early breast cancer - prevention and genetic screening

Prevention and Genetic Screening
Approximately 10% of breast cancer patients have familial breast cancer, typically defined as breast cancer showing an autosomal dominant inheritance pattern (625). During the 1990s, germline mutations in three important tumor suppressor genes—p53, BRCA1, and BRCA2—were discovered in family members of individuals with familial breast cancer (100,354,462). All three genes have been shown unequivocally to predispose a woman to breast cancer.
Mutations in p53
Germline mutations in the p53 gene are very rare and result in Li-Fraumeni syndrome, named after two investigators who made significant contributions to the understanding of this condition (354). The p53 gene is one of the most important tumor suppressor genes and has been called the guardian of the genome because of its critical role in cellular pathways that recognize and direct a response to DNA injury. One consequence of a germline mutation in p53 is an increased risk for a variety of cancers, including childhood sarcomas, gynecologic tumors, and breast cancer. Breast cancer is the most common malignancy in patients with Li-Fraumeni syndrome; the lifetime risk is estimated to be 90% (354).
Mutations in BRCA1 and BRCA2
Studies of patients with familial breast cancer led to the discovery of BRCA1 in 1995 and BRCA2 in 1996. Similar to p53, both BRCA1 and BRCA2 are tumor suppressor genes that contribute to the stability of the genome by mediating the effects of the cellular response to DNA injury. Individuals with a germline mutation in BRCA1 have a lifetime risk of breast cancer of 65% to 85% (171,172). In addition, these individuals have an elevated lifetime risk of ovarian cancer, which may approach 50%. Other types of cancer that develop more frequently in BRCA1 carriers include colon cancer and prostate cancer. The lifetime risk of breast cancer for women with germline BRCA2 mutations mirrors that for women with BRCA1 mutations. BRCA2 mutation carriers are also at increased risk for ovarian cancer compared with the general population, but their risk is much less than the risk in women with BRCA1 mutations. BRCA2 is also associated with pancreatic cancer and male breast cancer. Genetic screening for germline mutations in BRCA1 and BRCA2 is now possible. Testing should be performed in centers equipped with genetic counseling programs designed to properly inform individuals of the social, economic, and legal consequences associated with genetic testing. Germline mutations in BRCA1 and BRCA2 are rare, occurring in fewer than 7% of patients with breast cancer. Thus, only a minority of breast cancer patients with a family history of the disease would be predicted to carry a mutation in one of these genes. Table 53.2 contains data concerning the probability of carrying a BRCA1 mutation based on an individual's age at cancer diagnosis, personal cancer history, and family cancer history and whether the individual is of Ashkenazic Jewish descent (677).

No definitive data exist on which to base screening recommendations for individuals with a proven germline mutation in a gene predisposing to the development of breast cancer. The American Society for Clinical Oncology (ASCO) has published a consensus statement recommending that these individuals undergo annual mammography and clinical and self-breast examination beginning at the age of 25 to 35 years (17). In addition, annual pelvic examinations with transvaginal sonography, color Doppler examinations of the ovaries, and measurement of serum CA-125 levels are recommended beginning at age 25 to 35 years. As will be discussed later, magnetic resonance imaging (MRI) has emerged as a promising screening tool for women at high risk for breast cancer.
Breast Cancer Prevention Strategies
Tamoxifen
Understanding of the role of estrogen and progesterone in breast cancer development has led to the development of pharmacologic strategies that could significantly decrease the incidence of breast cancer over the next two decades.
P.1182


Several pharmaceuticals that affect the estrogenic pathways have been studied as chemopreventive agents, but the only agent for which mature data from clinical trials are available is tamoxifen (40,133,136,137,209,780). Interest in tamoxifen as a chemopreventive agent arose after a number of randomized trials designed to test the efficacy of hormonal therapy for invasive breast cancer reported that tamoxifen reduced the incidence of contralateral breast cancer. On the basis of these data, in 1992 the National Surgical Adjuvant Breast and Bowel Project (NSABP) began a randomized, placebo-controlled study (the P-1 trial) to test the efficacy of 5 years of tamoxifen use in the prevention of breast cancer (209). Between 1992 and 1997, 13,388 women with a 1.67% or greater predicted risk of developing breast cancer within 5 years were enrolled in this trial. Risk was assessed using a modification of the Gail model, which permitted enrollment of any woman older than 60 years of age and selected women younger than 60 years with additional risk factors that increased their annual risk to at least that of a 60-year-old. In addition, women with a history of lobular carcinoma in situ (LCIS) were included because their risk was believed to exceed the cutoff risk of 1.67%. Women were not allowed to use estrogen replacement therapy during their participation in the trial. The results of the P1 trial indicated that tamoxifen reduced the rates of invasive and noninvasive breast cancer by 49% and 50%, respectively. The benefit of tamoxifen was seen in all age groups (≤49 years, 50 to 59 years, ≥60 years). In addition, women with a history of atypical ductal hyperplasia had an 86% risk reduction, and women with a history of LCIS had a 56% risk reduction. Finally, the benefit was seen across all subgroups specified according to family history of breast cancer. Tamoxifen selectively reduced the incidence of estrogen receptor–positive tumors; estrogen receptor–negative tumors developed at an equal rate in the tamoxifen and placebo groups. No evidence was shown of a cardioprotective effect of tamoxifen in this trial, but the number of osteoporosis-related fractures was reduced in the tamoxifen-treated cohort. Tamoxifen increased the risk of developing stage I endometrial cancer (risk ratio of 2.53).
Although this study indicated that 5 years of tamoxifen use decreased the 5-year risk of developing breast cancer by about 50%, whether this result warrants the widespread use of tamoxifen for breast cancer prevention is controversial. In the NSABP report, 5 years of tamoxifen therapy in 6,576 women reduced the number of invasive or noninvasive breast cancer by 120 compared with the expected number. With the relatively short follow-up period of this important study, however, it was not possible to determine whether tamoxifen actually prevented this number of cases of breast cancer or rather simply delayed the onset of the disease.

More recently, a landmark study was published that compared tamoxifen to raloxifene as a preventative agent in postmenopausal women with breast cancer. Raloxifene, a drug that is primarily used in prevention of osteoporosis, had been shown in prior studies to decrease the incidence of breast cancers. The NSABP study of tamoxifen and raloxifene trial was a prospective, double-blind, randomized clinical trial (804). There were 19,747 postmenopausal women of mean age 58.5 years with increased 5-year breast cancer risk. Patients were randomized to oral tamoxifen (20 mg per day) or raloxifene (60 mg per day) for 5 years. There were 163 cases of invasive breast cancer in women assigned to tamoxifen and 168 in those assigned to raloxifene, which was not significantly different between the two arms. The main benefit of raloxifene was in toxicity. There were 36 cases of uterine cancer with tamoxifen and 23 with raloxifene. No differences were found for other invasive cancer sites, for ischemic heart disease events, or for stroke. Thromboembolic events also occurred less often in the raloxifene group. The number of osteoporotic fractures in the groups was similar, and there were fewer cataract surgeries in the group using raloxifene. There was no difference in the total number of deaths (101 for tamoxifen vs. 96 for raloxifene) or in causes of death. It appears from this study that raloxifene is as effective as tamoxifen in reducing the risk of invasive breast cancer and has a lower risk of thromboembolic events. Of note there were slightly more noninvasive cancers in the raloxifene group, but that was not statistically significant.
A comparison of the tamoxifen P1, P2, and other tamoxifen prevention trials is outlined in Table 53.3.
Prophylactic Surgery
An alternative strategy used to prevent breast cancer development is prophylactic surgical intervention. Hartmann et al. (335) analyzed outcomes in women with a family history of
P.1183

breast cancer who underwent bilateral prophylactic mastectomy at the Mayo Clinic between 1960 and 1993. With a median follow-up time of 14 years, only four of the 639 treated patients developed breast cancer. According to the Gail model, 37.4 cases of breast cancer would have been expected to develop in this population, so the prophylactic surgery resulted in an 89.5% risk reduction (p <.001) (335).
As will be discussed later in the section on management of patients with BRCA1 and BRCA2 mutations, prophylactic oophorectomy in BRCA carriers also significantly reduces the risk of subsequent breast cancers in this population.
Breast cancer prevention strategies will continue to be a dynamic area of preclinical and clinical research in the near future.
Natural History and Origins
All forms of breast cancer are believed to develop as a consequence of unregulated cell growth and the development of phenotypic changes such as the ability to invade, recruit a new blood supply, and metastasize. These changes in phenotypes are secondary to the development of aberrations in genetic pathways. Some of these aberrations are inherited (germline mutations), whereas others develop during the life of a breast cell (somatic mutations). It is currently believed that most breast cancer is a consequence of a series of somatic mutations. As previously noted, only 20% to 25% of breast cancer patients have a history of breast cancer in a first-degree relative. However, it is possible that some women without a first-degree relative with breast cancer still inherit a genetic background that predisposes to breast cancer. These mutations may be insufficient to cause breast cancer unless accompanied by other mutations and, therefore, would be predicted to have a low penetrance. Historically, it has been much more difficult to discover low-penetrance mutations than to discover germline mutations that result in an autosomal dominant pattern of breast cancer development. However, with newer molecular techniques, such as deoxyribonucleic acid (DNA)–array assays, the identification of low-penetrance predisposing mutations may be more feasible.
Left untreated, breast cancer can have a variable clinical course. A classic paper by Bloom et al. (65) outlined the natural history of breast cancer patients seen between 1805 and 1933, not treated by surgery or irradiation, 250 of whom had a pathologic diagnosis of cancer. There were no patients with stage I disease, 2.4% with stage II, 23% with stage III, and 74% with stage IV. Survival in the untreated group was 3.6% compared to an overall survival of 34% in patients treated with radical or modified radical mastectomy with or without radiation.
Concepts regarding the natural history of breast cancer have undergone great evolution over the past 100 years, with a profound impact on the management of these patients. The Halsted (315) model was based on an orderly progression to the regional lymph nodes and from there to distant metastatic sites. Later, Keynes (399) and Crile et al. (130) suggested that breast cancer is a systemic disease, and that extensive surgery to achieve local tumor control was not as important as originally believed. This alternative hypothesis was fully demonstrated in both laboratory and clinical studies by Fisher (216), who advanced the concept that breast cancer, as a systemic process involving host–tumor interactions, would not show substantial effects on survival with variations in locoregional treatment. A third hypothesis put forward by Hellman (344) considers breast cancer as a heterogeneous disease with a spectrum extending from a tumor that remains localized throughout its course to one that disseminates systemically even when detected as a small lesion, suggesting that metastases are a function of tumor growth and progression factors.
The growth rate of a tumor in the breast is thought to be constant from the date of origin. Using estimates of doubling time, it would take an average of approximately 5 years for a tumor to reach palpable size, and those lesions with slower doubling time would have an even longer latent period (293).
The most common site of origin of breast cancer is the upper-outer quadrant (38.5%), followed by the central area (29%), the upper-inner quadrant (14.2%), the lower-outer quadrant (8.8%), and the lower-inner quadrant (5%) (293). These rates correlate with the amount of breast tissue in the various quadrants. Cancer is somewhat more common in the left than in the right breast and may appear in both breasts simultaneously (1% to 2%). As noted above, women with a history of breast cancer have a 10% to 15% risk of developing a new primary in the contralateral breast.
As the cancer grows, it travels along the ducts, eventually breaking through the basement membrane of the duct, invading adjacent lobules, ducts, fascial strands, and the mammary fat, spreading through the breast lymphatics and into the peripheral lymphatics. The tumor can grow through the wall of blood vessels, spread into the deep lymphatics of the dermis, and eventually produce edema of the skin (peau d'orange), which usually indicates that the superficial as well as the deep lymphatics are involved. Skin dimpling can be caused by involvement of Cooper's ligament. Ulceration and infiltration of overlying skin, which may develop late in the course of the disease, are usually preceded by fixation and localized redness of the skin over the tumor and are less frequently seen because of the current emphasis on screening and early diagnosis (293).
Axillary Spread
A common route of spread of breast carcinoma is first through the axillary lymph nodes, with the incidence increasing with larger tumors. Depending on mode of detection, tumor size, histology, and other clinical-pathological factors, between 10% and 40% of newly diagnosed stage T1 and T2 breast cancers have pathologic evidence of axillary nodal metastases. Voogd et al. (806) assessed 7,680 patients with documented invasive breast cancer; of 5,125 patients known to have clinically negative lymph nodes who underwent axillary dissection, 1,748 (34%) had positive lymph nodes at pathologic examination. Univariate analysis showed that lymph node metastases were associated with tumors larger than 1 cm (p = .001), moderate or poorly differentiated nuclear grade (p = .005), high fraction of cells in the growth phase (S phase) of the cell cycle (p = .041), presence of lymphatic vascular invasion (p <.001), and age younger than 60 years (p = .01).
Table 53.4 demonstrates the strong relationship between primary tumor size and axillary nodal involvement. Even patients with T1a and T1b disease have significant nodal involvement. Mustafa et al. (519) noted an overall frequency of axillary lymph node metastases in T1a and T1b lesions of 16%; integrating age, tumor size, and grade predicted the frequency of nodal metastases. Overall, patients with all three poor prognostic indicators had a 34% incidence of nodal involvement, and those with no poor prognostic factors had a 7% or less probability of nodal metastases. Gann et al. (271) reviewed 18,025 patients with a diagnosis of breast carcinoma from the American College of Surgeons database. On multivariate analysis, the following factors were independently associated with a greater likelihood of one or more positive lymph nodes: larger tumor size, young age, African American or Hispanic race, outer-half tumor location, poor or moderate differentiation, aneuploidy, and infiltrating ductal histologic type.
Although up to 30% to 40% of T1–2 clinically node-negative breast cancers may have pathologically involved lymph nodes, data from NSABP-04 suggest that less than half of clinically negative but pathologically positive axilla will experience a clinical
P.1184

relapse in the axilla (212). In this study, operable breast cancer patients, who were primarily diagnosed with palpable breast tumors in the premammography era, were randomized to one of three arms: simple mastectomy without axillary dissection, simple mastectomy with axillary dissection, or simple mastectomy with comprehensive chest wall and regional nodal irradiation. In the arm undergoing axillary dissection, nodal positivity was approximately 40%. Nodal control was excellent (>97%) in this arm as well as in the arm treated with radiation. In the simple mastectomy arm, where no nodal treatment by radiation or dissection was administered, the axillary failure rate was approximately 20%. Assuming equal distribution among the arms, it is presumed the pathological involvement was approximately 40%, indicating that less than half of those with pathological involvement eventually failed clinically.

Nodal involvement has more recently been assessed by sentinel node techniques. Kamath et al. (388) analyzed 101 women with sentinel lymph node metastases on whom subsequent complete lymph node dissection was performed. Sentinel lymph node micrometastases (<2 mm) detected by cytokeratin staining were associated with a 7.6% (2/26) incidence of positive complete lymph node dissection, compared with a 25% (5/20) incidence when micrometastases were detected initially by routine hematoxylin-and-eosin (H&E) staining. Sentinel lymph node micrometastases, regardless of identification technique, conferred a risk of 15.2% (7/46) for nonsentinel lymph node involvement, which increased with larger tumors. Sentinel nodes are discussed in more detail later in the surgical management of breast cancer.
Internal Mammary Spread
Metastases to the internal mammary nodes (IMNs) are correlated with tumor size, are more frequent from medial half and central lesions, and occur more frequently when there is axillary node involvement (Table 53.5) (323). Veronesi et al. (777) found that, among women with tumors larger than 2 cm who were younger than 40 years of age and had positive axillary nodes, there was a 41% risk of having positive IMNs on IMN dissection; the corresponding risk for patients of that age with negative nodes was 16%. Sugg et al. (723) reviewed 286 patients with breast cancer who underwent IMN dissection. Positive IMNs were associated with primary tumor size (p <.0001) and the number of positive axillary nodes (p <.0001), but not with age or primary tumor location. Patients who had positive IMNs (25% of all patients) had a significantly worse overall 20-year disease-free survival rate than did patients with negative IMNs (p <.0001). Clinical failure of the internal mammary nodes is extremely rare, despite the evidence of pathological involvement from these studies. Most studies looking at nodal failure patterns report failure in the internal mammary region of <1% (242,269,299,318,606).

Supraclavicular Spread
Spread to supraclavicular lymph nodes usually allows involvement in the high axillary lymph nodes or IMNs depending on the location of the primary lesion. Chen et al. (110) reviewed 2,658 patients with invasive breast cancer who underwent surgery and adjuvant therapy. With a median follow-up period of 39 months, supraclavicular lymph node metastasis developed in 113 (4.3%). Young age (≤40 years), tumor size >3 cm, angiolymphatic invasion, negative estrogen receptor (ER) status, and DNA synthetic phase fraction >4% were significant for predicting supraclavicular metastasis on univariate analysis. Three predictive factors were significant after multivariate analysis: high histologic grade, more than four positive nodes, and axillary level II or III involved nodes. In patients
P.1185

with axillary level I involved nodes and four or fewer positive nodes, the incidence of supraclavicular lymph node metastasis was 4.4%, but if axillary level III was involved, it increased to 15.1%.
Clinical failure in the supraclavicular fossa is relatively rare in patients with early stage breast cancer and is dependent on the degree of axillary involvement. For patients with no or minimal nodal involvement (<3 involved axillary nodes), supraclavicular failure is extremely rare. In an analysis of 691 patients with 0 to 3 nodes involved undergoing breast-conserving surgery and radiation therapy to tangential fields only without regional nodal irradiation, Galper et al. (269) reported failure in the supraclavicular fossa in 1.3% of patients.
Several studies have demonstrated that the failure rate in supraclavicular nodes, left untreated, may be as high as 20% in patients with advanced disease and/or more than four lymph nodes involved (207,219,220,722,797). In a cohort of 1,031 patients with operable breast cancer treated with mastectomy and level I or II node dissection plus adriamycin-based chemotherapy, but no radiation, Strom et al. (722) reported failure in the supraclavicular fossa was 8% at 10 years. Predictors of supraclavicular failure included four or more involved nodes and gross extranodal extension. In these subgroups, supraclavicular failure ranged from 14% to 19%. Radiation to the supraclavicular fossa in these higher-risk patients results in high local control rates, with isolated supraclavicular failures occurring in less that 1% of prophylactically treated nodes.
Systemic Spread
Using monoclonal antibodies to epithelial cytokeratins or tumor-associated cell membrane glycoproteins, carcinoma cells can be detected on cytologic bone marrow (or lymph node) preparations. Braun et al. (77) combined patient data from nine studies involving 4,703 patients with stage I, II, or III breast cancer. Micrometastasis was detected in 30.6% of the patients. With a median follow-up of 5.2 years, patients with bone marrow micrometastasis had larger tumors and tumors with a higher histologic grade and more often had lymph node metastases and hormone receptor-negative tumors, compared to those without bone marrow micrometastasis. The presence of micrometastasis was a significant prognostic factor for poorer overall survival (RR = 2.15; p <.001), breast cancer–specific survival (RR = 2.44; p <.001), disease-free survival (RR = 2.13), and distant-disease–free survival (RR = 2.33; p <.001 for all outcomes measures). In multivariate analysis, micrometastasis was an independent predictor of a poor outcome.
Local Control and Systemic Metastasis
Patients treated for breast cancer are at risk for local–regional failure as well as systemic metastasis. It is evident from the available literature that optimizing local control can impact systemic metastasis and survival, and similarly, systemic therapy has an impact on local control (124,300,358,803). Integration of systemic therapy with radiation will be discussed in detail later, but numerous studies have clearly demonstrated a significant improvement in local control with the use of radiation therapy and systemic therapy (both cytotoxic and hormonal) compared with radiation therapy without the use of systemic therapy (88,201,204,208,300,402). Appropriate integration of both local and systemic treatments through a multidisciplinary approach is thus essential to optimize outcome. Although there is some overlap, prognostic factors for local–regional control and systemic metastasis often differ.
For patients with early stage invasive breast cancer, even with appropriate systemic therapy, development of metastasis can vary from <5% in women with T1a disease and favorable histology, to over 40% for women with T2 tumors and pathologically involved lymph nodes. Similarly, local–regional failure rates can vary from <5% to over 40% depending on local treatment and prognostic factors for local failure (201,204,208,246,302,768,778).
The impact of local control on systemic metastasis in breast cancer as well as other malignancies has been the subject of considerable debate and controversy. Although the benefits of local control with respect to cosmesis and quality of life are apparent, the independent effect of local control on systemic disease and survival has been questioned. Several studies have identified local control as an independent predictor of disease free and/or overall survival (111,202,241,302,307,778). Fisher et al. (202), in an analysis of patients treated in NSABP Protocol B-06, concluded that ipsilateral breast tumor recurrence was a harbinger, but not a cause, of distant metastases. Although mastectomy or breast irradiation after lumpectomy prevented expression of the marker (breast relapse), neither lowered the risk of distant metastases, which was determined by a host of prognostic factors.
More recent meta-analyses, however, have demonstrated a small but significant impact of local control on systemic metastasis and overall survival (124,803). A recent meta-analysis of randomized trials by Vinh-Hung et al. (803) comparing breast-conserving surgery without radiation to breast-conserving surgery with radiation confirms an approximate threefold reduction in local relapse with radiation therapy and an 8.6% improvement in mortality in the radiated cohorts.
One of the most convincing and authoritative studies related to this subject is the recent analysis of the Early Breast Cancer Trialists Collaborative Group (EBCTCG) (124). In this analysis over 42,000 women were enrolled in 78 randomized trials that compared 24 types of local treatment (radiotherapy vs. no radiotherapy, more vs. less surgery, or more surgery vs. radiotherapy). The EBCTCG attempted to relate the effect on local control to breast cancer mortality by grouping studies into whether the 5-year local relapse risk difference between the two comparisons of local therapy exceeded 10%. In those comparisons in which the difference in 5-year local recurrence risk was <10%, there was no impact on 15-year breast cancer mortality. However, there were 25,000 women enrolled in trials in whom the comparisons involved >10% differences in local control. In those studies, the difference in local recurrence risks at 5 years were 7% versus 26%, and the 15-year mortality risks were 44.6% versus 49.5% (p <.00001). The authors concluded that avoidance of a local recurrence in the conservatively managed breast or avoidance of a local–regional relapse after mastectomy had similar impact on breast cancer mortality. In the absence of any other cause of death, differences in local treatment that substantially affect local recurrence would avoid about one breast cancer death over 15 years for every four local relapses avoided. Figure 53.6 summarizes the results of this meta-analysis with respect to the impact of radiation on breast cancer mortality in both breast conservation and following mastectomy.
Another study, which supports the notion that local control influences survival, even in patients with known metastatic disease, is a report by Rapiti et al. (594). In 300 metastatic breast cancer patients recorded at the Geneva Cancer Registry between 1977 and 1996, they compared mortality risks from breast cancer between patients who had surgery of the primary breast tumor to those who had not and adjusted these risks for other prognostic factors. Women who had complete excision of the primary breast tumor with negative surgical margins had a 40% reduced risk of death due to breast cancer (hazard ratio [HR] = 0.6; 95% CI, 0.4 to 1.0; p = .049) compared with women who did not have surgery. The effect was most evident for women with bone metastasis only (HR, 0.2; 95% CI, 0.1 to 0.4; p = .001).