Prof. Gareth Evans, Manchester Academic Health Science Centre, Genetic Medicine, St Mary's Hospital, Central Manchester Hospitals Foundation Trust, Manchester M13 9WL, UK. Tel.: +44 161 276 6206; fax: +44 161 276 6145; e-mail: email@example.com
Since the localization and discovery of the first high-risk breast cancer (BC) genes in 1990, there has been a substantial progress in unravelling its familial component. Increasing numbers of women at risk of BC are coming forward requesting advice on their risk and what they can do about it. Three groups of genetic predisposition alleles have so far been identified with high-risk genes conferring 40–85% lifetime risk including BRCA1, BRCA2 and TP53. Moderate risk genes (20–40% risk) including PALB1, BRIP, ATM and CHEK2, and a host of low-risk common alleles identified largely through genome-wide association studies. Currently, only BRCA1, BRCA2 and TP53 are used in clinical practice on a wide scale, although testing of up to 50–100 gene loci may be possible in the future utilizing next-generation technology.
The authors declare that there are no conflicts of interest.
Breast cancer (BC) is the most common form of cancer affecting women. One in 9–12 women will develop BC in their lifetime in the developed world. There are a number of recognized risk factors for BC development including hormonal, both endogenous and exogenous, reproductive and obesity. However, the strongest factor is family history (FH). Individual risk increases with increasing number of relatives affected with BC and the decreasing age at which it was diagnosed.
Whilst twin studies estimate that around 27% of BC is because of hereditary factors (1), only 5–10% of BC has a strong inherited component with only 4–5% being due to high penetrance genes transmitted in an autosomal dominant fashion (2–5). Rates of mutation in the best known high penetrance genes BRCA1 and BRCA2 vary across populations because of founder effects (6). A further proportion is caused by a number of moderate penetrance genes (7–10). At present, around 20 low penetrance genes have been identified (11) cumulatively; when all have been discovered, these may contribute a higher proportion of familial BC. These genes appear to interact, although this has yet to be completely elucidated (12).
Individuals with an FH in the UK (and most of the developed world) are referred to FH clinics for an assessment of their BC risk. The FH clinics aim to provide a network of services from primary, through secondary and into tertiary care. The referral pathways depend upon the level of assessed risk according to the NICE guidelines for familial BC (13). The subsequent management options available to an individual woman including screening or prevention are then dependent on this level of assessed risk.
Genetics of BC
High penetrance genes
In the last 17 years, advances in the genetics of BC have resulted in the cloning of BRCA1 and BRCA2(2, 3). Whilst pathogenic mutations in these genes have high penetrance, they occur relatively rarely, with a combined frequency of about 0.4% (6, 14). Pathogenic mutations in BRCA1/BRCA2 are known to increase the risk of BC by 10- to 20-fold. Mutations in TP53 also give a high risk of BC, although the frequency of germline mutations is even lower.
In 1990, following international collaborative studies, BRCA1 was linked to chromosome 17q and was subsequently cloned in 1994 (2). About 1 in 500–1000 individuals carry a pathogenic mutation in BRCA1(15) (outside founder populations), which accounts for about 7–10% of familial BC (Fig. 1) (16).
Pathogenic mutations in BRCA1 confer a lifetime risk of BC between 60% and 85%, with increased relative risks (RR) at younger ages (17–20). For example, the RR of BC between 30 and 39 years is 33, but decreases to 14 between 60 and 69 years (19).
Pathogenic mutations also confer an increased lifetime risk of ovarian cancer of 40–60% (17–20). Ovarian risk is not as age-dependant. Women may also have increased risk of pancreatic malignancy (21). Men with BRCA1 mutations have an RR of cancer of 0.95% (21). The age at which BRCA1 mutation carriers are affected with both breast and ovarian cancer is substantially younger than the general population, approximately 3% risk of BC by 30 years (17).
BRCA1 is a large gene with 24 exons: the largest exon 11. Mutations are found throughout the coding sequence of the gene, with the majority being frameshift mutations resulting in truncated proteins. Missense mutations account for approximately 2% of pathogenic mutations in BRCA1, but may be difficult to interpret or distinguish from polymorphisms. Between 15% and 27% of mutations may be due to large rearrangements, including large deletions (whole exon) and insertion/duplications (22). There does not appear to be any useful genotype–phenotype correlation clinically, although mutations in the 5′ portion of the gene have been associated with a higher risk of ovarian cancer (21).
Although hundreds of unique pathogenic BRCA1 mutations have been described, within certain populations particular mutations are more common (founder mutations). For example, within the Ashkenazi Jewish population, two mutations, c.68_69delAG and c.5266dupC (previously 185delAG/5382insC) occur in about 1.2% of the population. The c.5266dupC mutation is also found in other eastern European populations particularly in Poland (23). The identification of these founder mutations facilitates mutation population screening.
The major role of BRCA1 appears to be DNA repair including homologous recombination and nucleotide excision repair (24). However, it also has a function in the regulation of cell-cycle progression in particular checkpoint control (24).
Clinical features of BRCA1
BCs in women with BRCA1 mutations often exhibit different pathology to those of BRCA2 mutation carriers and non-familial BCs. These cancers have been noted to have an increased frequency of pushing margins, high degree of nuclear pleomorphism and mitotic frequency, suggestive of medullary carcinoma (25). Indeed atypical medullary BCs have been observed more frequently with BRCA1 mutations approximately 13% vs 3% in sporadic BCs. Breast malignancies in BRCA1 mutation carriers are also more likely to be steroid receptor and Her2-negative than sporadic cancers. Ductal carcinoma in situ (DCIS) is noted less frequently in BRCA1 mutation carriers. It has recently been noted (25) that BRCA1 cancers have a similar immunohistological profile to sporadic basal carcinomas (carcinomas expressing molecules normally seen in the basal/myoepithelial cells of the normal breast) and include positivity for CK5/6+/CK14+.
The prognosis of BRCA1 associated BCs has been reported as both better and worse than sporadic tumours, although the association with ER tumours and basal carcinomas would support the hypothesis of worse prognosis.
In addition to a high risk of a primary BC, women with BRCA1 mutations also have an increased contralateral risk with cumulative risk of 64% by 70 years (20).
Ovarian tumours associated with BRCA1 mutations are usually high-grade serous epithelial carcinomas. Endometrioid and less frequently clear cell carcinomas have been reported, but mucinous and borderline tumours are not seen (26). Two granulosa cell tumours in BRCA1 mutation carriers have occurred (26). Primary peritoneal malignancies are also frequent.
BRCA2 mutations account for about 10% of families with breast and ovarian cancer with between 1 in 600 and 1 in 800 women having a pathogenic mutation in outbred populations (15, 27). Mutations in this gene confer a BC lifetime risk of around 40–85% (15–20). The range of risk is much higher in BRCA2 and the method of ascertainment from high-risk families or population studies clearly has an effect. This is demonstrated by the common Ashkenazi Jewish population mutation (6174delT) c.5946delT, which has much lower penetrance with some studies suggesting a lifetime risk of BC of approximately 30–40%. However, the literature is biased by ascertainment, and as such risks appropriate to individual families should be quoted (17). This should be based on closeness of BC FH, other known risk factors and possibly even assessment of common genetic variants (28). The BOADICEA model takes into account FH in assessing BC risk in BRCA2 carriers within a wide range (14).
The ovarian risk associated with pathogenic mutations is up to 30% (17). There is more variability of risks associated with mutations in BRCA2, which suggests that this is a more modifiable gene. The RR of cholangiocarcinoma, melanoma, pancreatic (RR 4.1, 95% CI 1.9–7.8) and gastric cancers (RR 2.7, 95% CI 1.3–4.8) are also increased (29, 30).
Male carriers of BRCA2 mutations have an increased risk of prostate cancer with lifetime risks of 14–20% along with an increased risk of male BC. The RR of male BC associated with a BRCA2 mutation is 80- to 100-fold with about 10% of male BCs being due to mutations in this gene (31) and 8–10% developing BC in their lifetime (32).
As with BRCA1, BRCA2 is a large gene with 27 exons encoding a 3418 amino acid protein, with exon 11 being the largest. Mutations occur throughout the gene, again the majority being frameshifts. There are a large number of missense mutations found within BRCA2, but the pathogenicity of these may be difficult to establish. Large gene rearrangements also occur in BRCA2, but are less frequent than BRCA1, only 19/336 (6%) families in our service. An area within exon 11 called the ovarian cluster regions (OCR) flanked by nucleotides 3035–6629. Within the OCR, there is higher reported risk if ovarian cancer (33), although potentially a sampling anomaly as other investigators have not found an effect (17).
BRCA2 is known to be involved in DNA repair. It facilitates homologous recombination and is involved with double-strand break repair (24). It interacts directly with RAD51 forming a complex and holding it in an inactive state. Cells that lack BRCA1/BRCA2 are hypersensitive to DNA-damaging agents with resulting double-stranded breaks. These are then repaired by error-prone mechanisms such as non-homologous end joining, resulting in chromosomal rearrangements and instability. This chromosomal instability is a crucial feature of carcinogenesis.
Biallelic mutations in BRCA2 have been shown to cause Fanconi anaemia (FANCD1), a condition causing developmental anomalies including short stature, microcephaly and radial ray abnormalities as well as predisposing to childhood solid tumours and haematological malignancies (34).
Clinical features of BRCA2
Specific BC pathology is not as characteristic with BRCA2 mutations as it is with BRCA1. The tumours appear to have less tubule differentiation and both increased and decreased mitotic rates compared with sporadic tumours have been reported. Lobular carcinoma has been reported more commonly with BRCA2 associated tumours than with BRCA1 associated tumours. DCIS is also more common in BRCA2.
These tumours are more frequently ER+ than controls, although higher grade. Overall, BRCA2 tumours tend to have similar features to sporadic BCs unlike BRCA1(35). Prognosis of BRCA2 related BC is similar to population BCs.
Ovarian carcinomas associated with BRCA2 have similar features to those associated with BRCA1 mutations (36). Borderline and mucinous tumours are not part of the clinical picture. The prognosis of BRCA2 associated ovarian cancers is better than the general population, probably due to a better response to platinum-based therapies (37).
TP53 was first identified in 1979 and now is known be the most frequently altered gene in human tumours. Somatic mutations in the TP53 gene are common in solid tumours. Inherited germline mutations are rare, but are known to result in Li-Fraumeni syndrome (LFS). LFS causes childhood tumours (typically soft tissue and osteosarcomas, gliomas and adrenocortical carcinoma) and very early onset BC (30% of female gene carriers have developed BC by 30 years of age). Over 70% of classical LFS families have inherited TP53 mutations. There is good in vitro evidence to suggest that patients with LFS have an abnormal response to low dose radiation with defective apoptosis. Recognition of this syndrome is important as these women should avoid radiotherapy if possible due to an increased risk of second primary malignancies.
LFS only accounts for <0.1% of BC, but mutations in TP53 confer an 18- to 60-fold increased risk of BC <45 years compared to the general population.
TP53 consists of 11 exons, with the core DNA binding domain being encoded by exons 4–8. TP53 is essential in cell-cycle control, resulting in either a delay in cell-cycle progression or apoptosis.
Other potential high-risk genes
A number of other rare syndromes have been associated with quoted high risk of 40–60% for BC. Mutations in PTEN that cause Cowden syndrome, STK11 that causes Peutz Jeghers syndrome and E-Cadherin (CDH1) that causes hereditary diffuse gastric cancer have all been associated with high risk of BC. There are, however, no comprehensive studies allowing for ascertainment bias that provide reliable risk estimates for BC in these conditions.
Moderate penetrance genes
There are four genes in which mutations have recently been identified associated with an RR of BC of two- to fourfold. These are rare genes with a population frequency of <0.6%. The phenotypes associated with these mutations have not been clearly delineated and therefore, the clinical utility of these genotypes has yet to be established.
Ataxia telangiectasia is an autosomal recessive condition due to homozygous mutations in ATM. Clinically this condition results in progressive cerebellar ataxia and oculomotor apraxia, conjunctival telangiectasia, immunodeficiency and increased risk of malignancy including BC.
It had been suggested for several years that ATM heterozygotes have increased BC risk (38), although this has been controversial. However, recent studies (39) have confirmed an increased RR of BC of 2.23 (95% CI 1.16–4.28) in ATM heterozygotes. This RR increases <50 years. Mutations described in ATM include truncating mutations, splice site and missense mutations. ATM is a protein kinase involved in the response to double-stranded DNA breaks in a pathway that includes TP53, BRCA1 and CHEK2.
It is difficult to assess the clinical utility of genetic testing for ATM at present. The penetrance of the gene is approximately 15% and estimating which mutation carriers will develop BC is not possible. However, these women may merit different approaches to treatment of BC due to the increased radiosensitivity associated with ATM mutations.
The checkpoint kinase gene CHEK2 encodes a protein that is a signalling component in the cellular response to DNA damage. It is involved in the same pathway as TP53 and BRCA1. CHEK2 is a tumour suppressor gene and somatic mutations have been identified in a number of malignancies. A particular germline mutation 1100delC has been shown to give an RR of BC of 2.34 (95% CI 1.72–3.2) (40). It is present in 0.2–1% of European populations and 4.2% of BC families, although the mutation frequency varies between populations. A number of other CHEK2 mutations have been reported in BC families, but the clinical significance of these is unclear.
Carriers of 1100delC mutation have an increased risk of bilateral BC. Originally it was suggested that it may also contribute to male BC, but this has not been verified. There does not appear to be an increased risk of other malignancies with heterozygous CHEK2 mutations.
A recent publication (41) has described families with homozygous CHEK2* 1100delC mutations. Women homozygous for the mutation have a much higher risk of BC – estimated to be sixfold. There also appears to be an increased risk of other malignancies within these families including colorectal cancer, although clearly further work needs to be undertaken.
BRIP1 encodes for a protein that was identified as a binding partner of BRCA1 and was therefore investigated as a BC predisposing gene. In 2006, truncating mutations were identified in BC families (8). Segregation analysis assessed an RR of BC of 2.0 (95% CI 1.2–3.2), although there are reports of higher risks in some families. There have been some suggestions that mutations in BRIP1 may also confer an increased ovarian cancer risk (42). Biallelic mutations of BRIP1 cause Fanconi anaemia complementation group J (FANC-J). The phenotype in FANC-J is different to that of biallelic mutations in BRCA2 and results in a much lower rate of childhood solid tumours.
PALB2 (partner and localizer of BRCA2) encodes for a protein that interacts with BRCA2 during homologous recombination and double-strand break repair. Mutations in this gene were identified in BC families negative for mutations in BRCA1/2. The RR of BC associated with PALB2 mutations is approximately 2.3 (95% CI 1.4–3.9) (9). A Finnish Founder mutation PALB2 is thought to result in a slightly higher RR of BC.
Biallelic PALB2 mutations have been shown to cause FANC-N. This is similar to that caused by biallelic BRCA2 mutations. As with BRCA2, heterozygotes PALB2 mutations have been associated with an increased risk of pancreatic cancer (43).
A summary of the high and moderate risk genes can be found in Table 1.
Table 1. High and moderate risk hereditary conditions predisposing to breast cancer
A number of common alleles have now been identified to be associated with a slightly increased or decreased risk of BC and that these may work in a polygenic multiplicative model to account for the remainder of familial BC. Some of the single nucleotide polymorphisms (SNPs) identified are genes that have also been investigated as modifiers of BRCA1 and BRCA2. At the time of writing, 19 validated SNPs have been identified (Table 2 in Ref. 11). Antoniou and others (29, 44–46) have determined nine of the common BC SNPs (TOX3, FGFR2, MAP3K, LSP1, 2q35, SLC4A7, 1p11.2, 5p12 and 6q25.1) that confer increased risks for BC in BRCA2 mutation carriers (44, 46). Conversely, only TOX3, 2q35 and 6q25.1 polymorphisms showed increased risk for BRCA1 mutation carriers out of the genetic variants examined. This may reflect that most SNPs discovered thus far only increase the risk of estrogen receptor (ER) + BC. One recent study suggested that the use of just five SNPs in BRCA2 mutation carriers varied the lifetime risk of BC from 45% to 95% (28). A summary of low-risk alleles can be found in Table 2.
Table 2. Validated common low-risk susceptibility alleles identified through genome-wide association studies
Recently, mutations in two genes in the RAD51 group: RAD51C and RAD51D have been identified in breast/ovarian cancer kindreds, but not BC only families (47, 48). The initial report on RAD51C suggested that this was a high-risk gene for both breast and ovarian cancer, but more detailed analysis on both genes (48, 49) suggest that these are predominantly ovarian cancer susceptibility genes and the risk of BC is not clearly elevated. The risk associated with neurofibromatosis1 may also be moderately increased (50). The next phase of an estimated 20 extra SNPs has started with three more, published in 2012 (51).
In 2004, NICE (National institute for clinical excellence) published guidelines for referral and management of familial BC, which were subsequently updated in 2006 (13). These guidelines manage the referral pathway for women from primary to secondary through to tertiary care. The aim of the guidelines is for women to be stratified according to average, moderate and high risk of BC, with only those at high risk with a high probability of mutations in BRCA1/2 being referred to the regional genetic services. Women at moderate risk should be assessed and managed in secondary care, ideally in association with breast units. Similar criteria exist in other European and North American countries.
Broadly speaking a woman's risk of BC increases with increasing number of relatives with associated cancers and decreasing age at which those relatives were diagnosed.
Important factors within the FH include
1Young age at onset of the disease.
3Multiple cases on one side of the family.
4Association of BC with other malignancies such as ovarian cancer or early onset prostate cancer in a male relative or early onset sarcoma.
5Number and age of unaffected females.
The most important tool for risk assessment is an accurate three generation pedigree. A paternal history is as important as a maternal history, especially considering that men are more likely to be non-penetrant gene carriers.
In a minority of families, the disease is clearly inherited in an autosomal dominant fashion. Risk assessment then becomes straight forward as it depends on the prior probability of inheriting the mutation and the penetrance of the gene. However, in the absence of a dominant FH, risk estimation is based on large epidemiological studies. These demonstrate a 1.5- to -3-fold increased risk with an FH of a single affected relative.
There are different ways of utilizing these models, either using them manually to estimate risk or using computer programs that utilize epidemiological data. Some of the computer programmes also include the likelihood of detecting a BRCA1/2 mutation within a given pedigree (60).
The models in wide use for risk estimation include the Claus model (3), Gail model (61), BRCAPRO (62) and the Tyrer-Cuzick model (63). The Claus model is used mainly in the UK for manual risk estimation, whereas the other three are computerized.
Currently, the only comparison of these models was carried out by Amir et al. (64). They assessed the accuracy of the different risk estimation models using data from 1933 women with an FH of BC in a screening programme. Fifty-two of these women developed a malignancy, which was detected during the screening programme. All models were applied to this group of women and the Tyrer-Cuzick model was the most consistently accurate in predicting BC risk. The other models significantly underestimated the risks. However, the Claus model can be modified by altering the risks according to hormonal factors (using the manual model of risk estimation).
A further model is now in use called BOADICEA, although this has yet to be validated in the familial clinic setting (14).
High-risk families should be referred to a regional genetics service to discuss the likelihood of developing a genetic test within a family. Predictive genetic testing (testing of unaffected at risk individuals) is only possible if a mutation has been identified within that family, usually using a sample of DNA from a person with a malignancy. In some countries, testing of unaffected relatives without first ascertaining a mutation in an affected family member is commonplace, but only has clinical utility in families with a very high a priori probability of a BRCA1/2 mutation (usually meaning multiple cases of breast and ovarian cancer). As such in most situations, a negative test should be considered ‘uninformative’. The exception to this is if the family is from a population with a high frequency of specific founder mutation(s) such as the Ashkenazi Jewish population. In this situation, a negative mutation screen is more useful than usual.
The NICE guidelines state that mutation screening should be offered in families if there is ≥20% probability of detecting a BRCA1/2 mutation, whereas in most other Western countries a 10% threshold is usual. There are several methods of determining the likelihood of detecting a BRCA1/2 mutation within a given family. These include computer models including BOADICEA and BRCAPRO. The computer models require inputting of data into the computer and may take around 5–10 minutes per family. Myriad provide prevalence tables using family histories and data obtained from their clinical testing service. The Manchester scoring system is a tabulated scoring system that can be used easily in 1–2 min (65). Several studies have assessed the best predicting model in various populations (50), with conflicting results. Both the BOADICEA and Manchester Score incorporate information about pancreatic and prostate cancers into their systems unlike the other models. All these models could be improved by incorporating tumour pathology information (66). The new wave of next generation sequencing, which will cut both cost and time of testing, is likely to loosen the testing criteria.
Once a mutation is known within a family, predictive genetic testing becomes available. This then allows the identification of mutation carriers with the potential for targeted screening and intervention. Patients undergoing predictive testing are seen in the regional genetics service at least twice so that they are fully informed about the cancer risk associated with mutations and the implications to themselves and the wider family. Some individuals feel that psychologically they are unable to cope with the information that they are at high risk of a malignancy and choose to remain at 50% risk availing themselves of screening and surgical options.
Predictive genetic testing for BRCA1/2 mutations is only offered to individuals over the age of consent.
The NICE guidelines clearly delineate the management of women at increased risk into:
1moderate risk – lifetime risk of 1 in 6 to 1 in 4 or 10 year risk between 40 and 49 years of 3–7.99
2high risk – lifetime risk of ≥30% or 10 year risk of ≥8%.
Moderate risk women should be managed in secondary care and high risk in tertiary care.
Currently, the most Western countries offer 2- to 3-yearly mammography for women in the general population between the ages of 50 and 73 years. Recently, both the US and Canada withdrew screening aged 40–49 years in the general population. Women with an increased risk of double the population risk or higher are eligible for screening on an annual basis in many countries from the age of 40 years.
Mammography screening has come under great scrutiny in the last 2–3 years because falling BC mortality rates have been more attributed to improvements in treatment than mammography. Furthermore, overdiagnosis with cancers that may never present clinically may outweigh much of the benefit of screening. However, this is less of a problem when targeted at a higher risk population and mammography screening from 40 to 49 years has been shown in a large multicentre study to be effective in the familial setting (67) and may be effective at younger ages (68).
Magnetic resonance imaging
There have been a number of trials assessing the utility of magnetic resonance imaging (MRI) screening for women at increased risk of BC. In women with BRCA1/2 mutations, mammography only detects about 40–50% of lesions due to a variety of factors including increased density of breast tissue in young women. MRI has greater sensitivity and, overall, the studies demonstrated that a combination of MRI and mammography detects 70–100% of malignancies in high-risk women (69–71). However, MRI does have limited sensitivity detecting DCIS, which may be an issue with BRCA2 mutation carriers. Availability of MRI varies across countries, but is generally available to BRCA1/2 mutation carriers and women at very high risk of BC from 30 to 50 years of age with some starting earlier at 25 years and finishing after 50 years.
Whilst women who carry a BRCA1/2 mutation also have an increased risk of ovarian cancer, screening for ovarian cancer is not effective (72). All patients are therefore advised to undergo bilateral salpingo-oophorectomy (BSO) at the appropriate age.
One of the options for women at high risk of BC is to consider risk reducing mastectomy as a prevention of BC. There is good evidence to suggest that this will give a risk reduction of 90% (73, 74), although prospective very long term follow-up of women with mutations undergoing surgery is not yet possible.
None of the different surgical procedures will completely remove all breast tissue, and there will therefore always be a small residual risk of breast malignancy. The prime aim of surgery is to remove breast tissue with cosmesis as a secondary aim. Both of these issues need to be discussed with individual women prior to surgery, along with issues surrounding general surgical and anaesthetic risks.
Whilst there have been few long-term studies on the psychological effects of surgery, most studies suggest significant benefit to women who choose this option compared to those that don't in terms of anxiety and cancer related worry (75). It is important that protocols including psychological support are in place for any women considering surgery. The Manchester protocol (76) includes two sessions with a geneticist to discuss issues around genetic testing and risk, a session with a psychiatrist/psychologist to discuss body image and then sessions with the surgeons to discuss the different surgical options. The aim of the protocol is to ensure that patients are fully informed and as prepared as possible for surgery. Uptake rates for Risk reducing mastectomy vary enormously and are dependant on country of origin (77) as well as being age, risk and time dependant (78).
The management of choice for ovarian risk in women with BRCA1/2 mutations is risk reducing oophorectomy, once they have completed their family. The tissues at risk include the ovaries and the fallopian tubes and therefore patients should be offered BSO. This reduces the risk of ovarian cancer by 80–90% (79, 80) as well as decreasing the risk of BC by ∼50% if performed in <40–45 years old (81) women.
Undergoing BSO at a young age will put women into the menopause at an early age. The early data regarding BSO in BRCA1 mutation carriers suggest that the protection against BC afforded by early surgery is irrespective of hormone replacement therapy (HRT) usage or the type used (79, 80). However, general population data suggest the risk of BC is lowest with oestrogen only HRT which can only be used following hysterectomy. Therefore, any discussion of risk reducing oophorectomy should include the risks and benefits of surgery, HRT and whether to include a hysterectomy. Women should be advised that at present the benefits of HRT prior to 50 years of age outweigh the risks in terms of reduction in endocrine symptoms, and osteoporosis (82) and the risk of heart disease.
Other modifiable risk factors
A number of other risk factors for BC are being further validated. Obesity diet and exercise are probably interlinked (83, 84). The evidence for specific diets is lacking.
Identification of a group of women at high risk provides the possibility of obtaining sufficient events (development of BC) to make prevention trials worthwhile. Four major trials of prevention with tamoxifen have now been published (85–87). Tamoxifen had already been shown to reduce the risk of contralateral BC in affected women and the large American NSABP trial was the first to show a reduction in risk of BC in asymptomatic women (at increased risk) by 40–50% (85). Tamoxifen is by and large well tolerated, although hot flushes and other menopausal symptoms are common and there are increased risks of thromboembolic events and endometrial cancer (85–87). The IBIS1 study showed a 30–40% reduction in BC risk, but a rise in all cause mortality (86). As a result, tamoxifen is not currently licensed for prevention in the UK and Europe, but does have a license in North America. Nevertheless, longer term follow saw a sustained reduction in BC risk to the 10-year point, but a drop in the adverse effects to normal after stopping 5 years treatment (87). A study comparing tamoxifen with raloxifene in America, the STAR trial showed no overall difference in prevention between the two drugs in early follow-up (88), but extended follow-up revealed an advantage for tamoxifen with a higher risk of BC with raloxifene RR 1.26. This was balanced by a higher rate of endometrial cancer and thromboembolic disease with tamoxifen (89). Finally, a recent trial has shown a 60% reduction of BC risk with the aromatase inhibitor exemestane in post-menopausal women (90). Thus, tamoxifen has now sufficient evidence to offer to pre-menopausal women at moderate or higher BC risk whilst women can be offered a choice of three drugs post-menopausally with advice on which may suit them best balancing the risks and benefits and tailoring to their personal BC risk and situation. Nevertheless, uptake of tamoxifen even when offered remains low with only 10–15% in Manchester starting treatment.
Recently, the utility of genetic testing for BC predisposition to the oncologist treating the cancer in the index case has come sharply into focus with the advent of therapies such as the poly (ADP-ribose) polymerase (PARP) inhibitors (91). A number of phase II clinical trials are underway to investigate the therapeutic effects of PARP inhibitors in the treatment of cancer in BRCA1 and BRCA2 mutation carriers. PARP-1 is an enzyme that repairs single-strand DNA breaks by base-excision repair. Inhibition of PARP-1 leads to the formation of double-strand breaks as a consequence of the lack of ability to repair the single-strand break effectively. In BRCA1 and BRCA2 null cells, such double-strand breaks cannot be repaired and thus PARP inhibition leads to apoptosis of such cells (91). This led to the hypothesis that their use may be of benefit in women with BRCA1/BRCA2 mutations who have tumours without BRCA1 or BRCA2 functional protein. Early results suggest efficacy in advanced breast and ovarian cancer in BRCA1/2 carriers (92), and potential in triple negative BC (93). However, reports of the possible failure of a large phase 3 study of inaparib may extend the time to potential licensing of these drugs.
The department of Genetic Medicine in Manchester is supported by the NIHR Manchester Biomedical Research Centre.