Division of Tissue Pathology, Institute of Medical and
We all make assumptions about the genetic composition of individuals based on their appearance. Gender, racial heritage and some genetic diseases produce recognisable outward features in individuals. Pathologists extend this activity to the microscopic appearances of tumours. The histologic diagnosis of certain tumours raises the possibility of associations with hereditary syndromes. The diagnosis of medullary carcinoma of the thyroid in young individuals, for example, requires exclusion of the familial form of this tumour and of the multiple endocrine neoplasia syndrome. For the common human malignancies, such as breast and colon cancer, there is increasing interest in whether tumour morphology may be predictive of an inherited cancer disposition. The purpose of this review is to summarise this information with regards to BRCA1-associated breast cancer and to discuss the utility of this morphological data in the current clinical context.
Breast cancer is not a single disease. As illustrated in figure one even routine histologic examination shows that the various tumours included under the rubric of ‘breast cancer’ appear very different from each other and follow-up studies demonstrate innate differences in the biological potential of some of these neoplasms. The classification of these seemingly disparate tumours as ‘breast cancers’ is a managerial decision, necessitated by the relative paucity of our present therapeutic options. Given this heterogeneity of breast cancers, are there some variants that are more frequently seen in the hereditary setting?
Table one summarises the most common histologic features of BRCA1-associated breast cancers.
Data from several studies1-10, including those by the Breast Cancer Linkage Consortium, show that almost all types of breast cancer can occur in BRCA1 and BRCA2 mutation carriers and like sporadic cancers, ductal carcinoma of no special type is the most common variant. BRCA1 mutation carriers have been found to have an excess of medullary and atypical medullary phenotypes. In practice, the poor inter-observer concordance in the diagnosis of medullary carcinoma limits the clinical significance of this diagnosis. Some of the criteria used for the diagnosis of medullary carcinoma, such as smooth, non-infiltrative borders were however independently associated with BRCA1 mutation1. The presence of pushing margins was also associated with BRCA2 mutation.
High-grade cancers are over-represented among BRCA1 and 2 mutation carriers (see figure one: middle panel, right). Whereas only one-third of sporadic breast cancers are of high grade, approximately two-thirds of BRCA1 cancers are grade III. Conversely, only one in 10 BRCA1-associated cancers is of low grade.
Grade is assigned histologically by following specific rules regarding the extent of tubule formation, level of nuclear atypia and the degree of mitotic activity in a tumour. BRCA1 tumours score three out of three for each of these components of grade, whereas BRCA2 tumours score significantly higher than controls only for tubule formation.
Ductal carcinoma in situ (DCIS)
The rate of proliferative activity in BRCA1-associated breast cancers is striking in some cases. This is manifested as a high mitotic rate in H&E stained sections and can also be highlighted with the use of proliferation markers, such as Mib-1 or Ki-67. This high-growth fraction is consistent with a role in regulation of cellular proliferation postulated for the BRCA1 protein.
Hormone receptor markers
Approximately 10% of BRCA1-associated breast cancers lack oestrogen or progesterone receptors (ER or PR). This compares to 35% of sporadic and BRCA2-associated breast cancers. This paucity of hormone receptors is a fundamental feature of BRCA1-associated cancers. It is seen even in low-grade cancers and at the earliest stages in which these cells are recognised as being malignant, that is, in their in situ component11. In a regression analysis ER negativity was found to be the single best predictor of BRCA1 positivity5.
While high grade and ER negative sporadic cancers are more likely to over-express HER-2, familial breast cancers are usually HER-2 negative when assessed by immunohistochemistry5,9,12. Like BRCA1, HER-2 is located on chromosome 17. It is possible that co-deletion may occur in some cases. Alternatively functional or structural suppression of the HER-2 gene may result from mutation in BRCA1.
Mutations in the TP53 gene are common in breast cancers of BRCA1 and 2 mutation carriers. Some have even found high rates of mutations in the somatic cells of BRCA1 carriers. The normal p53 protein has a short half-life and is not routinely detectable by immunohistochemistry. BRCA1 is postulated to be involved in DNA repair. If faulty BRCA1 function leads to an inability to repair mutations in the TP53 gene, the resulting products may be abnormally stable and thus detectable by immunohistochemistry.
The use of immunohistochemical analysis of cancer associated gene and encoded proteins has become important in identifying cases of hereditary non-polyposis colon cancer (HNPCC). In this context, loss of expression of MSH2, ascertained by immunohistochemistry, is highly predictive of an underlying pathological mutation in this gene. Antibodies to the BRCA1 gene product are available and loss of nuclear staining for BRCA1 has been reported in breast cancers of mutation carriers13 but to date very few reports have claimed success in using these reagents14.
a. Morphology as a guide for the selection of individuals for genetic testing
Limitations of current means of selecting patients for genetic testing
BRCA1 and 2 are large genes. Over 200 mutations have been described in each gene and these mutations are distributed throughout the length of each gene, without mutation ’hot-spots’15. This necessitates full direct sequencing of both genes to exclude an obvious abnormality. Such tests are expensive. Genetic testing also raises complex ethical and legal considerations. Careful patient selection is required in order to optimise the use of these tests.
Through the study of large, multi-case families, criteria have been developed to help select high-risk individuals for genetic testing. The criteria currently used are based principally on patient age and family history of cancer. There is no consensus as to the relative importance of various features of the family history, so that several different sets of selection criteria exist, including computer-generated risk calculators15-19. The American Society of Clinical Oncology suggests that patients whose risk of carrying a mutation exceeds 10% be considered for testing20. There are significant discrepancies in the risk estimates of these risk models, such that genetic testing may be advocated by one model, while others estimate the risk to be less than the threshold value of 10%21. Shannon et al reported 22% of routine patients attending a multidisciplinary breast cancer clinic were estimated to have a 10% probability of carrying a BRCA1/2 mutation by at least one model and should have been offered genetic counselling that included the discussion of genetic testing22. As expected, the specificity of the selection criteria is low with only 25-30% of families screened being found to be mutation carriers19,23.
Conversely, the current strong reliance on family history may deny some women the chance to be offered genetic testing. It has been pointed out that the families with an obvious cancer syndrome are likely to represent only a small fraction of individuals with inherited predisposition to cancer24. Data emerging from population-based series of early onset breast cancer suggest that a high proportion of patients with BRCA-associated cancers present as sporadic cancers25. Genetic risk calculators are not applicable to women from cancer-free families and even when a family history exists, the trend towards smaller family size may render it unimpressive. In Frank’s study 9.5% of women with breast cancer diagnosed before age 50 and mutations in BRCA1 or BRCA2, had no family history of early breast cancer or any history of ovarian cancer26.
1. Enhancement of current selection criteria for genetic testing
Attention to the morphologic features of an individual’s breast cancer can enhance current selection criteria for genetic testing. Even using the most obvious differences in oestrogen receptor expression and grade, breast cancer patients can be stratified into risk groups for likelihood of BRCA1 mutations. Table two summarises some of the promising reports into the application of morphologic data for improved selection of at-risk women. Women with high-grade, ER negative cancers are those most likely to harbour mutations in BRCA1. Even disregarding all information regarding family history, 25% of unselected pre-menopausal women with breast cancers of this phenotype were found to carry germ line BRCA1 mutations27. Others have reported similar results when focusing on high-risk groups such as women 35 years or younger28 or Ashkanazi women29. The combination of high-risk cancer phenotype and significant family history led to a BRCA1 mutation detection rate of 53%30.
Lakhani et al have published useful tables that detail the interplay between age, tumour grade and ER status in altering the individual’s likelihood of carrying a BRCA1 mutation5. It may well be time that such information was incorporated in the assessment of likelihood of BRCA1 mutations.
Improved sensitivity of selection criteria
It should be noted that in the above studies, a positive family history was infrequent in women who were found to be BRCA1 positive after being selected on the basis of phenotype27,28. Because morphologic criteria widen screening to a larger group than those with a strong family history, the sensitivity of genetic screening is enhanced, with obvious economic and ethical benefits.
Who is unlikely to carry a germ line BRCA 1 mutation? Improved specificity
The corollary of identifying high-risk patients on the basis of morphologic data is that certain tumours are unlikely to be observed in association with BRCA1 mutations (see table two). Disregarding age and family history, a woman with an ER positive, low-grade breast cancer has less than 5% chance of being a mutation carrier5. Even when women with positive family histories are tested, the magnitude of the risk for this group does not exceed 10%30. This is lower than the threshold recommended by ASCO.
2. Genetic polymorphisms versus deleterious mutation
Some sequence variations in BRCA1 do not portend an increased propensity for cancer. Determining the significance of genetic variations is problematic. It is interesting to note that in the small number of cases studied, mutations considered non-pathogenic were found to be unassociated with the usual BRCA1 cancer phenotype30.
3. A more targeted search
Which gene to test first?
Current criteria do not segregate risk of BRCA1 versus BRCA2 mutations with confidence23, such that both genes have to be sequenced until a pathogenic mutation is found. An additional benefit of taking note of tumour morphology is that the phenotype may suggest one gene over the other, thereby permitting a more targeted search. This scenario may also extend to the so-called BRCAX families, in who family histories are highly suggestive of an inherited cancer predisposition, but a mutation has not been detected to date. If morphology is strongly suggestive of BRCA1 mutation in some of these patients, a closer analysis of that gene may be attempted, since it is known that some abnormalities, for example large genomic alterations, are undetected by direct sequencing23. Alternatively, morphologic sub-groups may be found among the breast cancers of BRCAX families, potentially leading to the discovery of presently unknown genetic factors. Some early reports suggest heterogeneity among this group, but with a preponderance of low-grade cancers31.
4. More precise hereditary risk assessment
Until now, for the purposes of estimating risk of familial breast cancer, a history of DCIS has not been distinguished from a history of invasive breast cancer. Frank’s data demonstrate the importance of maintaining such a distinction26. While a history of DCIS represents some increased likelihood of mutations in BRCA1 and BRCA2, the magnitude of this risk is lower than for invasive cancer. They suggest that a history of DCIS at a particular age be given as much merit as that of an invasive cancer diagnosed 10 years later.
b. Implications for the management of affected women
Breast cancer patients who are carriers of BRCA1 mutations have a propensity for recurrence and are at risk for further primaries in both breasts32. In this setting the prevention and early detection of future breast cancers is a worthy aim. Yet certain features of these tumours pose formidable challenges in the achievement of this goal.
1. Biological perspective on the likely value of anti-oestrogens in treatment of BRCA1-associated breast cancer and in cancer risk reduction
The use of tamoxifen has been shown to halve the incidence of breast cancer among women who are at increased risk for developing breast cancer33. This includes women who have already experienced breast cancer. At first glance, women with a genetic predisposition to breast cancer would be expected to benefit from this therapy. Unfortunately, the efficacy of tamoxifen in cancer prevention is largely limited to patients whose tumours express ER. Little or no benefit is documented for the smaller proportion of sporadic patients who have ER negative tumours.
An estimated 80-90% of breast cancers in women with BRCA1 mutations are ER negative. In this setting even DCIS lacks ER11. On the basis of this fundamental ER-resistant phenotype, from a biological standpoint, one would predict that most breast cancers in BRCA1 mutation carriers would be relatively resistant to such hormonal therapy. Also, the use of tamoxifen for prevention of future cancers in cancer-free mutation carriers would be anticipated to be limited to the minority of these women who develop ER positive tumours.
The limited data available on the value of tamoxifen for cancer-free mutation carriers need to be read carefully to distinguish the outcome between BRCA1 and BRCA2 mutation carriers. This distinction is important because whereas up to 90% of BRCA1-associated cancers are ER negative, BRCA2 mutation carriers have similar rates of ER positivity to sporadic cancers, and would be expected to draw similar benefits from this medication. Some incongruous data have been presented in this regard. King performed a subset analysis of BRCA1 mutation carriers participating in the NSABP- P1 trial and found no evidence that the use of tamoxifen beginning at age 35 years or older reduced the incidence of breast cancer in those patients34. However Rebbeck35 reported a 50% reduction in the incidence of breast cancer in BRCA1 mutation carriers who underwent prophylactic oophorectomy. It is possible that earlier use of tamoxifen in cancer-free women may duplicate the effectiveness of oophorectomy. The finding of risk lesions in mastectomy specimens of these patients, would support this notion, since it raises the possibility of a pre-invasive phase for some BRCA1-associated tumours36,37.
Until results of primary prevention trials in women with BRCA1 are available, the use of anti-oestrogens in these women should be considered carefully and be accompanied by disclosure that such drugs may not be effective in reducing the risk of breast cancer.
2. Surveillance for early detection of hereditary breast cancer
Attempts at chemoprevention aside, even the early detection of these tumours poses formidable challenges. Morphologic studies of BRCA1-associated breast cancers reveal significantly lower incidence of an in situ component in these tumours. The paucity of a significant in situ component, and the high proliferation rate of these tumours, implies rapid carcinonogenesis. These observations would suggest that BRCA1 mutation carriers might not be optimal candidates for routine screening mammography. The young age of the at-risk population and the increased density of breast tissue in young women further compound the obstacles to screening mammography. These morphologic predictions are unfortunately borne out by the limited data available.
Surveillance programs in BRCA1 mutation carriers have shown that only approximately 50% of their breast cancers are detected by annual screening mammography37,38. The remaining tumours were radiologically occult a few months before, or even at presentation. Of more concern is the observation that these were not all ‘early cancers’. At least some, and in one report over half, of the interval cancers, were node positive38. Self-breast examination and clinical examination were the most common methods of detection in this cohort of women. These data suggest that the transition from a radiologically undetectable stage to clinically detectable mass is too rapid for annual screening mammography to be reliable.
These sobering data call into question the reliance on annual screening mammography for the detection of cancers in these women. Reducing the interval between screening episodes and the addition of frequent breast examinations and magnetic resonance imaging (MRI) are worthy of investigation, but at least in one of the above-mentioned surveillance programs annual mammograms and MRI were combined39. Other reports are more encouraging40,41.
3. Prophylactic surgery
Given the high risks for the development of breast cancer among BRCA1 mutation carriers, estimated to be between 50-80% for BRCA1, and the unproven efficacy of attempts at prevention or early detection of these tumours, bilateral prophylactic mastectomy is a difficult choice that should be discussed with these women. Intuitively, this procedure would be expected to be effective in cancer reduction and the limited data available suggest that this is indeed the case39,42.
Incidentally, a small proportion of women who undergo prophylactic mastectomy are found to have significant lesions, including occult DCIS in the mastectomy specimen, further underscoring the high level of risk faced by these patients36,37.
Prophylactic salpingo-oophorectomy is also being advocated in BRCA1 mutation carriers43 since not only does it offer protection against ovarian cancer, to which these women are predisposed, it has also been found to reduce the risk of subsequent breast cancer among BRCA1 mutation carriers by approximately 50%35.
Differences exist between BRCA1-associated breast cancers and BRCA2-associated and sporadic cancers. Attention to even the most rudimentary of these features opens new avenues for the identification of possible mutation carriers and offers perspectives towards more effective care of those individuals who are known to carry these mutations.
Failure to take note of these biological differences would represent a regrettable loss of opportunity for the families involved. The challenge for pathologists is to demonstrate the reproducibility and clinical validity of these distinctions. It is also hoped that highlighting the discriminative clinical value of an increasingly sophisticated array of phenotypic features can extend these benefits. This task requires study of larger numbers of cases and access to relevant tissue is currently limited. The prospective collection and proper storage of normal and tumour tissue from all appropriately consented individuals would represent a valuable resource for further investigation. In Australia we are fortunate to have already in existence organisations such as kConFab that serve these families and the scientific community by documenting important aspects of the pedigree, environmental and psychological influences in these patients. Closer involvement of pathologists in these efforts is to be encouraged.
1. Breast Cancer Linkage Consortium. “Pathology of familial breast cancer: differences between breast cancers in carriers of BRCA1 or BRCA2 mutations and sporadic cases”. Lancet, 349, 9064 (1997): 1505-10.
4. SR Lakhani, J Jacquemier, J Sloane. “Multifactorial analysis of differences between sporadic breast cancers and cancers involving BRCA1 and BRCA2 mutations”. J Natl Cancer Inst, 90 (1998): 1138-45.
5. SR Lakhani, MJ Van De Vijver, J Jacquemier, et al. “The pathology of familial breast cancer: predictive value of immunohistochemical markers estrogen receptor, progesterone receptor, HER-2, and p53 in patients with mutations in BRCA1 and BRCA2”. J Clin Oncol, 20, 9 (2002): 2310-8.
8. J Armes, A Egan, M Southey. “The histologic phenotype of breast carcinomas occurring before age 40 years in women with and without BRCA1 or BRCA2 germline mutations: a population based study”. Cancer, 83 (1998): 2335-45.
16. D Berry, J Iverson, D Gudbjartsson, et al. “BRACPRO validation, sensitivity of genetic testing of BRCA1/BRCA2, and prevalence of other breast cancer susceptibility genes”. J Clin Oncol, 20 (2002): 2701-12.
26. TS Frank, AM Deffenbaugh, JE Reid, et al. “Clinical characteristics of individuals with germline mutations in BRCA1 and BRCA2: analysis of 10,000 individuals”. J Clin Oncol, 20, 6 (2002): 1480-90.
31. SR Lakhani, BA Gusterson, J Jacquemier, et al. “The pathology of familial breast cancer: histological features of cancers in families not attributable to mutations in BRCA1 or BRCA2”. Clin Cancer Res, 6, 3 (2000): 782-9.
33. B Fisher, J Constantino, D Wickerham, et al. “Tamoxifen for prevention of breast cancer: report of the National Surgical Adjuvant Breast and Bowel project PI Study”. J Natl Cancer Inst, 90 (1998): 1371-88.
34. M King, S Wieand, K Hale. “Tamoxifen and breast cancer incidence among women with inherited mutations in BRCA1 and BRCA2. National Surgical Adjuvant Breast and Bowel Project (NSABP-P1) breast cancer prevention trial”. JAMA, 286,18 (2001): 2251-6.
41. E Warner, DB Plewes, RS Shumak, et al. “Comparison of breast magnetic resonance imaging, mammography and ultrasound for surveillance of women at high risk for hereditary breast cancer”. J Clin Oncol, 19, 15 (2001): 3524-31.
43. MS van Roosmalen, LC Verhoef, PF Stalmeier, et al. “Decision analysis of prophylactic surgery or screening for BRCA1 mutation carriers: a more prominent role for oophorectomy”. J Clin Oncol, 20, 8 (2002): 2092-100.