Abstract
Genetic testing for BRCA1 and BRCA2 mutations has become an important part of the practice of medical oncology and clinical genetics over the past decade. Increasing numbers of women are requesting a genetic test so that they may better understand their personal risks of breast and ovarian cancer, and so that they may take appropriate measures to reduce the risk. Several of the risk factors can be modified, including breastfeeding and the use of oral contraceptives. A significant number of women opt for preventive mastectomy or oophorectomy, which will dramatically reduce the risks of breast and ovarian cancer. Chemoprevention with tamoxifen is still uncommon, largely due to women's fears of the side effects of the drug. A number of studies have shown that magnetic resonance imaging is superior to conventional mammography in terms of the early detection of breast cancer in the high-risk population. This article explores what is known about assessing genetic risk and the evidence supporting a range of preventive strategies.
Keywords
It has been 10 years since Mark Skolnick and his colleagues at Myriad Genetics in Salt Lake City, USA announced that they had identified the BRCA1 gene [1]. A year later a second breast cancer gene, BRCA2, was identified by Mike Stratton and his team of researchers in England [2]. The announcements were greeted with enthusiasm – optimism was expressed by the research community (and by those who funded it) that these discoveries would lead to important advances in the prevention and treatment of breast cancer. It is now well established that both BRCA1 and BRCA2 confer high risks of breast cancer among women who are born with a mutation [3]. There is still some uncertainty about the precise risk, but almost all would agree that in the absence of any specific preventive measure the risk for breast cancer for women who carry a mutation in either gene exceeds 70%. Some mutations are exceptional in that they appear to be associated with a lower lifetime risk than this, notably the 6174delT BRCA2 mutation (which is found in Ashkenazi Jews) [3] and the (Polish) 4153delA BRCA1 mutation [4]. One of the purposes of this article is to discuss strategies that can be used to reduce the risk to an acceptable level.
It is believed that multiple genes and environmental factors are important in influencing breast cancer susceptibility [5]. However, no other breast genes have been identified since 1995 that rival BRCA1 or BRCA2 in terms of clinical importance. The identification of BRCA1 and BRCA2 sets the stage for genetic testing for breast cancer susceptibility, which is currently restricted to mutations in these two genes. A genetic test made it possible to identify women who carry a mutant gene prior to developing cancer, and to catalog women with hereditary forms of breast cancer. Testing was initially promoted with caution because many women were concerned about the possible negative consequences of testing. Genetic testing is now well established, although uptake rates vary from country to country. In some countries, such as the USA, testing is available to women through commercial laboratories, provided they have the means to pay (in most cases through private health insurance). In Canadian provinces such as Ontario, testing is available without cost through the provincial health plan to all women who satisfy the criteria for elevated risk.
The ultimate goal of genetic testing is to reduce the risk of cancer in women who are found to be predisposed. Preventive options which are now available to women include early detection (screening), preventive surgery and drug therapies. Prophylactic mastectomy was initially viewed as mutilating and thought to be excessive, but over the past decade women have shown increasing acceptance of preventive breast surgery. There is now compelling evidence that both preventive mastectomy and preventive oophorectomy are capable of dramatically reducing the cancer risk. Screening with magnetic resonance imaging (MRI) is gradually replacing the much less sensitive technique of mammography, but chemoprevention with tamoxifen (or any other drug) remains unpopular. A recent survey of North American women who received a positive genetic test result found that 60% underwent preventive oophorectomy and 25% opted for prophylactic mastectomy, but that only 12% had taken tamoxifen [6]. These options are reviewed below.
Genetic counseling
Based on a gene frequency of 1 in 250 women, there are probably 250,000 women in the USA who carry a BRCA mutation. Mutations in BRCA1 or BRCA2 are either protein-truncating (the majority) or missense mutations (relatively rare). In the first case, the mutated gene is used as a template to generate a protein that is shorter than intended (truncated). Since both BRCA1 and BRCA2 fall into the class of tumor suppressor genes, it is believed that any mutation that reduces protein function will be pathogenic. In the case of a missense mutation the mutant protein will contain one or more erroneous amino acids situated along its length. However, not all single amino acid substitutions in BRCA1 or BRCA2 are deleterious – most simple changes of this type are innocuous polymorphic variants. This makes it difficult to counsel women who are found to have a mutation of this type. Mutations are identified by comparing the DNA sequence of an individual (usually from a blood test) with the expected known sequence. If a family contains more than two cases of early-onset breast cancer and two cases of ovarian cancer, then chances are a BRCA mutation will be found. However, families as dramatic as this are rare and in practice it is only possible to identify a BRCA1 or BRCA2 mutation for the minority of women who present for genetic counseling. Factors which increase the probability that a mutation will be found include the number of relatives in the family with breast or ovarian cancer and the ages of diagnosis of the breast cancers. In particular, two or more breast cancers under the age of 40 should raise the suspicion that a genetic factor is present. Furthermore, certain characteristics of the breast and ovarian cancers themselves may help predict the presence of a mutation. High-risk pathologies include medullary breast cancer [7]. Ductal breast carcinoma in situ (DCIS) is relatively common in BRCA2 carriers and may be present in BRCA1 carriers, albeit at a lower frequency [8,9]. Only rarely will women with mucinous ovarian cancers be found to carry a mutation [10–12].
Another important factor in predicting the presence of a mutation is ethnic origin. Rather than being equally distributed in all ethnic groups, BRCA1 and BRCA2 mutations often show patterns which reflect the shared ancestry of the members of that group. For example, two mutations in BRCA1 and one in BRCA2 account for the great majority of BRCA mutations in Ashkenazi Jews [13]. If one of these three founder mutations is not present then it is highly unlikely that a different mutation will be found [14]. Other countries with founder mutations include Iceland and Poland [15,16]. In these countries genetic testing is both cheap and sensitive, because it is necessary to look for only a small number of well-characterized mutations. However, in countries with ethnically-mixed populations, such as the UK or the USA, the range of mutations is wide and it is not possible to shortcut the comprehensive search.
In general, it is reasonable to offer genetic testing to women with a significantly increased risk of hereditary breast or ovarian cancer by virtue of multiple cases of the disease, particularly if cases are early-onset or if breast cancer occurs in a male. While criteria vary, one approach is to offer testing to all women with invasive ovarian or fallopian cancer [10,17] and to women with familial breast cancer (two or more cases of breast cancer diagnosed under age 50 years or family histories of breast and ovarian cancer). The testing thresholds are less stringent for Jewish women. It is reasonable to offer genetic testing to all women of Ashkenazi ancestry with a personal or family history of breast or ovarian cancer [11,13,18].
There has been a great deal of effort expended over the last decade to try to expand the proportion of families which can be accounted for by mutations of these (and other) genes [5]. Of course, small family cancer clusters might be due to chance; however, chance alone cannot explain large families with many cases of breast cancer but no mutation. Other explanations include genomic deletions in BRCA1 or BRCA2 (that escape detection by conventional screening techniques) or mutations in yet to be identified genes [5].
Pathology
BRCA1-related breast cancers are usually infiltrating ductal carcinomas of high grade. An atypical medullary phenotype is more common in BRCA1-related breast cancer than in matched controls. These cancers show medullary features, lymphocytic infiltration, and syncytial growth patterns [7]. The most specific characteristic of a BRCA1-associated breast cancer is the presence of a lymphocytic stromal infiltrate. BRCA1-linked tumors also tend to be of higher grade than non-hereditary cancers, and are more frequently negative for estrogen (ER) and progesterone receptors (PR) [19]. DCIS is relatively rare in carriers of BRCA1 mutations, but may occur [8,9]. Tumors associated with BRCA2 mutations tend to be similar to their non-hereditary counterparts. Ovarian cancers, which occur in women with a BRCA1 mutation, also appear to be similar to sporadic cases [10–12] with the exception that mucinous tumors and tumors of low malignant potential (or ‘borderline’ tumors) are under-represented. The majority of BRCA-linked ovarian cancers show moderate to poor differentiation.
Risk factors for breast & ovarian cancers
The risks of breast and ovarian cancer in mutation carriers are not fixed but are influenced by non-genetic factors. In particular, several reproductive and hormonal factors have been found to modify the risk of breast cancer in BRCA1 and BRCA2 carriers. Some of these (such as oophorectomy and oral contraceptives) are amenable to behavioral modification and may be useful in managing breast cancer risk. Others, such as age of menarche, are not readily modified, but nevertheless may provide clues regarding cancer etiology in these women. Breastfeeding for 1 year or more (in total) reduced the risk of breast cancer by about half in BRCA1 carriers, but had no effect in BRCA2 carriers [20]. Increasing parity is a risk factor for breast cancer in BRCA2 carriers, but not in BRCA1 carriers [21]. Late menarche is protective against breast cancer in BRCA carriers [22]. A marginal increase in breast cancer risk was reported among users of oral contraceptives, but only for women who began pill use before 1975 and who reported 5 or more years of use [23]. However, a strong protective effect of oral contraceptives against ovarian cancer has also been reported in BRCA carriers [24,25]. In a recent study of 232 ovarian cancer cases and 232 controls, oral contraceptive use was associated with a 56% reduction in the risk of ovarian cancer (p = 0.002) [25]. Three years of oral contraceptive use initiated after age 25 years should provide protection effect against ovarian cancer and result in little or no increase in the risk of breast cancer. Tubal ligation has been found to be protective against ovarian cancer in the general population and among BRCA1 carriers [25]. A risk reduction of 60% was seen in the high-risk population. The combination of tubal ligation and oral contraceptives provided 72% protection for BRCA1 carriers [25].
Prophylactic mastectomy
Preventive mastectomy is clearly an effective way to prevent breast cancer, and reduces the risk of breast cancer by more than 90%. This has now been shown in several independent studies [26,27]. Total mastectomy has generally been recommended over subcutaneous, or nipple-sparing surgery, because of a high risk of cancer in the residual breast tissue. However, data about the failures of subcutaneous mastectomy are largely anecdotal. Reasons for promoting nipple-sparing surgery have recently been presented [28,29]. In principal, most women are much more favorably disposed to surgery that has minimal impact on body image and it is important that they be offered a range of procedures and be informed about the risks and benefits of each.
Contralateral breast cancer
After the initial diagnosis of breast cancer in a BRCA1 or BRCA2 carrier, the risk of cancer in the opposite breast is approximately 3% per year [30]. The risk is similar for BRCA1 and BRCA2 carriers [30]. The risk appears to be slightly greater in women who are diagnosed with their first breast cancer before age 50 years [30,31]. Because of this elevated risk many carriers with breast cancer may opt to be treated initially with bilateral mastectomy [32]. The incidence of contralateral cancer appears to be reduced by oophorectomy and tamoxifen, each of which appears to reduce risk by approximately 50% [30].
Prophylactic oophorectomy
Prophylactic oophorectomy is used to prevent both breast and ovarian cancer in BRCA1 and BRCA2 mutation carriers [33–35]. Several studies have shown that prophylactic oophorectomy is effective in preventing breast cancers. The largest of these reported that premenopausal oophorectomy reduced the risk of breast cancer by about half, and that the duration of protection was approximately 15 years [35]. Oophorectomy also appears to be effective in reducing the risk of ovarian cancer [33]. It seems logical that prophylactic oophorectomy should eliminate the incidence of ovarian cancer, but there are two reasons for possible failure of prophylactic oophorectomy. First, it is possible that the removed ovaries contain foci of occult carcinoma and that cancer has spread locally to the peritoneum at the time of the resection. In this case, the peritoneal cancer is not a primary cancer, but a metastatic ovarian cancer. Second, it is possible that de novo cancer of the peritoneum, or of the fallopian tube arises in the peritoneum after oophorectomy. The peritoneum is derived from coelomic epithelium, of the same embryologic origin as the surface epithelium of the ovary.
In about 6% of women who undergo prophylactic oophorectomy, early or occult ovarian cancers have been identified in the pathology specimens [36–40]. In many cases these occult malignancies are only detected by thorough pathology review of the ovaries and fallopian tubes. The majority of the lesions appear to arise in the fallopian tubes. It is of interest to know the survival rate of the women diagnosed with these small ovarian cancers as this will have important implications for ovarian cancer screening.
It is also not known to what extent peritoneal cancers represent new primary cancers or are the result of latent metastases of subclinical ovarian or tubal cancers. Based on the age distribution of latent cancers in oophorectomy patients and the age-specific prevalence rates of mutations in ovarian cancer patients, surgery oophorectomy should be advised at age 35 years. It is important that the fallopian tubes be removed in their entirety. However, early surgical oophorectomy results in induced menopause and the medical and psychologic consequences of this need to be discussed.
Screening for hereditary ovarian cancer
Screening for ovarian cancer using serial CA-125 levels and abdominal ultrasound has been proposed as a method of reducing mortality through early detection [41,42]. There have been no randomized trials of screening in BRCA1 carriers, but observational cohort studies have been disappointing. Liede and colleagues identified seven incident ovarian/peritoneal cancers in a historical cohort of 33 BRCA carriers who underwent regular screening examinations [43]. Six of the seven cases were Stage III at the time of diagnosis. For the majority of cases, the ultrasound findings were normal prior to diagnosis and the women presented with pain or abdominal distension.
Tamoxifen
In a large case-control study tamoxifen was found to reduce the incidence of contralateral breast cancer in affected BRCA1 and BRCA2 carriers by approximately half [44]. A follow-up study on this population documented significant reductions in breast cancer risk in both mutation subgroups [45]. To the extent that contralateral cancers in carriers are representative of all new primary breast cancers, the results of this study might be extrapolated to the prevention of first primary breast cancers. But this conclusion has generally not been accepted and the use of tamoxifen for primary prevention of hereditary breast cancer remains rare [6].
Breast cancer screening
The goal of screening is to identify a breast cancer at a stage when a surgical cure is likely. Traditionally, this includes small breast cancers (<1 cm) that are node-negative and with no evidence of distant spread. For these women cure can be expected in the majority of cases. But BRCA1-associated breast cancers are typically of high grade and are estrogen-receptor-negative and thus prognosis might be expected to be worse than average. Among BRCA1 carriers there was little correlation between tumor size and lymph-node positivity in one study; about a third of BRCA1 carriers had lymph node metastases detected at diagnosis, regardless of tumor size [46]. Therefore, it may be problematic to predict the benefits of screening using survival data generated from a comparison group of non-carriers.
BRCA-associated tumors may be particularly hard to detect mammographically. Pushing margins, breast density, and mutation status contribute independently to false-negative mammograms in BRCA carriers [47]. Goffin and colleagues found that only two of eight breast cancers (25%) in BRCA1 carriers were detectable by mammogram at diagnosis, versus 27 of 35 (77%) from non-carrier controls (p = 0.01) [48]. In a large cohort at a single center, of 12 breast tumors diagnosed in BRCA mutation carriers, fewer than half were found by mammogram [49].
Breast MRI offers the promise of greatly improving the early detection of breast cancers in women at high risk. The sensitivity of MRI greatly exceeds that of mammography. In the largest series reported to date, the sensitivity of MRI was 83% for invasive breast cancer [50]. Only two of the 22 women with breast cancer (9%) detected by MRI in a Canadian trial had lymph node metastases, and both of these were detected in the initial screen [51]. However, when MRI is used, false-positive findings have been noted; in the first year of screening the false-positive rate is on the order of 10% at most centers.
Breast cancer treatment
Currently, the typical management of most women with hereditary breast cancer differs little from the management of nonhereditary cancers. However, because of the 32% estimated 10-year risk of contralateral breast cancer in BRCA1 carriers [30] and the 13% risk of ovarian cancer [52] some women with Stage I or II breast cancer may choose to undergo prophylactic oophorectomy and/or contralateral mastectomy as part of their initial treatment plan. With regard to radiation or chemotherapy, relatively little research has been completed on the assessment of various treatment approaches for hereditary breast cancer.
Concern has been expressed that ionizing radiation may pose a special hazard for women with BRCA mutations, who are deficient in their ability to repair radiation-induced DNA breaks [53]. However, there are no empirical data yet to suggest that this is the case for therapeutic radiation or for mammography. Radiotherapy appears not to increase the risk of cancer in the opposite breast [30] and the incidence of local reactions to radiation has not been found to be exceptional in BRCA carriers [54].
Several of the functions of the BRCA1 and BRCA2 have been discovered over the past decade. Both proteins are integral to the DNA damage response pathway and facilitate DNA repair through homologous recombination [55]. Most women with BRCA1-associated breast cancer will have high-grade, estrogen-receptor-negative tumors, and are therefore candidates for chemotherapy. It has been suggested that BRCA1-associated tumors are highly sensitive to certain chemotherapy agents such as mitomycin and platinum, or to chemotherapy in the setting of adjuvant or neoadjuvant administration [56], but it has not yet been shown that these treatments reduce mortality. The majority of BRCA1-associated tumors are estrogen-receptor-negative and in general hormonal ablative treatments are not indicated for these patients. However, oophorectomy has been shown to prevent primary breast cancers, local recurrences and contralateral breast cancers. Studies of the effect of oophorectomy with mortality as the end point are underway.
The survival of patients with BRCA-associated ovarian cancer appears to be better than that of women with sporadic ovarian cancer [57]. A recent study of consecutive cases of ovarian cancers, which compared BRCA associated with sporadic ovarian cancer from the same institution, found that BRCA mutation status was a favorable and independent predictor of survival for women with advanced disease [58]. It is not yet clear if the improved survival rates are the result of a difference in the natural history of ovarian cancer in the two subgroups, or is the result of a better response of BRCA-associated tumors to current therapies. Laboratory data suggests that enhanced response to chemotherapy underlies the better survival. In support of this position, Cass and colleagues reported that BRCA1 carriers with ovarian cancer had a higher response rate to primary therapy than did matched noncarriers, and carrier patients with advanced disease had improved survival (91 months for BRCA1 carriers vs 54 months for non-carriers; p = 0.05) [58].
Conclusions
The discovery of the BRCA1 and BRCA2 genes 10 years ago enhanced our ability to identify women at risk for breast cancer due to an inherited risk. It has also allowed us to subdivide the category of familial cancers into more meaningful groupings, each with a different molecular basis. No other gene has been identified in the past decade which has modified cancer care – some positive associations have been reported, but for these, either the frequency of mutations is too low or the relative risk incurred is too modest to lead to the dissemination of a clinical test. We have seen important advances in our understanding of the nature of hereditary breast cancer, in terms of basic science and in the clinical arena. Both prophylactic mastectomy and prophylactic oophorectomy have come to be accepted and are capable of significantly reducing the cancer risk in women who carry a mutation. Women who retain their breasts should be screened annually by MRI examination.
Future perspective
Over the next 10 years I expect progress to continue in the areas of prevention, screening and treatment. We should see a gradual acceptance of subcutaneous mastectomy as an alternative to total mastectomy and increasing numbers of mutation carriers who opt for this type of surgery. It is important that the women who develop breast cancer following prophylactic mastectomy be documented in order that we may better understand the causes for failure.
Currently, chemoprevention remains unpopular among BRCA carriers because of a lack of prospective data. It is important that the use of tamoxifen or other antihormonal chemopreventive agents be evaluated by observation of cohorts of women with mutations. Possibly the greatest degree of protection will come from the combination of oophorectomy and an aromatase inhibitor.
The majority of ovarian cancers in BRCA1 carriers appear to originate in the fallopian tubes, but the ovary and the peritoneum may also be the sites of origin. More and more ovarian/tubal cancers are being diagnosed in the early stages as a result of prophylactic oophorectomy. It is important that these women be followed for all outcomes in order that the benefits of preventive surgery be evaluated. These data will have important implications for determining the optimal timing of surgery and for screening as well. If the survival rate of these early stage cancers is not good then it is unrealistic to expect that we can make an impact on ovarian cancer mortality through screening.
Increasing attention is being paid to the possible hazards of exposing the breasts of young BRCA1 and BRCA2 carriers to ionizing radiation. MRI examination is gradually replacing mammography as the screening tool of choice, and methods will be devised to better distinguish benign from malignant lesions that are detected by MRI. However, despite the shift in stage and nodal status in breast tumors identified with MRI, versus mammography or clinical examination, it has not yet been proven to reduce mortality. It will be useful to identify a cohort of MRI detected asymptomatic breast cancers so that 5- and 10-year survival rates can be constructed.
There are several reasons to believe that BRCA1-associated breast cancers will respond to hormonal treatments, despite their being estrogen-receptor-negative. It is important to establish if there is an improvement in survival of women with estrogen-receptor-negative cancers in the context of BRCA1 or BRCA2 mutations. Studies of chemosensitivity on tumors in BRCA carriers, and on cultured cells, will lead to more directed treatments for women with cancers in this category.
Executive summary
BRCA1 and BRCA2 are the only two genes for which clinical testing for breast cancer susceptibility is clinically accepted. Testing should be offered when there are two cases of early-onset (< 50) breast cancer, or ovarian cancer (any age) in a family. It is also important to take into account ethnic origin of the patient if genetic testing services are offered. Populations with founder effects, including those in Poland, Iceland and in Ashkenazi Jews, have contributed greatly to our understanding of the epidemiology of hereditary breast cancer.
BRCA1-associated breast cancers typically present as high-grade ductal or medullary cancers with a lymphocytic stromal infiltrate. BRCA2-associated cancers do not have characteristic features.
The risk of breast cancer in BRCA1 carriers can be reduced by breastfeeding and oophorectomy. The risk of contralateral breast cancer can be reduced by tamoxifen.
Preventive mastectomy will reduce the risk of breast cancer to levels below population risk. Subcutaneous mastectomy should be offered to women who do not wish to undergo total mastectomy. The effectiveness of both procedures needs to be compared.
Breast screening by MRI is much more sensitive than mammography and does not result in increased radiation exposure. In almost every case, MRI will detect a breast cancer prior to it being detectable by mammography.
The risk of contralateral breast cancer in a BRCA1 or BRCA2 carriers is approximately 3% a year and can be cut in half by tamoxifen or oophorectomy. The risk is nearly eliminated by contralateral mastectomy.