
Consultant Breast Surgeon
Minimum criteria required for testing to be appropriate as stated in the Gene Dossier:
Criteria
Tick if this patient
meets criteria
Woman with breast cancer who has ONE of the following:
1. Bilateral invasive ductal breast cancer and both cancers diagnosed <40
years
2. Grade 3 triple negative breast cancer diagnosed <40 years or <50 if family
history unclear or unknown
3. Non- mucinous epithelial ovarian cancer
4. A first-degree relative* with breast cancer and both diagnosed <40 years
5. A first-degree relative* with a histologically confirmed non-mucinous
epithelial ovarian cancer
6. A family history with a pathology adjusted Manchester score greater than or
equal to 15
OR Woman with ovarian cancer who has:
1. Histology consistent with a high grade serous epithelial carcinoma OR
2. A family history with a pathology adjusted Manchester score greater than or
equal to 15
33
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Appendix 1
OR Man with a BRCA-related (prostate, breast or pancreas) cancer who has:
1. A family history with a Manchester score greater than or equal to 15
OR Affected individual with Ashkenazi Jewish/ Polish ancestry who has:
1. Female breast cancer diagnosed <50 or a male BRCA-related cancer
(founder mutation screen)
OR Unaffected individuals:
Referrals only accepted from Consultant Clinical Geneticist or Registered
Genetic Counsellor
1. Unaffected individual who has a family history with a Manchester score
greater than or equal to 20 AND a first-degree relative with breast/ ovarian/
prostate/ pancreatic cancer where there are no affected relatives available
for testing (ovarian cancer and cancer in first-degree relative should be
confirmed)
2. Unaffected individual with Ashkenazi Jewish/ Polish ancestry who has a
first-degree relative with female breast cancer diagnosed <50 or a male
BRCA-related cancer and a Manchester score greater than or equal to 10
(founder mutation screen$)
* Or a second-degree relative via a father
$This should only be done if testing cannot be performed in an affected relative
OR At risk family members where familial mutation is known.
Additional Information:
If the sample does not fulfil the clinical criteria or you are not one of the specified types of referrer
and you still feel that testing should be performed please contact the laboratory to discuss testing of
the sample.
First-degree relative: parent, sibling, child
Second-degree relative: uncle, aunt, nephew, niece, grandparent, grandchild, half-sibling
34
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Appendix 2 Meeting programme
DEVELOPING UKGTN TESTING CRITERIA FOR FAMILIAL BREAST/OVARIAN CANCER :
INCORPORATING NICE GUIDELINES
15th July 2014
Chandos House, The Royal Society of Medicine, London
Chair: Fiona Lalloo
10.00
Registration and Coffee
10.30
Introduction and Welcome
Ros Skinner
10.40
Background and Scope of Workshop
Shehla Mohammed
10.45
Current practice:
10.45-11.00
Clinical Background
Gareth Evans
11.00-11.30
View and Impact of Recent NICE Guidelines
Julian Adlard
11.30-11.45
Review of Current Practice for Genetic Testing
Marc Tischkowitz
11.45-12.00
Perspective from the Oncology Clinic
Andrew Tutt
12.00-12.15
Perspective from the Pan London Approach
Adam Shaw
12.15 -1.00
LUNCH
1.00-1.15
Patient Perspective
Emma Pennery
1.15-1.30
Impact of Emerging Technology Platforms
Andrew Wallace
1.30 -3.00
Development of Recommendations on Consensus
Referral and testing guidelines
Shehla Mohammed
3.00-3.15
Closing Remarks
35
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Appendix 3 List of delegates
Name
Job Title
Organisation
Julian Adlard
Consultant in Cancer
Genetics
Yorkshire Regional Genetics
Service, Leeds
Louise Cameron
Project Manager (Science)
PHG Foundation
Treena Cranston
Principal Clinical Scientist
and Lead for Cancer and
Endocrine Services
Oxford Medical Genetics
Laboratories
Caroline Dalton
Genetics Policy Lead
Breakthrough Breast Cancer
Rosemarie Davidson
Consultant in Clinical
Genetics
Southern General Hospital,
Glasgow
Diana Eccles
Professor of Cancer Genetics
University of Southampton
Gareth Evans
Professor of Genetic
Medicine
Central Manchester University
Hospitals
Jacqui Hoyle
Knowledge &
Communications Manager
Consultant in Public Health
Medicine
and UKGTN Public Health
Advisor
UK Genetic Testing Network
Consultant Clinical
Geneticist, Clinical Director
Clinical Genetics Service and
Chair, Cancer Genetics
Group
Principal Genetic Counsellor
Manchester Centre for Genomic
Medicine
Mark Kroese
Fiona Lalloo (Chair)
Mark Longmuir
PHG Foundation
West of Scotland Genetic Service
Shehla Mohammed
Consultant Clinical Geneticist Guy's & St Thomas' NHS
and UKGTN Clinical Advisor
Foundation Trust
Alexandra J Murray
Consultant Clinical Geneticist University Hospital of Wales
Kai-ren Ong
Consultant in Clinical and
Cancer Genetics
Principal Clinical Scientist
Sheila Palmer-Smith
Birmingham Women's NHS
Foundation Trust
Institute of Medical Genetics
University Hospital of Wales
36
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London
Emma Pennery
Clinical Director
Breast Cancer Care
Rachel Robinson
Principal Scientist
Yorkshire Regional DNA Laboratory
Clinical Genetics Service
Adam Shaw
Consultant in Clinical
Genetics
Chair
Guy's & St Thomas' NHS
Foundation Trust
UK Genetic Testing Network
Head of Strategy and Public
Affairs
University Lecturer and
Honorary Consultant
Physician in Medical
Genetics
Ovarian Cancer Action
Sue Tomkins
Consultant in Clinical
Genetics
University Hospital, Bristol
Vishakha Tripathi
St George’s University of London
Andrew Wallace
Principal Genetic Counsellor
(Macmillan)
Honorary Senior Lecturer
Consultant Clinical
Oncologist
Consultant Clinical Scientist
Jacquie Westwood
Director
UK Genetic Testing Network
Jo Whittaker
Scientific Developments
Advisor
UK Genetic Testing Network
Rosalind Skinner
Katherine Taylor
Marc Tischkowitz
Andrew Tutt
Department of Medical
Genetics, University of Cambridge
Institute of Cancer Research,
London
Genomic Diagnostics Laboratory
Manchester Centre for Genomic
Medicine
37
UK Genetic Testing Network workshop. Developing testing criteria for familial breast and ovarian cancer: incorporating NICE guidelines
15 July 2014 | Chandos House, London