Main

Triple-negative (TN) breast cancer describes a subgroup of tumours that lack expression of oestrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor (HER2) (Foulkes et al, 2010). Overall, TN cancers account for about 15% of all breast cancers, but occur more frequently in younger women and are the predominant cancer subtype in individuals with a germline BRCA1 mutation (Bauer et al, 2007; Atchley et al, 2008; Blows et al, 2010; Foulkes et al, 2010).

The identification of a BRCA mutation has profound consequences for clinical management; impacting on the likelihood of developing contralateral breast cancer and/or ovarian cancer and increasingly having implications for optimal therapy (Antoniou et al, 2003; Fong et al, 2009; Tutt et al, 2010; Nathanson and Domchek, 2011). Due to financial and logistical constraints, BRCA testing is currently rationed in most countries. In the US and much of Europe, BRCA testing is typically undertaken if the likelihood of detecting a mutation is >10% (American Society of Clinical Oncology, 2003; Gadzicki et al, 2011). In the UK, the National Institute for Health and Clinical Excellence (NICE) recommended that testing should minimally be offered if the likelihood of detecting a mutation is >20%, though many UK centres also offer testing if the likelihood is between 10–20%, (McIntosh et al, 2004; NICE, 2006). Several different methods to determine which cases are eligible for testing are utilised in clinical practice, most of which require specialised knowledge and/or software (Antoniou et al, 2008).

The recognition of the strong association of the TN phenotype and BRCA1 mutations has led to efforts for establishing the frequency of BRCA1 mutations in individuals with TN breast cancer. To date, several small studies have evaluated this in both unselected series and case series selected on the basis of family history and/or age (Table 1). Additionally, it was recently proposed that BRCA testing all women with TN cancers diagnosed below 50 years would be cost-effective with respect to overall health spending at a national level (Kwon et al, 2010), based on an estimated mutation prevalence of 10–25%. However, this study did not address the practical and cost implications for the local services that undertake testing.

Table 1 Studies with over 50 cases that have evaluated BRCA1 mutation prevalence in TN cancers
Full size table

In this study, we have undertaken BRCA1 analysis in 308 individuals with TN breast cancer; the largest study to date. We have used the data to further evaluate the mutation frequency and to consider the practical ramifications of undertaking BRCA testing in individuals with TN breast cancer.

Materials and methods

Cases

We included 308 TN breast cancers from UK. Oestrogen receptor, PR and HER2 status were confirmed either in a histopathology report and/or a clinician's referral letter. When not explicitly stated, ER and PR status were scored as negative when there was absent expression (equivalent to a Quickscore of 0 out of 8). Human epidermal growth factor receptor was regarded as negative when scored as 0 or 1+ for HER2 by immunohistochemistry and/or when there was non-amplification of HER2 by fluorescent in situ hybridisation.

The cases were either from case series unselected with respect to genetic susceptibility (the unselected series n=159) or from case series that were specifically ascertained because of young age at diagnosis and/or a family history of breast cancer (the selected series n=149). The unselected series came from either the ongoing TNT trial ISRCTN97330959 (Kilburn, 2008), a UK-wide randomised phase III trial of carboplatin compared with docetaxel for patients with metastatic or recurrent locally advanced TN breast cancer (n=81); or the Marsden sample series, which was a collection of samples from breast cancer patients attending the oncology clinics at the Marsden Hospital (n=78). The selected series came from either the Familial Breast Cancer Study (n=90), from referrals to our regional genetics department (n=25) or from the Prospective study of Outcomes in Sporadic versus Hereditary breast cancer (POSH, n=34). The latter was a UK-wide study that recruited individuals with invasive breast cancer aged ⩽40 years (Eccles et al, 2007). None of the cases have been included in any other published study on TN breast cancer. The study was undertaken as part of our research into the genetic causes of breast cancer, which has been approved by the London Multicentre Research Ethics Committee (MREC/01/2/18).

BRCA1 analysis

BRCA1 mutation analysis, including multiplex ligation-dependent probe amplification (MLPA) analysis for large deletions/duplications, was performed in DNA from all cases. This was either performed through a clinical BRCA test by the local centre, or was undertaken by ourselves by sequencing genomic DNA through the 24 coding exons and intron–exon boundaries of BRCA1 and undertaking MLPA using probe mix P002. All mutations were confirmed by separate bi-directional sequencing in a second sample. All copy number changes were confirmed in a fresh aliquot of DNA with a different probe mix (P087). The mutation nomenclature is in accordance with HGVS convention with numbering starting at the first A of the ATG initiation site, using U14680.1 as the reference sequence.

Assessment of eligibility for clinical BRCA testing

Currently, in UK there is variability in the threshold used for BRCA testing; in the South West Thames Regional Genetics and Royal Marsden Cancer Genetics services we use a threshold of 10% and primarily use the Manchester score to assess this (Evans et al, 2005). Family history information was available for 271 individuals. We calculated the Manchester score, and designated those with a score ⩾15, as eligible for clinical testing (Supplementary Tables 1 and 2). Of note, this classification does not necessarily mean that these cases actually had clinical BRCA testing; for many, this information was not known and in some cases we were aware that testing had not occurred. The classification denotes whether the patient would have been eligible for testing by genetics departments operating a 10% mutation detection threshold.

Age-specific TN breast cancer incidence

National breast cancer figures in England are not subclassified by receptor status. Therefore, to estimate the annual age-specific incidence of TN breast cancer in England we used the national figures for age-specific breast cancer incidence published by the Office for National Statistics for England (Office for National Statistics, 2006) and data from a study of 10 159 cases of breast cancer subtyped by immunohistochemistry (Blows et al, 2010). This allowed us to estimate the annual age-specific incidence of TN breast cancer in England.

Results

BRCA1 mutation frequency

The full results of all 308 cases are given in Supplementary Table 1. Overall, there were 45 BRCA1 mutations in the 308 individuals (14.6%). This included 15 (9.4%) BRCA1 mutations in 159 individuals in the unselected series, and 30 (20.1%) BRCA1 mutations in 149 individuals in the selected series (Table 2). There was a strong age-effect with marked decrease in mutation frequency in individuals aged over 50 years in both the unselected and selected series (Table 2). If one considers just individuals with sporadic TN breast cancer, that is, those without a first or second degree relative with breast or ovarian cancer, 8 out of 103 (8%) had a BRCA1 mutation, and all were under 50 years of age.

Table 2 Summary of BRCA1 mutations in 308 TN breast cancer cases
Full size table

Eligibility for clinical BRCA testing

Family history data was available in 271 individuals, which included 122 out of 159 individuals in the unselected series and all individuals in the selected series. We used these data to calculate the Manchester score for each family to evaluate whether the individual would be eligible for clinical BRCA testing (Manchester score ⩾15, Supplementary Table 2). Overall, approximately a third of cases (85 out of 271; 31%) were eligible for clinical BRCA gene testing, though as expected this was lower in the unselected series (20 out of 122; 16%), particularly at older ages (Supplementary Table 1). If one considers the individuals with BRCA1 mutations for whom we could calculate the Manchester score, 15 out of 42 (36%) were not eligible for clinical BRCA testing (Table 3).

Table 3 Eligibility for clinical BRCA testing by age and mutation status
Full size table

Impact of using age-specific testing threshold on BRCA testing

To estimate the number of BRCA tests that would be undertaken in TN cases, if age-specific criteria were employed, we used the cancer registration statistics from the Office for National Statistics for England (Office for National Statistics, 2006), together with the estimates of the proportion of breast cancers that are TN by age (Blows et al, 2010). These suggest that ∼6000 women with TN breast cancer are diagnosed each year in England, of which ∼500 are <40 years of age and ∼1500 are <50 years of age (Table 4). If one assumes that the data from the unselected series are broadly representative of the proportion of these cases that would not currently be eligible for clinical BRCA testing, 69% (11 out of 16) of TN <40 years, 77% (33 out of 43) <50 years, 79% (72 out of 91) <60 years and (84%) (102 out of 122) of all TN cases would not be eligible for clinical BRCA testing (percentages reflect the number of individuals not eligible for testing/total number of individuals in that age group for whom eligibility status was known, see Supplementary Table 1). If all individuals with TN breast cancer diagnosed before 50 years of age were tested, these data suggest that an additional ∼1200 BRCA tests would be performed per year in England.

Table 4 Age-specific incidence of TN breast cancer and impact on BRCA testing
Full size table

Discussion

We have undertaken the largest analysis of BRCA1 in TN breast cancer to date, and showed that the frequency of BRCA1 mutations in unselected individuals with TN breast cancer is ∼10%, increasing to ∼19% of individuals diagnosed below 50 years. The latter is similar to the frequency of BRCA1 mutations (23%) in individuals diagnosed before 50 years that were selected for inclusion because of a family history of breast cancer and/or young age at diagnosis. Our study included samples from four different sources and more precise figures would be obtainable from larger, prospective studies. Nevertheless, our results are similar to those previously reported and we believe that they are likely to be broadly accurate (refs in Table 1).

We estimated the proportion of individuals included in this study that would currently qualify for a clinical BRCA test. Our analysis suggests that over a third of the BRCA1 mutation-positive individuals we identified would not have been eligible for clinical genetic testing in departments that use a 10% mutation detection threshold calculated without consideration of histological parameters.

Taken together, these data strongly indicate that histological parameters should be included in deciding which individuals should be offered BRCA testing. There have been efforts to incorporate histological parameters into existing BRCA-testing selection systems, such as BOADICEA and the Manchester system, (Evans et al, 2009; Mavaddat et al, 2010). Additionally, studies to define testing criteria in individuals that are not eligible for testing using current systems (e.g., individuals with TN breast cancer but without a family history) have been undertaken (Young et al, 2009; Evans et al, 2011). We believe a drawback to these approaches is that complex evaluation by specialist practitioners is typically required for most cases, but is unnecessary in the sizeable proportion eligible for testing on the basis of age alone.

A simpler approach, which would be readily comprehensible by clinicians and patients, would be to define a BRCA testing eligibility criteria for women with TN breast cancer based on age. Only individuals above the age threshold would require the more detailed, specialist review incorporating family history and scoring algorithms to decide whether they were eligible for BRCA testing.

Our data suggests that diagnosis of TN cancer below 50 years would be a suitable age threshold for BRCA testing. It is also consistent with recent simulation data, suggesting that this testing threshold would be a cost-effective strategy and would result in substantial reduction in subsequent breast and ovarian cancer in mutation-positive women (Kwon et al, 2010). However, implementation of routine BRCA testing in all TN breast cancers diagnosed before 50 years would increase the logistical and financial burdens on genetic departments. We estimate that it would lead to ∼1200 extra tests in England each year, which may be challenging for some departments to immediately implement with current resources and procedures. However, new sequencing technologies are leading us into an era of fast, affordable gene testing. Together with procedural reorganisation to allow BRCA testing in affected individuals to be undertaken though oncology services (with support from genetics as required), this should enable genetic services to introduce BRCA testing to women with TN breast cancer diagnosed below 50 years of age, within the next few years.