Risk Factors

Risk factors for breast cancer can be divided into those that are modifiable and those that are not. Nonmodifiable risk factors include gender, age, family history, age at menarche, age at menopause, race, and history of prior benign breast biopsy. Modifiable factors include parity, age at first live birth, mammographic density, breast-feeding, obesity and weight gain, exogenous hormones, radiation, alcohol consumption, and diet.

Aside from being female, age is the single most important breast cancer risk factor. According to the National Cancer Institute, the risk between ages 30 and 39 is 0.43% (1 in 233), 40 and 49 is 1.44% (1 in 69), 50 and 59 is 2.63% (1 in 38), and between 60 and 69 is 3.65% (1 in 27) based on probabilities for the whole population and not individual risk factors.¹⁰

A family history of breast cancer particularly in a first-degree relative is a significant risk factor, and the risk escalates with the number of relatives affected and younger age at diagnosis. This pattern suggests an inherited genetic mutation that predisposes to the development of breast cancer. Approximately 5% to 10% of breast cancer patients have a familial form of the disease.²¹ Many of these cases contain an alteration in the breast cancer genes, BRAC1 and BRAC2. More than 100 distinct mutations have been identified in high-risk families and it is not clear if all carry an equal cancer risk. Some populations have a higher likelihood of carrying germline mutations such as family members of Ashkenazi Jewish (Eastern European) heritage and families with multiple cases of breast and/or ovarian cancers. The estimated lifetime risk of developing a breast cancer is up to 80% (36% to 85%), with a near 40% risk of developing a contralateral breast cancer. The risk of developing an ovarian cancer is 40% in BRCA1 carriers and 20% for BRAC2 carriers.^22,23 Genetic counseling should be offered to these patients including those of young age at diagnosis, two primary breast cancers (ipsilateral or contralateral) or with breast and ovarian cancer and male breast cancer.²⁴ There are other rare familial genetic syndromes that display a predisposition to breast cancer, the most well-known are Li-Fraumeni syndrome with germline mutations in TP53 and Cowden and Bannayan-Riley-Ruvalcaba syndrome due to PTEN mutations.

The absolute risk of a contralateral breast cancer in women with a personal history is 0.5% to 1% per year or up to 10% during the 10 years following diagnosis.^25,26 Biopsy-proven atypical proliferative disorders, including atypical lobular hyperplasia (ALH), lobular carcinoma in situ (LCIS), and atypical ductal hyperplasia (ADH), may increase the risk by a range of fourfold to tenfold with a further increase in a patient with a family history.^27–30

Mammographic density is a strong independent risk factor with a fourfold to sixfold increase for postmenopausal women with high breast density compared with those with least dense breasts.^31–36 Breast density refers to the amount of white area (fibrous and glandular tissue) on a black (primarily fat tissue) mammogram. While methods of measurement and definitions of breast density vary among studies, women with breast densities of more than 60% to 75% have been found to have an increased risk of breast cancer.³⁷ Higher breast density is more common in Caucasian women and younger women and decreases during menopause. Hereditary factors may account for the majority of highly dense breasts. Several studies have identified genes that influence mammographic density.^38–40 Tamoxifen has been shown to decrease breast density especially during the first 18 months of treatment.⁴¹

The risk of developing breast cancer after exposure to ionizing radiation is dose and age dependent and has been demonstrated from data collected from the Japanese atomic bomb survivors and patients exposed to radiation for nonmalignant conditions, such as thymus enlargement, multiple chest fluoroscopies for tuberculosis, and mastitis examinations.^42–45 Particularly susceptible are adolescents who demonstrate the greatest risk of breast carcinogenesis over a lifetime. Secondary breast cancer has been described in young women who underwent mantle irradiation for Hodgkin disease with doses ranging from 20 to 44 Gy.⁴⁶ In a multi-institutional review of 1380 adolescents treated before age 16 years old, a cumulative probability of breast cancer at age 40 years was 35%. This cohort of survivors had a risk of breast cancer 75 times higher than that of the general population.⁴⁷ The risk of breast cancer significantly increases 15 to 30 years after treatment for those women exposed between the ages of 10 to 30 years; however, relative risk begins to increase several years following radiation exposure.⁴⁸ Current practice using lower doses of radiation and limited fields for Hodgkin disease may decrease the risk of breast cancer for these patients.⁴⁹

A moderate relative risk is associated with factors which affect circulating hormone levels such as delayed childbirth, nulliparity, early or late menarche and exogenous hormones. Body mass index (BMI) or postmenopausal obesity, has clearly been associated with breast cancer risk likely due to higher estradiol levels associated with aromatase in adipose tissue, which converts androgens to estradiol. A pooled analysis of prospective studies demonstrated a 30% higher risk in postmenopausal woman with a BMI more than 31 kg/m² compared with a BMI of less or equal to 20 kg/m².⁵⁰ Weight gain at menopause is associated with increased risk; however, weight loss with maintenance is associated with a substantial lowering of breast cancer risk.⁵¹

Alcohol consumption increases the risk of breast cancer. In the Oxford meta-analysis of 53 epidemiologic studies, 58,515 patients with breast cancer and 95,067 women without breast cancer demonstrated that two drinks a day (defined as 24 g of alcohol) can increase breast cancer risk by 21%.⁵² The relative risk of breast cancer was dose-dependent and increased with daily amount. Another analysis of 184,418 postmenopausal women showed that moderate intake of one to two drinks a day can increase risk by about 32% and three or more by 51%.⁵³

Environmental factors, pollutants, tobacco, nutrition and physical activity have not clearly been linked with breast cancer risk to date. Diet, nutrition, and physical activity are clearly interrelated with obesity and BMI but are difficult to dissect apart. Environmental pollutants and other environmental factors are an ongoing topic of investigation.

Two commonly used models to estimate an individual’s breast cancer risk are the Gail and Claus models which heavily weigh family history with a combination of other risk factors.^54,55 A breast cancer risk assessment tool is available on the National Cancer Institute website based on the Gail model (cancer.gov/bcrisktool). Other decision models have been developed to estimate the likelihood of a BRCA mutation include the BRCAPRO.^56,57 and the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA).⁵⁸

Clinical Presentation and Detection

The most frequent presentation of early stage breast cancer is an asymptomatic, nonpalpable mass, which is detected as an abnormality on screening mammogram. The most common physical sign is a nonpainful mobile mass.⁵⁹ A detailed physical examination includes evaluation of the ipsilateral and contralateral breast tissue and regional lymph nodes (bilateral axillae, supraclavicular, infraclavicular, anterior cervical/neck, and submental and submandibular lymph nodes chains). The treatment approach is determined, in part, by the clinical presentation such as tumor size, location, and skin involvement. A large tumor can be a contraindication to breast conservation; inner quadrant tumors require lymphoscintigraphy to identify sentinel nodes located in the internal mammary chain.

Physical signs of a locally advanced breast tumor include enlarged or matted axillary or supraclavicular lymphadenopathy, an abnormal breast contour, asymmetrical size, and a palpable mass larger than 5 cm with or without fixation to the chest wall. Skin erythema or abnormal color changes, thickening, peau d’orange, ulceration, and the presence of satellite skin nodules are signs of advanced disease. Focal dimpling of the skin or erythema or nipple retraction does not by itself imply locally advanced disease. Inflammatory breast cancer is a subcategory of locally advanced disease characterized by rapid onset, skin erythema, induration, warmth, asymmetric enlargement, often without a discrete breast mass. Pretreatment photographs are encouraged for patients with obvious physical signs of cancer particularly in the era of preoperative chemotherapy for locally advanced breast cancer where initial clinical stage, presentation, and subsequent clinical response are integral to management and outcome.

Mammographic signs of cancer consist of two primary findings: (1) a mass with ill-defined, irregular, or spiculated edges and/or (2) irregular, pleomorphic calcifications. Further diagnostic imaging including targeted ultrasonography and mammographic magnification and compression views of all suspicious areas should be completed before a biopsy procedure. The imaging reports should include size and location of the primary tumor and a description of the findings in accordance with the American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) guidelines.⁶⁰ A pathologic diagnosis of cancer is necessary before proceeding to definitive surgery. Fine-needle aspiration of a palpable mass or lymph node can be performed at the time of the initial evaluation and before a core needle biopsy. A benign FNA is an incomplete evaluation if discordant with the clinical findings and must be followed with an ultrasound-guided core needle biopsy, stereotactic biopsy, or an excisional biopsy. Clinically enlarged lymph nodes are also evaluated by FNA or core needle biopsy and computerized tomography (CT)/positron emission tomography (PET). Once a breast cancer diagnosis has been made, magnetic resonance imaging (MRI) can be a diagnostic tool for assessing the extent and pattern of cancer in the breast parenchyma and detecting occult contralateral cancers.^61,62

Mammography as a screening tool has been proven to reduce mortality from early stage breast cancer.^63–65 Additional imaging tools, such as breast ultrasound and dynamic/contrast-enhanced MRI, have not been shown to impact mortality, and given the technical and practical limitations have no universal screening role but are useful supplemental techniques. MRI has been shown to have superior sensitivity for detecting breast cancers in high-risk populations over mammography, clinical examination, and ultrasound. The cost of higher sensitivity is the variable specificity that is lower than mammography in all studies.⁶⁶

In woman with a high risk of breast cancer due to a strong family history or a BRCA mutation, routine surveillance with clinical breast examination and mammography is suboptimal compared with MRI. Six prospective nonrandomized studies have demonstrated the role of MRI screening in this high-risk population.^67–72 One of these studies followed 236 women with BRCA mutations with breast examination, mammography, ultrasound, and MRI. More than twice as many cancers were detected by MRI (77%) than by mammography (36%), and the combined sensitivity of all four modalities was 95% compared with the 45% from breast examination and mammography alone.⁶⁸ Similar results were obtained from a larger study, which surveyed 1909 high-risk women (15% or more cumulative lifetime breast cancer risk) with annual breast examination, mammography, and MRI. With a median follow-up of 2.9 years, 51 tumors were detected with a sensitivity of 18%, 33%, and 79%, respectively, and specificity of 98%, 95%, and 90%, respectively.⁶⁷

The American Cancer Society published guidelines for the use of screening magnetic resonance imaging (MRI) as an adjunct to mammography in 2007.⁶⁶ Annual screening MRI was recommended for proven carriers of a BRCA mutation, untested first-degree relatives of a proven mutation carrier and women with a lifetime risk of breast cancer at 20% as determined by risk assessment models such as BRCAPRO. It was also recommended for women with a prior history of chest radiation between the ages of 10 to 30 years and for women with Li-Fraumeni syndrome or other rare genetic syndromes and their first-degree relatives. The role of annual MRI screening in women with a 15% to 20% lifetime risk for breast cancer (i.e., those with a personal history or extremely dense breasts on mammography) was not clarified. Annual MRI screening was not recommended for women with less than 15% lifetime risk of breast cancer.

Mastectomy

Not all patients will be candidates for BCS even after neoadjuvant therapy, and many will choose mastectomy regardless for personal reasons. Mastectomy offers an excellent oncologic treatment for breast cancer, and with new surgical techniques, the cosmetic results have greatly improved if the patient desires breast reconstruction. The most recent advance in surgical technique has been the introduction of the total skin-sparing mastectomy or nipple-sparing mastectomy.^185,186 Leaving the nipple-areolar complex (NAC) intact has led to considerable improvements in cosmetic results without compromising the oncologic procedure itself. The tissue directly underneath the areola is sharply dissected and the nipple is inverted to core out all ductal tissue with an approximate 5% rate of nipple loss secondary to ischemia. Contraindications to nipple-sparing mastectomy are inability to achieve acceptable margins at the NAC or direct involvement of the nipple tissue itself. Patients can choose either autologous tissue transfer or implant reconstruction following nipple-sparing mastectomy.

It is generally accepted that immediate reconstruction after mastectomy is safe in terms of local recurrence rates and does not significantly prolong the start of adjuvant therapy.^187,188 Most plastic surgeons agree that immediate reconstruction offers a better aesthetic outcome in conjunction with the psychologic benefit of awakening from surgery with an intact reconstructed breast.^189,190 However, there is considerable debate surrounding immediate reconstruction for those patients who are at a higher risk of requiring postmastectomy radiation therapy (PMRT) with concern for an increased rate of complications. Numerous studies have shown that radiation increases the rate of TRAM flap complications, namely fat necrosis and flap necrosis.^191,192 These complications can occur both with immediate reconstruction and delayed reconstruction. Some would argue that the immediate reconstruction with skin-sparing or nipple-sparing mastectomy leads to a superior cosmetic result and that it is beneficial in light of the potential for volume loss and fibrosis that can accompany PMRT. In an attempt to avoid these complications, but also to achieve excellent cosmesis, some have implemented a “two-stage” approach.¹⁹³ If the patient is deemed high risk for PMRT, they undergo either a skin-sparing mastectomy or nipple-sparing mastectomy with placement of a subpectoral tissue expander to preserve the skin envelope. After review of the final pathology, patients who do not require PMRT can undergo immediate reconstruction, while those who require PMRT can have this treatment with the expander in place. Following radiation the patient can then undergo delayed reconstruction with the preserved skin envelope.

Axilla

Sentinel lymph node biopsy has proven to be an accurate technique for staging the axilla in breast cancer patients and is associated with less morbidity when compared with complete axillary node dissection. Several prospective trials in the United States and Europe have randomized women to sentinel node biopsy and axillary dissection versus sentinel node biopsy and axillary dissection only if the sentinel node is positive. Veronesi et al.¹⁹⁴ reported the results of the European Institute of Oncology trial for stage I breast cancer. The false-negative rate was 8.8% and the negative predictive value was 95.4%. There were no axillary recurrences in the women who had sentinel node biopsy only. In the NSABP B-32 trial of 5611 women, sentinel nodes were identified in 97.1%, the false-negative rate was 9.8%, and the negative predictive value was 96.1%.⁴ When performed by experienced hands, the false-negative rate has been shown to be less than 5%, with long-term recurrence rates in sentinel node-negative patients approaching 0.3%.¹⁹⁵

Completion axillary dissection is recommended for patients with positive sentinel nodes.¹⁹⁶ This recommendation is based on the finding from a meta-analysis of 69 trials of sentinel node biopsy of 48% nonsentinel node positivity in women with positive sentinel nodes.¹⁹⁵ However, the role of completion axillary dissection in women with immunohistochemical (IHC) only positive sentinel nodes or micrometastases (>0.2 mm and <2 mm) is unknown. A number of predictive models have been developed to estimate the risk of nonsentinel node positivity in patients with positive sentinel nodes. These include three nomograms^197–199 and three scoring systems.^200–202 Factors evaluated in these models include method of detection (frozen section, routine H+E, serial H+E, or IHC), primary tumor size, nuclear grade, multifocality, presence of lymphovascular invasion, estrogen receptor status, size of sentinel node metastasis, number of positive sentinel nodes, and number of sentinel nodes removed. It has been suggested by some investigators that women with a less than 10% risk of nonsentinel node positivity may forego completion axillary dissection.^203,204 The role of axillary irradiation in patients with positive sentinel nodes who do not undergo an axillary dissection has been evaluated in eight single institution studies (Table 58-3).

It should be noted that in a significant number of patients, the sentinel node had isolated tumors cells or was IHC positive only. Overall axillary recurrence rates are 3% or less with median follow-up times of 2.5 years. Axillary recurrence appears to be low whether the low axilla is included in the tangential fields or with axillary radiation. This issue is being investigated in an EORTC phase III randomized trial²¹³ in which sentinel node-positive patients with primary tumors 0.5 to 3 cm are randomized to axillary dissection or axillary radiation.²¹³ Accrual is estimated to be complete in 2010. The ACOSOG Z0011 trial was designed to ask a similar question but closed prematurely due to poor accrual in the United States.²¹⁴

Sentinel lymph node biopsy, which began as a staging tool in early stage primary breast cancer, has now been expanded to patients undergoing neoadjuvant therapy and patients undergoing a reoperation for recurrent breast cancer with a previous negative sentinel node biopsy.²¹⁵ Patients with an ipsilateral recurrence or a new primary breast cancer after previous SLNB may be spared a complete axillary node dissection by undergoing remapping with lymphoscintigraphy. Of course, in the 20% to 25% of patients who fail mapping in this setting, the axilla should be dissected for complete staging and eradication of potential disease.

The use of SLNB in the neoadjuvant setting has become more widespread with increased use of preoperative chemotherapy, leading to controversy surrounding the issue of timing of the sentinel node procedure. Proponents of performing SLNB after neoadjuvant therapy argue that complete axillary response is possible in 20% to 40% of women harboring lymph node metastases.^216,217 Eradication of this disease allows the woman to undergo SLNB only, foregoing the morbidity associated with a complete axillary node dissection. Other advantages to SLNB following neoadjuvant therapy include limiting surgery to one combined procedure and avoiding a delay in starting neoadjuvant therapy secondary to an axillary node dissection in the event a positive SLN is identified in a preadjuvant setting.

Those who advocate performing the SLNB before initiation of neoadjuvant therapy would argue that the procedure’s accuracy is compromised if done at completion of neoadjuvant therapy. The false-negative rate for SLNB following neoadjuvant therapy has been shown to vary from 9% to 30% in published series with a recent meta-analysis of 21 published studies investigating more than 1200 women reporting a false-negative rate of 12% and an identification rate of 90%.^218,219 Some would argue that these latter statistics are within the expected range for SLNB accuracy in a nonneoadjuvant setting; however, with more experience now, these numbers are more likely to be in the 2% and 98% range, respectively. Another reason to argue for SLNB upfront involves accurate staging of the axilla to be used in determining the extent of radiation therapy. Decisions on treatment rely on knowing the number of positive nodes, and this information could be lost if axillary staging is done after neoadjuvant therapy.

One issue that most surgeons agree on is the use of SLNB only in patients who are deemed clinically positive in the axilla at the time of diagnosis, or have a biopsy proven positive axillary node. The false-negative rate for SLNB following neoadjuvant therapy for these patients is unacceptably high, and should only be done within the confines of a clinical trial.²²⁰ Axillary dissection with or without a SLNB is recommended in this setting. Whether to perform SLNB before or after neoadjuvant therapy is surgeon dependent, but should be discussed within a multidisciplinary team to achieve the best treatment plan for the patient.

Risk Stratification for Adjuvant Therapy

Recommendations for adjuvant therapy are generally based on estimation of a patient’s risk for recurrence and the expected benefit of therapy. Factors such as patient age, comorbidities, and an individual’s concerns over toxicity and side effects also play a role in the choice of treatment. In the past, recurrence risk was predicted based on characteristics such as tumor size, tumor grade, and lymph node involvement. More recently, however, prognostic information derived from assays such as Oncotype DX and MammaPrint have become more frequently incorporated into clinical decision-making.

Benefit of Adjuvant Chemotherapy and Hormonal Therapy—Evidence from the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG)

The Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) was established in 1984 to conduct meta-analyses of data from randomized trials of the treatment of early breast cancer. The first meta-analysis published in 1988 served as the foundation of evidence supporting the benefits of adjuvant chemotherapy and endocrine therapy for early breast cancer. In the first systematic overview, 28 trials of tamoxifen and 40 chemotherapy trials demonstrated significant reductions in mortality when tamoxifen was compared with no tamoxifen and any chemotherapy was compared with no chemotherapy. In tamoxifen trials, there was a reduction in mortality of about 20% in the first 5 years among women 50 or older. For women under 50, there was a 25% reduction in the annual odds of death during the first 5 years after treatment with combination chemotherapy. This meta-analysis also included direct comparisons showing that combination chemotherapy was significantly more effective than single-agent therapy. However, no survival advantage was observed for administration of chemotherapy for 8 to 24 months compared with use of the same chemotherapy for 4 to 6 months.

The most recent EBCTCG results published in 2005 combined six meta-analyses: anthracycline-based versus no chemotherapy; CMF-based versus no chemotherapy; anthracycline-based versus CMF-based chemotherapy; 5 years of tamoxifen versus none; 1 to 2 years of tamoxifen versus none; and 5 years versus 1 to 2 years of tamoxifen.²²⁶ Treatment with approximately 6 months of anthracycline-based chemotherapy with a regimen such as 5-fluorouracil, Adriamycin, and cyclophosphamide (FAC) reduces the annual breast cancer death rate by about 38% for women younger than 50 years of age and by about 20% for those of age 50 to 69. This meta-analysis also confirmed a moderate but highly significant advantage for anthracycline-based regimens (those including Adriamycin or epirubicin) to regimens consisting of cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) in trials involving more than 14,000 patients. Since few women greater than 70 years of age participated in chemotherapy trials in the past, the benefits of chemotherapy in this patient population are not as well defined.

The Evolution of Adjuvant Chemotherapy—Practice Changing Trials

Systemic chemotherapy has become an integral component of the adjuvant treatment of early stage breast cancer since investigators began reporting significant improvements in disease-free survival (DFS) with single-agent chemotherapy after radical mastectomy in the 1970s. Polychemotherapy was first evaluated by Bonadonna, who randomized women with node-positive breast cancer to 12 monthly cycles of cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) chemotherapy or no further therapy after radical mastectomy²²⁷ (Table 58-4).

After more than 19 years of follow-up, significant improvements in relapse-free survival (RFS, relative risk 0.65) and overall survival (OS, relative risk 0.76) were observed.²³⁹ The benefit of postoperative chemotherapy for ER-negative, node-negative breast cancer patients was further reinforced through a series of clinical trials by the National Surgical Adjuvant Breast and Bowel Project (NSABP). The B-13 trial randomized 760 patients to surgery alone or methotrexate and 5-fluorouracil (MF), and with 16 years of follow-up, an overall benefit was seen with MF relative to surgery alone (RFS: HR = 0.59, P < .001; OS: HR = 0.75, P = .03).²²⁸ The B-19 trial randomized 1095 patients to MF or cyclophosphamide, methotrexate, and 5-fluorouracil (CMF), and with 13 years of follow-up, an overall benefit was seen for CMF relative to MF (RFS: HR = 0.59, P < .001; OS: HR = 0.71, P = .01).²²⁸ In the NSABP B-20 study, 2306 ER-positive, node-negative women were randomized to adjuvant tamoxifen alone, MF plus tamoxifen (MFT) or CMF plus tamoxifen (CMFT).^240,241 The study was updated after 12 years of follow-up revealing an RFS of 89% versus 79% and an OS of 87% versus 83%, demonstrating a benefit for the addition of chemotherapy (either MF or CMF) to tamoxifen.²⁴²

A number of trials have compared anthracycline-based to CMF-based regimens. The NSABP B-15 trial compared 2 months of doxorubicin and cyclophosphamide (AC) with 6 months of CMF in 2194 patients with ER-negative breast cancer. This trial also evaluated whether AC followed in 6 months by intravenous (IV) CMF was more effective than AC without reinduction therapy. Through 3 years of follow-up, there were no significant differences in disease-free survival (DFS), distant disease-free survival (DDFS), or overall survival among the three groups.²⁴¹ In NSABP B-23, 2008 ER-negative breast cancer patients were assigned to CMF administered every 4 weeks for six cycles or doxorubicin and cyclophosphamide (AC) administered every 3 weeks for four cycles.²³⁰ Patients were also randomized to receive tamoxifen to determine if there was any benefit from hormonal therapy in this ER-negative patient population. With 8 years of follow-up, there were no statistically significant differences between the CMF and AC groups (RFS: HR = 1.00, P = .97; OS, HR = 0.92, P = .51). As expected, tamoxifen with either chemotherapy regimen resulted in no significant advantage over that achieved from chemotherapy alone. The National Cancer Institute of Canada (NCIC) MA5 trial compared the efficacy of an intensive cyclophosphamide, epirubicin, and fluorouracil (CEF) adjuvant chemotherapy regimen with CMF in 710 node-positive premenopausal patients. With a median follow-up of 10 years, RFS was 52% for patients who received CEF compared with 45% for CMF patients (HR for CMF versus CEF = 1.31; P = .007), and the 10-year OS for patients who received CEF and CMF were 62% and 58%, respectively (HR for CMF vs. CEF = 1.18; P = .085).²³¹

Adjuvant Hormonal Therapy

The NSABP B-14 investigators first demonstrated the role of tamoxifen in the management of ER-positive, node-negative breast cancer when they randomized approximately 3000 women to receive either adjuvant tamoxifen or a placebo.²⁴⁷ The study was updated in 2004 after 15 years of follow-up, with a recurrence-free survival (78% versus 65%) and overall survival (71% versus 65%) benefit with the addition of tamoxifen.²⁴² In the updated EBCTG analysis with 15 years of follow-up time, 5 years of adjuvant tamoxifen reduces the risk of breast cancer recurrence by 41% and the annual breast cancer death rate by 33% in estrogen receptor-positive breast cancer. Furthermore, 5 years of tamoxifen treatment was shown to be significantly more effective than 1 to 2 years of tamoxifen.²²⁶

While tamoxifen remains the appropriate adjuvant hormonal treatment for premenopausal patients, a number of large, randomized trials have evaluated the role of the aromatase inhibitors (AIs) in postmenopausal women with hormone receptor-positive breast cancer (Table 58-5). These studies have compared tamoxifen to AIs in a number of treatment approaches. Two large trials have reported outcomes comparing initial treatment with tamoxifen for 5 years versus an AI for 5 years. The Arimidex, Tamoxifen, Alone or in Combination (ATAC) trial randomized more than 9000 postmenopausal women to receive tamoxifen, anastrozole, or the combination. The trial results were first presented in 2002 at which time treatment with 5 years of anastrozole was superior to tamoxifen in terms of disease-free survival.²⁵⁶ When anastrozole was administered concurrently with tamoxifen, the benefit of the AI was lost so subsequent follow-up data have focused on the AI versus tamoxifen arms of the trial. Additionally, since a significant percentage (16%) of patients on this trial were estrogen receptor-negative or unknown, analyses of treatment benefit have been presented for both the intent to treat population and for the estrogen receptor-positive population for whom the benefit of hormonal therapy is proven. With a median follow-up of 100 months, disease-free survival was improved by 15%, time to recurrence by 24%, time to distant recurrence by 16%, and contralateral breast cancer by 40% in the patients receiving an AI.²⁴⁸ Despite the fact that absolute differences in time to recurrence increased and recurrence rates remained significantly lower on anastrozole compared with tamoxifen after treatment completion, there was no difference in overall survival between tamoxifen and the AI. The second trial that compared upfront AI to 5 years of tamoxifen is the Breast International Group (BIG) 1-98 study in which women were randomly assigned to tamoxifen for 5 years, letrozole for 5 years, 2 years of tamoxifen followed by 3 years of letrozole or 2 years of letrozole followed by 3 years of tamoxifen.²⁵⁷ At a median follow-up time of 51 months, the use of upfront letrozole for 5 years resulted in a significant reduction in the risk of an event (HR.82; P = .007) compared with upfront tamoxifen for 5 years.²⁴⁹ Preliminary data from the arms comparing sequencing of tamoxifen and letrozole (2 years of tamoxifen followed by 3 years of letrozole or vice versa) to 5 years of letrozole were recently reported at the 2008 San Antonio Breast Cancer Symposium. Due to a significant portion of patients in the tamoxifen only arm crossing over and starting letrozole, a comparison of 5 years of tamoxifen to the sequential arms was not performed in this analysis. With a median follow-up of 71 months, 5-year DFS was 87.9% for the patients receiving 5 years of letrozole, 87.6% for patients who received letrozole for 2 years followed by tamoxifen for 3 years, and 86.2% for patients who received 2 years of tamoxifen followed by 3 years of letrozole.²⁵⁸ Statistically, there was no difference in DFS, OS, or time to distant recurrence for these three arms.

Several trials have reported more mature outcomes from switching strategies. The largest of the switching trials is the Intergroup Exemestane Study (IES), which randomly assigned 4724 postmenopausal women who were disease free after 2 to 3 years of tamoxifen to continue tamoxifen or switch to exemestane for a total duration of 5 years of endocrine therapy. After a median follow-up of 55.7 months, switching to exemestane was associated with a significant improvement in DFS (HR = 0.76; P = .0001) and when hormone receptor-negative patients were excluded, an improved overall survival (HR = 0.83; P = .05).²⁵¹ The Austrian Breast and Colorectal Cancer Study Group (ABCSG) trial 8 and the Arimidex-Nolvadex (ARNO 95) are two multicenter, open-label trials that randomized postmenopausal women with hormone-sensitive early breast cancer who had completed 2 years of adjuvant oral tamoxifen (20 or 30 mg daily) to receive 1 mg oral anastrozole (n = 1618) or 20 or 30 mg tamoxifen (n = 1606) daily for the remainder of their adjuvant therapy. At a median follow-up of 28 months, a 40% reduction in the risk for an event was observed in the anastrozole group as compared with the tamoxifen group (HR = 0.60; P = .0009).²⁵⁰

Data from the EBCTG meta-analysis reveals that more than half of all recurrences occur in years 6 to 15 after diagnosis²²⁶ and therefore interventions to reduce the risk of late recurrences have been explored. Studies extending the course of tamoxifen beyond 5 years have yielded inconclusive results. In the NSABP B-14 study,²⁴² where patients were reassigned to either continuation of tamoxifen beyond 5 years or a placebo, a slight advantage was observed in patients who discontinued tamoxifen relative to those who continued to receive it: DFS = 82% versus 78% (P = .03), RFS = 94% versus 92% (P = .13), and survival = 94% versus 91% (P = .07), respectively. Two small combined ECOG studies of patients with node-positive breast cancer (E4181 and E5181) found no statistically significant differences in time to relapse or survival between women continuing to receive tamoxifen beyond 5 years and those on observation. With a median follow-up of 5.6 years since randomization to continue tamoxifen, 85% of the women receiving tamoxifen were disease free compared with 73% of those on observation (P = .10); survival was 86% for those continuing to receive tamoxifen and 89% for those on observation (P = .52).²⁵⁹ Patients with estrogen receptor-positive tumors appeared to experience a longer time to relapse with continued tamoxifen therapy (P = .014), but there was no the survival difference for this subgroup. Lastly, in the Scottish Adjuvant Tamoxifen Trial patients who were disease free at 5 years were randomly assigned either to stop taking tamoxifen or to continue taking it indefinitely until relapse or death. With a median follow-up of 15 years, no additional benefit was observed in those randomly assigned to continue taking tamoxifen beyond 5 years.²⁶⁰ Despite the preponderance of evidence suggesting that extension of tamoxifen therapy beyond 5 years is not beneficial, two large ongoing trials, ATLAS (Adjuvant Tamoxifen-Longer Against Shorter) and aTTom (Adjuvant Tamoxifen Treatment-Offer More) may offer new insight into the optimal duration of tamoxifen therapy.

In contrast to the trials evaluating prolonged duration of tamoxifen, several other trials have shown that extension of hormonal therapy by switching to an aromatase inhibitor can be beneficial. The National Cancer Institute of Canada (NCIC) MA-17 was a double-blind, placebo-controlled trial evaluating the effectiveness of 5 years of letrozole therapy in postmenopausal women with breast cancer who completed 5 years of tamoxifen therapy.²⁶¹ At the first interim analysis, the estimated 4-year disease-free survival rates were 93% for the letrozole group and 87% for the placebo group (P = .001). These findings led to study termination and unblinding. With a median follow-up of 30 months, women in the letrozole arm had statistically significantly better DFS and distant DFS than women in the placebo arm. Overall survival was not different for both arms, but among lymph node-positive patients, overall survival was statistically significantly improved with letrozole (HR = 0.61, 95% CI = 0.38 to 0.98; P = .04).²⁵⁴ In a similar design, the NSABP B-33 trial evaluated the steroidal aromatase inhibitor exemestane as extended adjuvant therapy in postmenopausal breast cancer patients who were disease free after 5 years of tamoxifen. Due to the unblinding of MA-17, patients in the placebo arm of B-33 were offered exemestane, and 44% of these patients elected to receive exemestane. With 30 months of median follow-up and despite early closure of the study, patients originally assigned to exemestane experienced a borderline statistically significant improvement in 4-year DFS and in a statistically significant improvement in 4-year relapse-free survival compared with patients originally assigned to the placebo group.²⁵⁵ The Austrian Breast and Colorectal Cancer Study Group (ABCSG) Trial 6a evaluated 3 years of anastrozole or no further treatment among women who were disease free at the end of the ABCSG Trial 6, in which they had received 5 years of adjuvant tamoxifen, with or without 2 years of the aromatase inhibitor aminoglutethimide. At a median follow-up of 62.3 months, women who received anastrozole had a statistically significantly reduced risk of recurrence compared with women who received no further treatment (HR = 0.62; 95% CI = 0.40 to 0.96, P = .031).²⁵³ Taken together, these three trials provide strong support for extending adjuvant tamoxifen therapy beyond 5 years with an AI in postmenopausal patients. The optimum length of extended adjuvant therapy requires further research, and patients who have completed 5 years of letrozole on the MA-17 study are being rerandomly assigned to continuation of letrozole versus a placebo. This study will provide evidence of safety for extension of treatment with AIs beyond 5 years.

Biologic Therapies—Trastuzumab

HER-2 is overexpressed or gene amplified in 20% to 25% of breast cancers. The anti-HER-2 monoclonal antibody trastuzumab targets HER-2-positive tumors, inhibits proliferation, and induces cell death through a variety of mechanisms. Four large phase III adjuvant trials (NSABP B-31, NCCTG N9831, HERA, and BCIRG 006) including more than 11,500 patients have investigated the efficacy and safety of 1 or 2 years of trastuzumab given in combination with or after standard adjuvant chemotherapy. The addition of 1 year of trastuzumab to adjuvant chemotherapy has significantly improved disease-free survival (DFS) by 33% to 52% and overall survival by 34% to 41% in the four trials. The DFS benefits were observed regardless of age, nodal status, hormonal status, or tumor size in all trials.

The NSABP B-31 trial compared four cycles of doxorubicin and cyclophosphamide (AC), followed by four cycles of every-3-week paclitaxel (P) (AC→P, arm 1) with AC followed by P with 52 weeks of trastuzumab beginning with the first cycle of paclitaxel (AC→PH→H, arm 2) . The N9831 trial was a three-arm study comparing four cycles of AC followed by 12 weekly doses of paclitaxel (AC→P, arm A) with AC→P followed by 52 weeks of H beginning after P (AC→P →H, arm B) and AC→P plus 52 weeks of H beginning with the first P cycle (AC→PH →H, arm C). Because arm 1 and arm 2 of the B-31 trial are similar to arms A and C of the N9831 trial, the studies were amended to include a joint statistical analysis combining arm 1 and arm A for comparison with arm 2 and arm C. Arm B of the N9831 trial was excluded from the analysis because H was not given concurrently with P. In both trials, only patients with tumors scored as with 3+ or more staining for HER-2 by immunohistochemistry (IHC) or gene amplification by fluorescence in situ hybridization (FISH) were eligible. In an intent to treat analysis with a combined median follow-up of 2 years (2.4 years for B-31 and 1.5 years for N-9831), there was a highly significant 52% reduction in the risk of disease recurrence with sequential trastuzumab (DFS 87% versus 75%, HR 0.48), and despite the short follow-up, a 33% reduction in the risk of death (3-year OS 94.3 versus 91.7%, HR 0.67).²⁶²

A different design was employed in the HERA trial, which compared 1 year or 2 years of trastuzumab given every 3 weeks with observation (no trastuzumab) in patients with HER-2/neu-positive early breast cancer who had completed local-regional therapy and at least four cycles of neoadjuvant or adjuvant chemotherapy. The results of 1 year of trastuzumab versus observation were reported and included 1698 patients receiving only chemotherapy and 1703 patients treated with 1 year of trastuzumab after chemotherapy who were followed for an average of 24 months.²⁶³ The addition of trastuzumab after chemotherapy resulted in a statistically significant 36% reduction in disease recurrence (HR 0.64, 3-year DFS of 81% versus 74%) and a significant improvement in overall survival (HR 0.66, 92% versus 90% in the trastuzumab and nontrastuzumab groups, respectively).

Patients enrolled in the smaller FinHer trial were randomized to three cycles of docetaxel or vinorelbine followed by three cycles of fluorouracil, epirubicin, and cyclophosphamide. The primary aim of that trial was to compare treatment using docetaxel with treatment using vinorelbine. The subset of women with HER-2-positive tumors (n = 232) was further randomized to either receive or not receive trastuzumab for 9 weeks together with the first three cycles of docetaxel or vinorelbine. Within this subgroup, DFS was significantly better among those who received trastuzumab (89% versus 78%, P = .01) and there was a trend toward better OS (96% versus 90%, P = .07).²⁶⁴ Surprisingly, the magnitude of those benefits is similar to those seen in the other trials that used 1 or 2 years of trastuzumab treatment.

The BCIRG-006 trial evaluated the efficacy and safety of three regimens as adjuvant systemic therapy: (1) doxorubicin and cyclophosphamide followed by trastuzumab plus docetaxel chemotherapy (AC-TH), (2) docetaxel and carboplatin plus trastuzumab (TCH), and (3) AC followed by docetaxel alone (AC-T) as the control arm. The results of the first interim analysis of 3222 patients was presented at the San Antonio Breast Cancer Symposium in 2005.²⁶⁵ At a median of 23 months of follow-up, DFS was significantly improved in patients treated with any trastuzumab-containing regimen, compared with chemotherapy alone, with a hazard ratio of 0.49 for AC-TH compared with AC-T (P < .0001) and 0.61 for TCH (P = .0002). No significant difference in DFS was found between the two treatment arms that included trastuzumab (P = .16). However, the absolute benefit in DFS from an anthracycline plus trastuzumab compared with TCH was 4%. The second planned interim analysis was presented in 2006, and with median follow-up at 36 months, both AC-TH and TCH significantly improved the DFS and OS over the control (relative reduction risk of relapse 39% (P < .0001) and 33% (P = .0003) respectively, for AC-TH and TCH versus control).²⁴⁶ Relative reduction in the risk of death was 41% (P = .0041) and 34% (P = .017), respectively, for AC-TH and TCH versus control. At both time points, there were fewer symptomatic cardiac events and a lower incidence of asymptomatic left ventricular ejection fraction decline with TCH compared with either anthracycline group.

At the 2007 SABCS, however, the first trial that did not demonstrate a benefit for adjuvant trastuzumab therapy in HER-2-positive disease was presented. The PACS 04 study randomized 3010 women with node-positive disease to either six cycles of epirubicin and docetaxel (ED) or six cycles of 5-FU-epirubicin-cyclophosphamide (FEC-100). The 526 women with HER-2-positive disease underwent a second randomization to either 1 year of adjuvant trastuzumab or no adjuvant trastuzumab after the completion of their primary chemotherapy. After a median follow-up of 40 months, there was no difference in either DFS or OS for the trastuzumab versus the nontrastuzumab containing arms (HR for DFS 0.86, P = .41).²⁶⁶ It is not clear why the PACS 04 trial did not demonstrate a benefit similar to the other adjuvant trastuzumab trials, although it is important to note that 10% of women randomized to the trastuzumab arm were never treated with adjuvant trastuzumab therapy.

Other Biologic Therapies and Future Directions of Adjuvant Therapy

Despite significant reduction in recurrences with adjuvant chemotherapy, hormonal therapy, and trastuzumab, many women still experience distant recurrence and ultimately die of metastatic disease. Recent molecular data including expression and genomic profiling have confirmed that breast cancer consists of distinct phenotypes, including luminal, HER-2, and basal-like subtypes. Tumors that lack expression of the estrogen receptor, progesterone receptor, and HER-2/neu (so called “triple negative” breast cancer) represent a distinct challenge. Triple negative tumors most commonly correspond to the basal-like subtype of breast cancer. Since triple negative tumors are not amenable to hormonal or HER-2-targeted therapies, they typically require chemotherapy-based treatment. Specific targets associated with basal-like or triple negative breast cancer are under investigation. The use of inhibitors of poly (ADP-ribose) polymerase in triple negative metastatic breast cancer has been reported in a randomized phase II trial of gemcitabine/carboplatin with or without PARP inhibition.²⁶⁷ In an analysis of the first 86 of the 120 patients enrolled on the trial, the addition of the PARP inhibitor resulted in a significant improvement in median progression free and overall survival. Fong et al.²⁶⁸ reported the results of a phase 1 trial evaluating olaparib (AZD2281), an oral PARP inhibitor in the treatment of patients with metastatic previously treated breast (nine patients), ovarian, prostate or other cancers of which one third had BRCA 1 or 2 mutations. Dosing ranged from 10 mg daily for 2 out of 3 weeks to 600 mg twice daily. Antitumor activity was demonstrated only in the mutation carriers and toxicity was minimal. The investigators subsequently enrolled only mutation carriers who took olaparib 200 mg/day. The PARP inhibitors represent another example of targeted therapy. The underlying mechanism of action is a synthetic lethality in which inhibition of PARP1 activity, which is responsible for base-excision repair results in the elimination of the only remaining way in which the cancer cells of mutation carriers can repair DNA damage. These cancer cells have lost all BRCA 1 or 2 function and therefore do not have the ability to repair DNA damage by homologous recombination. The result is cell death. The study by O’Shaughnessy et al.²⁶⁷ suggest that the benefit of PARP inhibitors may not be limited to mutation carriers but may include triple negative nonmutation carriers. A number of ongoing trials are evaluating the role or PARP inhibitors in the metastatic setting and ultimately the adjuvant setting.

Targeting angiogenesis is one strategy that is being explored in the adjuvant setting to improve outcomes. Bevacizumab, a monoclonal antibody to the vascular endothelial growth factor, has been approved for use in a variety of tumor types, including colorectal cancer, non-small cell lung cancer, and metastatic breast cancer. Compared with paclitaxel alone, the combination of paclitaxel and bevacizumab improved overall response rates and prolonged progression-free survival and in patients with breast cancer receiving their first line of chemotherapy for metastatic disease.²⁶⁹ Based on this trial result, bevacizumab ECOG 5103 was designed as a randomized phase III trial studying doxorubicin, cyclophosphamide, and paclitaxel with or without bevacizumab in the adjuvant treatment of patients with HER-2-negative lymph node-positive or high-risk lymph node-negative breast cancer. Also based on results in the metastatic setting, lapatinib appears to be another promising agent among patients with HER-2-positive disease. The addition of lapatinib to capecitabine extended time to progression in patients who experienced disease progression after trastuzumab-based therapy.²⁷⁰ The Adjuvant Lapatinib and/or Trastuzumab Treatment Optimisation (ALTTO) trial is an international research study that will randomize more than 8000 patients to receive trastuzumab alone for 52 weeks, lapatinib alone for 52 weeks, trastuzumab for 12 weeks, followed by a 6-week break, followed by lapatinib for 34 weeks, or lapatinib in combination with trastuzumab for 52 weeks.

Each of these trials will require substantial commitment on the part of patients to receive such prolonged courses of adjuvant treatment and will demand that providers including surgeons, medical oncologists, and radiation oncologists communicate and coordinate patient care plans to minimize risks of multiple overlapping therapies.

Treatment

Ductal carcinoma in situ (DCIS) currently represents approximately 25% of all breast cancers and 20% to 30% of those that are mammographically detected. The American Cancer Society estimates that there will be 67,770 new cases of in situ cancer of the breast in the United States in 2008, of which the majority will be DCIS.¹ Local treatment options for ductal carcinoma in situ include mastectomy with or without reconstruction and breast-conserving surgery with or without radiation. Patterns of care studies suggest that approximately two thirds of women will undergo breast-conserving surgery and half of these will receive radiation.²⁷¹ Systemic options include tamoxifen or aromatase inhibitors.

The goals of treatment for DCIS include the eradication of the initial cancer, prevention of the appearance of an invasive or noninvasive cancer in the ipsilateral breast, and the ability to minimize the risk of death from breast cancer. Endpoints commonly used in assessing the outcome of cancer treatment include survival (cause-specific, i.e., death due to cancer or overall, i.e., deaths from all causes) and recurrence (local and/or regional and distant). Mortality from DCIS is related to the presence of a clinically occult invasive cancer or the subsequent development of an ipsilateral invasive cancer, which becomes the source of distant metastases. Breast cancer mortality from DCIS is less than 5% and does not vary significantly with initial treatment (Tables 58-6 and 58-7). This lack of variation is related to the finding that 50% of ipsilateral breast tumor recurrences after breast-conserving surgery are ductal carcinoma in situ and 50% are invasive cancers. Salvage of a noninvasive recurrence is almost 100% and salvage of an invasive recurrence is approximately 80%.^272,273 Breast cancer mortality following breast-conserving surgery for DCIS is approximately 10% of the overall ipsilateral breast tumor recurrent rate (IBTR). For example, if the ipsilateral breast tumor recurrence rate was 30% following conservative surgery alone and 15% following conservative surgery and radiation, the respective breast cancer mortality rates would be 3% and 1.5%. Following mastectomy for DCIS, chest wall recurrences are usually invasive cancers and salvage is approximately 50%. Therefore, breast cancer mortality following mastectomy for DCIS is one half the local recurrence rate. Since breast cancer mortality for DCIS is low regardless of the treatment, the primary consideration is the risk of an ipsilateral breast tumor recurrence and, in particular, an invasive recurrence. However, any local recurrence is significant in that the individual patient must face cancer again and subsequent treatment may not include breast-conserving surgery. An invasive recurrence has the potential to be life threatening. Treatment decisions may reflect the individual patient’s acceptance of the risk of recurrence and the perceived benefit of treatment.

Information regarding the results of treatment for DCIS may be obtained from prospective randomized trials, single-arm perspective studies, and retrospective series. There have been no prospective randomized trials comparing mastectomy to breast-conserving surgery for DCIS. However, 76 patients with DCIS were inadvertently entered into the National Surgical Adjuvant Breast and Bowel Project (NSABP) Trial B-06 for invasive cancer.^274,275 Ipsilateral breast tumor recurrence was reported in 43% of the 21 women who underwent conservative surgery alone and 24% of these were invasive cancers. Seven percent of the women who received radiation experienced an ipsilateral breast tumor recurrence and 4% of these were invasive cancers. There were no chest wall recurrences in the women who underwent mastectomy. Similar results have been obtained from retrospective series; however, they are clearly subject to treatment selection bias. The results of six of these series are presented in Tables 58-6 and 58-7.

With median follow-up averaging 7 years, ipsilateral breast tumor recurrence rates following conservative surgery without radiation range from 16% to 42% and with radiation 9% to 22%. Chest wall recurrence rates following mastectomy range from 0.6% to 10%.

Routine sentinel node biopsy and/or axillary dissection is not recommended for women with DCIS. Axillary node metastases are infrequent in DCIS (i.e., less than 3%).²⁸¹ Sentinel node metastases were identified in 3.1% of 223 women with DCIS in a European Institute of Oncology study.²⁸² However, selective use of sentinel node biopsy has been recommended for DCIS more than 4 cm, palpable tumors, high-grade disease, and patients undergoing mastectomy. These features have been associated with an increased risk of occult invasive disease^{92,95,283,284} and therefore an increased risk of positive axillary nodes.^285–287

Five prospective randomized trials have compared breast-conserving surgery to breast-conserving surgery and radiation. Descriptions of the trials and their results are presented in Tables 58-8 and 58-9. These trials have been criticized for their lack of detailed pathologic and clinical correlation. There was considerable variation among the participating institutions in specimen processing and pathologic assessment. Central pathology review was not performed for all slides in each individual case. The trials were designed and initiated at a time when the importance of various factors (margin width, size, nuclear grade) was not well appreciated. In addition, there were variations in surgical and radiotherapy technique. Despite these limitations, the trials are fairly consistent in their results. The addition of radiation decreased the risk of an ipsilateral breast tumor recurrence by 50% to 60% with an absolute benefit of 9%-16%. Invasive recurrences were decreased by 40% to 50% with an absolute benefit of 2%-9%. There were no significant differences in survival. It should be noted that the majority of tumors were low-grade or noncomedo DCIS in the EORTC trial and the United Kingdom-Australian-New Zealand Trial was a trial primarily of women greater than 50 years of age, with only 10% less than 50. A Cochrane review²⁹⁵ of four randomized trials of breast-conserving surgery with or without radiation (3925 women) reported a 51% decrease in ipsilateral breast tumor recurrence (invasive or DCIS) with the addition of radiation. There was no increase in non-breast cancer death with radiation and overall survival was similar for the two treatments.

The rationale for breast conservation therapy for ductal carcinoma in situ was based on the demonstrated effectiveness of the conservative approach for early stage invasive cancer. However, not all women are candidates for breast-conserving surgery. The current challenge is to identify women with DCIS whose risk of an ipsilateral breast tumor recurrence (primarily invasive) with breast-conserving surgery, with or without radiation, is significant enough to prompt a recommendation for a mastectomy and to identify women whose risk of an ipsilateral breast tumor recurrence following conservative surgery alone may be low enough to avoid radiation. A number of factors have been analyzed for their ability to predict an ipsilateral breast tumor recurrence following conservative surgery with or without radiation. These include the clinical factors of patient age, family history, method of detection, size, and residual calcifications on a postsurgery mammogram. Pathologic factors include architectural pattern, nuclear grade, necrosis, tumor size, margins of resection, angiogenesis, fibrosis, lymphocyte infiltrate, calcifications, multifocality, S phase fraction, estrogen and progesterone receptor and HER-2/neu. Table 58-10 presents outcomes reported from the prospective randomized trials of conservative surgery with or without radiation related to some of these factors. The relative importance of each factor has not been established. Difficulties in assessing the size and extent of DCIS, resection margin status and discordance among pathologists for the classification of DCIS and the reproducibility of any classification system have resulted in variations in the significance of each of these factors and the magnitude of risk of an ipsilateral breast tumor recurrence.

The single factor that is not subject to interpretation is patient age. Women 40 years of age or younger with DCIS have been reported to have ipsilateral breast tumor recurrence rates of approximately 50% in retrospective series^{100,277,296–298} and the EORTC prospective randomized trial.²⁹⁰ The addition of radiation to conservative surgery in the EORTC trial resulted in an ipsilateral breast tumor recurrence rate of 23% in women 40 years of age or younger. In the French Regional Cancer Center experience, the addition of radiation to conservative surgery decreased the ipsilateral breast tumor recurrence rate from 43% to 24% in women 40 years of age or younger with a median follow-up of 7.8 years.²⁷⁷ In the Rare Cancer Network study,²⁹⁶ the addition of radiation to conservative surgery decreased the ipsilateral breast tumor recurrence rate from 54% to 14% in women 45 years of age or younger with a median follow-up of 6 years. In addition, young age has been associated with an increased risk of an invasive recurrence following conservative surgery with or without radiation.^276,299 Therefore, conservative surgery alone does not appear to be a treatment option associated with a reasonable risk of ipsilateral breast tumor recurrence in young women. An additional consideration in young women is the possibility of a BRCA1 or BRCA2 mutation. The prevalence of DCIS in mutation carriers has been reported to be similar to noncarriers. Hwang et al.²⁰⁵ and Claus et al.³⁰⁰ reported a greater percentage of high-grade DCIS in BRCA1 carriers when compared with BRCA2 carriers or sporadic breast cancer. Currently, there are no studies reporting outcome of treatment for DCIS in women with BRCA1 or BRCA2 mutations. Mutation carriers with invasive cancers have been reported to have an ipsilateral breast tumor recurrence rate at 10 years of 34% when treated with conservative surgery alone³⁰¹ compared with 13% when treated with conservative surgery and radiation. Pierce et al.³⁰² noted that the addition of oophorectomy or tamoxifen to conservative surgery and radiation resulted in a decreased risk of ipsilateral breast tumor recurrence and contralateral breast cancer. Treatment options, therefore, for women with BRCA1- or BRCA2-linked DCIS include unilateral mastectomy versus bilateral mastectomy or conservative surgery and radiation, combined with tamoxifen or oophorectomy.

Most of the prospective randomized trials (see Table 58-10) suggest that increasing age is associated with a decreased risk of ipsilateral breast tumor recurrence in patients treated with conservative surgery alone or conservative surgery and radiation. However, for each of the age groups, the addition of radiation decreased the risk of an ipsilateral breast tumor recurrence when compared with conservative surgery alone. In the South Sweden trial,²⁹¹ the benefit of radiation appeared greatest in women over 50 years of age. In a recent analysis of the Surveillance Epidemiology and End Results (SEER)³⁰³ Medicare database of women 66 years of age or older with DCIS, the addition of radiation decreased the 8-year ipsilateral breast tumor recurrence rate from 11% to 3% and the 8-year risk of an ipsilateral invasive recurrence from 7% to 2%. A low-risk group was identified, being age 70 years or older, tumor size less than 2.5 cm, noncomedo histology and nuclear G₂. In this low risk group, the risk of an ipsilateral breast tumor recurrence at 5 years was 8% without radiation compared with 1% with radiation. Therefore, even in the most favorable subgroup based on age, the addition of radiation decreases the risk of an ipsilateral breast tumor recurrence, although the magnitude of the benefit may be small.

In each of the randomized trials, the detection of DCIS solely by mammography was associated with a lower risk of ipsilateral breast tumor recurrence when compared with clinical detection, with symptoms such as a palpable mass or bloody nipple discharge (see Table 58-10). An emerging area of interest is breast density and its association with an increased risk of ipsilateral breast tumor recurrence in women with DCIS. Habel et al.³⁰⁴ reviewed the mammograms of 392 of the 818 patients entered into the NSABP B-17 trial. The median follow-up was 11 years. Thirty-three percent of the women were 50 years of age or younger and 52% received radiation. A highly dense breast was seen on mammography (highly dense ≥ 75%) in 15.6% of the women less than 50 years of age compared with 2% of the women over 50. Women with mammographic density greater than or equal to 75% had a 40% ipsilateral breast tumor recurrence rate compared with 18% for mammographic density less than or equal to 25%, 27% for mammographic density 25% to 49% and 24% for mammographic density 50% to 74%.

The extent of DCIS impacts on the decision for breast-conserving surgery versus mastectomy. The extent can be assessed radiographically or by pathologic examination. Radiographic assessment includes the extent of calcifications on a mammogram; however, this may underestimate the pathologic extent of DCIS, especially for low-grade lesions.^87,305 More recently, MRI has been used to evaluate the extent of DCIS. In a study of 45 patients from the University of California, San Francisco, MRI overestimated the pathologic extent of disease more than twofold in 23% of patients and underestimated it in 9%.⁹³ The discrepancy between the pathologic size and MRI extent was more frequently observed in the MRI patterns of regional or multi-regional enhancement and in estrogen-receptor-positive DCIS. Clinical assessment of tumor size includes measurements of a palpable mass, the dimensions of the mammographic abnormalities, including calcifications and/or a mass. In the EORTC²⁹⁰ and South Sweden²⁹¹ prospective randomized trials, increasing clinical size was associated with an increased risk of ipsilateral breast tumor recurrence in patients treated with conservative surgery alone but not those treated with conservative surgery and radiation. In the NSABP B-17 randomized trial,²⁸⁸ the ipsilateral breast tumor recurrence rate was correlated with the extent of calcifications on the mammogram, both in women treated with conservative surgery alone or conservative surgery and radiation. It is of note that the appearance of scattered calcifications was associated with an ipsilateral breast tumor recurrence rate of 37% in patients treated with conservative surgery alone and 27% in those treated with conservative surgery and radiation.

In patients having calcifications, a postexcision mammogram before radiation or observation is essential to assure the removal of all malignant-appearing calcifications. If calcifications are not identified on routine views, magnification views of the surgical site should be performed. A specimen x-ray is not adequate proof that calcifications have been removed. All of the five patients reported in the literature^306,307 in whom residual malignant calcifications were not removed before radiation have reportedly recurred. Standards for the management of ductal carcinoma in situ have been developed by the American College of Radiology, the Society of Surgical Oncology, the American College of Surgeons, and the College of the American Pathologists³⁰⁸ and emphasize the importance of obtaining a postexcision mammogram.

The pathologic extent of DCIS can be assessed by gross examination, although most DCIS is not grossly evident. It can be determined by measuring the area on a given slide, the number of blocks with DCIS, or complete tissue processing with uniform section thickness and sequential analysis as performed by Silverstein et al. This latter method is not commonly employed in clinical practice. In the NSABP B-17 and B-24 trials,^288,293,309 size of DCIS as measured by the area on a given slide did not significantly correlate with ipsilateral breast tumor recurrence rates in patients treated with conservative surgery or conservative surgery and radiation. In a single-arm prospective study of conservative surgery alone,¹⁰² there was no correlation with ipsilateral breast tumor recurrence rates when size and extent were measured by the number of blocks with DCIS. MacDonald et al.³¹⁰ reported 8-year ipsilateral breast tumor recurrence rates of 14% for DCIS less than or equal to 1.5 cm treated with conservative surgery alone compared with 23% for those with DCIS more than 1.6 to 3 cm and 30% for those greater than 3 cm treated with conservative surgery alone (median follow-up 4.8 years). For patients undergoing conservative surgery and radiation, the ipsilateral breast tumor recurrence rate was 13% for DCIS less than or equal to 1.5 cm, 28% for DCIS 1.6 to 4 cm, and 32% for DCIS more than 4 cm, with a median follow-up at 8.6 years.³¹¹ In these series, pathologic size was assessed by detailed processing and analysis of sequential sections. The size was then computed by knowing the number of sections with DCIS, the location of the sections and the width of the blocks. For example, if DCIS appeared on 10 consecutive sections and the blocks were 3 mm in thickness, the size of the DCIS would be 3 cm.

Multifocal DCIS has been defined as multiple foci separated by intervening normal breast tissue. Its presence has been associated with an increased risk of ipsilateral breast tumor recurrence in the NSABP B-17,²⁸⁹ B-24,²⁹² and South Sweden²⁹¹ prospective randomized trials when compared with unifocal disease in patients treated with conservative surgery alone or conservative surgery and radiation (see Table 58-10). The presence of multifocal disease may be a contraindication to conservative surgery alone. In a retrospective series reported from the University of Toronto,³¹² the 10-year ipsilateral breast tumor recurrence rate for women undergoing conservative surgery alone was 20% for unifocal disease compared with 41% for multifocal DCIS. For women treated with conservative surgery and radiation, ipsilateral breast tumor recurrence rates at 10 years were 13% for unifocal disease compared with 20% for multifocal disease. In the NSABP B-24 trial, the presence of multifocal DCIS was a significant factor predicting for ipsilateral breast tumor recurrence rates in patients treated with conservative surgery and radiation with or without tamoxifen.²⁹²

A recent study from Memorial Sloan-Kettering Cancer Center reported an increased risk of IBTR in women with DCIS and atypical lobular hyperplasia (ALH) or LCIS undergoing breast-conserving surgery.³¹³ Fourteen percent of the 294 patients had associated ALH or LCIS. Eighty percent did not receive radiation and only 10% had tamoxifen. The 10-year cumulative incidence of an IBTR with LCIS was 47% and 21% for ALH compared with 19% for those with DCIS only. The presence of lobular neoplasia at an excision margin was associated with a 46% IBTR rate (12/26) compared with 25% (3/12) without lobular neoplasia at an excision margin. However, two additional series have not reported an increased risk of IBTR in women with DCIS and LCIS undergoing breast-conserving surgery.^314,315

Based on the above considerations of size and extent of DCIS, mastectomy is the recommended treatment for extensive areas of ductal carcinoma in situ (i.e., those manifesting as diffuse areas of malignant appearing calcification or evidence of DCIS in more than one quadrant in a single breast). Women with DCIS extending over a distance of more than 4 cm generally are not considered for breast conservation therapy, based on the outcome data from Silverstein et al.³¹¹

Resection margin status is a complex issue. Positive margins of resection have been associated with an increased risk of ipsilateral breast tumor recurrence in the NSABP B-17, EORTC, and the NSABP B-24 trials (see Table 58-10) in patients undergoing conservative surgery or conservative surgery and radiation. However, the magnitude of the risk of ipsilateral breast tumor recurrence in patients with positive margins has varied and reflects difficulties in assessing margins in the absence of a standardized method for assessing and reporting margins, especially in terms of the width of a negative margin. Margins cannot be assessed if the specimen is submitted in fragments, if it is not adequately sampled, if it is not oriented by the surgeon, and if it is not inked. There are six surgical margins (anterior, posterior, superior, inferior, medial, and lateral). The location of the positive margin influences recommendations for further surgery. A positive posterior margin usually does not require further surgery if the initial excision extended down to the pectoralis fascia. Similarly, a positive anterior margin does not require additional surgery if the excision extended to the skin. Ipsilateral breast tumor recurrence rates may be influenced by the extent of margin positivity and the number of positive margins and there appears to be an interaction between patient age and margin positivity. In a multi-institution study of breast-conserving surgery and radiation for mammographically detected DCIS, Solin et al.³¹⁶ reported ipsilateral breast tumor recurrence rates of 50% for women 39 years of age or younger with close margins compared with 7% for women 50 years of age or older. Women with positive margins of resection 50 years of age or older experienced a 13% ipsilateral breast tumor recurrence rate compared with 24% for those 40 to 49 years of age. The width of the negative margin that best minimizes the risk of an ipsilateral breast tumor recurrence following conservative surgery or conservative surgery and radiation is controversial. A recent report from the NSABP B-24 trial²⁹² reported no differences in ipsilateral breast tumor recurrence rates when comparing a negative margin width of less than or equal to 1 mm to more than 1 mm in patients undergoing conservative surgery and radiation with or without tamoxifen. However, in a series from Tufts New England Medical Center, negative margins more than 10 mm were associated with a 0% risk of an ipsilateral breast tumor recurrence in patients receiving conservative surgery and radiation.³¹⁷ Solin et al.³¹⁶ reported 10-year ipsilateral breast tumor recurrence rates of 8% for women with DCIS and more than 2 mm negative margins treated with conservative surgery and radiation. Dunne et al.³¹⁸ performed a meta-analysis of 21 retrospective studies and one randomized trial of DCIS in which the status of the excision margin was reported. The meta-analysis included 4660 women treated with breast-conserving surgery and radiation. Negative margins greater than or equal to 2 mm were associated with a decreased risk of ipsilateral breast tumor recurrence when compared with those less than 2 mm. Silverstein et al. have suggested that a 1 cm or greater negative margin is sufficient to prevent radiation in women undergoing breast-conserving surgery. The 8-year actuarial risk of ipsilateral breast tumor recurrence in 197 women who underwent conservative surgery alone with greater than or equal to 1 cm negative margins was 8%.³¹⁰ Margins less than 1 cm were associated with ipsilateral breast tumor recurrence rates ranging from 61% for a positive margin to 22% to 51% for negative margins in the less than 1 mm to 9.9 mm range. In a retrospective multi-institutional study of wide excision alone for DCIS, Kerlikowske et al.²⁹⁹ reported a 9%, 5-year ipsilateral breast tumor recurrence rate for women with negative margins greater than or equal to 1 cm. As width of the negative margin increased, the risk of an ipsilateral breast tumor recurrence with DCIS decreased but the risk of an ipsilateral breast tumor recurrence with an invasive cancer was not significantly changed. Therefore, negative margins of resection may have a greater impact on diminishing the risk of a noninvasive ipsilateral breast tumor recurrence rather than an invasive recurrence.

High nuclear grade and the presence of necrosis have been associated with an increased risk of ipsilateral breast tumor recurrence in patients undergoing conservative surgery in the NSABP B-17^288,289 and EORTC²⁹⁰ randomized trials (see Table 58-10). These factors have had less of an impact on ipsilateral breast tumor recurrence rates in patients undergoing conservative surgery and radiation. In the NSABP B-24 trial, high nuclear grade was not associated with an increased risk of ipsilateral breast tumor recurrence in patients undergoing conservative surgery and radiation with or without tamoxifen. However, the presence of a micropapillary pattern of DCIS and comedo necrosis were significant factors associated with an increased risk of ipsilateral breast tumor recurrence in these women.²⁹²

Silverstein et al.,¹⁰⁰ employing a multivariable analysis of prognostic factors for ipsilateral breast tumor recurrence in patients with DCIS treated in their practice, identified age, size of margins, nuclear grade, and necrosis as the most significant factors. Their work was based on detailed clinical (mammographic and operative) findings and pathologic correlation and required processing of the entire surgical specimen with a uniform thickness of 2 to 3 mm and sequential analysis, thereby providing the most accurate assessment of size and margin status. This analysis formed the basis for the University of Southern California/Van Nuys Prognostic Index (USC-VPNI), which is a scoring system with four categories: age, size, margins, and nuclear grade combined with necrosis. Within each category there are three scores (1, 2, 3). The age scores are 1 = more than 60 years, 2 = 40 to 60 years, and 3 = less than 40 years. The size scores are 1 = less than or equal to 1.5 cm, 2 = 1.6 to 4 cm, 3 = more than 4 cm. The margin scores are 1 = greater than or equal to 1 cm, 2 = 1 to 9 mm and 3 = less than 1 mm. The pathologic scores are 1= low grade and no necrosis, 2 = low grade with necrosis, and 3 = high grade. Summing each of the category scores results in total scores ranging from 4 to 12. The 12-year actuarial ipsilateral breast tumor recurrence rates were 4% for scores of 4 to 6, 40% for scores of 7 to 9, and 87% for scores of 10 to 12 in patients treated with conservative surgery alone, and 4% for scores of 4 to 6, 30% for scores 7 to 9, and 38% for scores of 10 to 12 in patients treated with conservative surgery and radiation.¹⁰⁰ The investigators concluded that patients with scores of 4 to 6 were candidates for wide excision alone; scores of 7 to 9, for excision and radiation; and scores of 10 to 12, mastectomy. Recognizing that their degree of pathologic processing and analysis was not routinely performed in clinical practice and that there was discordance among the pathologists in the classification of DCIS, the investigators ultimately concluded that margin status was the most important factor for ipsilateral breast tumor recurrence and represented a surrogate for the USC-VNPI index. The reproducibility of this system has been questioned by a number of investigators in retrospective and prospective studies.^101–104 Inaccuracies in calculating the score could result in overtreatment or undertreatment. Use of margin width as the sole prognostic factor may be an oversimplification of a complex issue especially in young women.

The relationship between the resection margin status and chest wall recurrences in patients with DCIS undergoing mastectomy has received little attention. Rashtian et al.³¹⁹ reported a 7.5% local recurrence rate in 80 women with DCIS and a less than 10 mm mastectomy margin. The median follow-up was 6.1 years. All six local recurrences were invasive. Local recurrence was 12% for less than 60 years compared with 0% for women 60 years old. Margin status of 2 mm with comedonecrosis or nuclear G₃ DCIS was associated with a 25% local recurrence. Local recurrence was 6% for total mastectomy compared with 11% for modified radical mastectomy. Carlson et al.³²⁰ reported a 3.3% recurrence rate in 223 women who had undergone skin-sparing mastectomy and immediate reconstruction. The median follow-up was 6.8 years. The local recurrence rate was 11% for margins of 1 mm, 6% for age 50 years old, 6% for nuclear G₃ DCIS, and 6% for DCIS of 4 cm. In contrast, Chan et al.³²¹ reported a local recurrence rate of 1.7% in 59 women with DCIS and a 5 mm mastectomy margin. Median follow-up was 8 years (range 1 to 21 years) and the median pathologic size of DCIS was 4.5 cm. The local recurrence rate was 3% for high-grade DCIS and 5% for skin-sparing mastectomy. None of the four patterns with a positive margin experienced a local recurrence. There were no chest wall recurrences in patients who underwent total mastectomy or total skin-sparing mastectomy. The single chest wall recurrence occurred in of the 14 patients who underwent skin-sparing mastectomy. The differences in the reported outcome in these series may reflect variations in surgical technique. Outcome of empiric treatment with radiation postmastectomy of 14 patients with DCIS with close or positive margins (1 mm) has been reported in a single series.³²² There were no local recurrences with a median follow-up of 4.8 years.

Two prospective single-arm studies were conducted in an attempt to identify a low-risk group of women for whom conservative surgery alone may be an acceptable treatment. The Dana Farber Cancer Center¹⁰² initiated a single-arm prospective study of conservative surgery alone for women with DCIS less than or equal to 2.5 cm (mammographic size), nuclear G₁-G₂, and margins greater than or equal to 1 cm. The study was terminated after 158 patients were enrolled because the observed ipsilateral breast tumor recurrence rate had exceeded the expected. With a median follow-up of 3.3 years, the crude ipsilateral breast tumor recurrence rate was 9% and the 5-year actuarial rate was 12.5% with a projected rate of 25% at 10 years and 30% at 12 years. The investigators were unable to duplicate the low ipsilateral breast tumor recurrence rate (8% at 8 years) reported by MacDonald et al.³²³ in a somewhat similar group of 212 with greater than or equal to 1 cm margins and no radiation. Not all of the patients in the Harvard study were required to have a postexcision mammogram to exclude residual calcifications and only 38% of the patients actually had one. It is possible that some of the recurrences were persistent disease, which was not completely resected. In addition, 40% of the 10 patients with G₃ tumors experienced an ipsilateral breast tumor recurrence compared with 6% of the G₁-G₂ tumors. The Eastern Cooperative Oncology Group³²⁴ completed a single-arm prospective study of conservative surgery alone with or without tamoxifen for nuclear G₁ or G₂ DCIS less than or equal to 2.5 cm, with margins of resection greater than or equal to 3 mm and DCIS less than or equal to 1 cm nuclear G₃ with necrosis and with margins greater than or equal to 3 mm. Size was defined as the largest dimension of the area of the DCIS histologically. The median follow-up was 6.2 years. For the 565 low-grade lesions, median lesion size was only 6 mm (only 24% >1 cm) and surgical margins were 5 to 10 mm in 21% and more than 1 cm in 49%. At 5 years the ipsilateral breast tumor recurrence rate was 6.1%, with 53% of these recurrences being invasive cancers. The contralateral breast cancer rate was 3.7%. For the 102 high-grade lesions, median lesion size was 5 mm with surgical margins 5 to 10 mm in 30% and more than 1 cm in 53%. The ipsilateral breast tumor recurrence rate at 5 years was 15.3% and 35% were invasive cancers. The contralateral breast cancer rate was 3.9%. For the low-grade lesions, ipsilateral breast tumor recurrence rates were not significantly influenced by age, lesion size, or margin widths. For high-grade DCIS, the ipsilateral breast tumor recurrence rate for the 11 women less than 45 years of age was 54.4% compared with 10.3% for those 45 years of age or older. Margin width greater than or equal to 1 cm was not associated with a decreased ipsilateral breast tumor recurrence rate (IBTR). The 5-year IBTR rate for 1-cm high-grade DCIS was 32.9% compared with 12.7% for size less than 1 cm. Therefore both the Dana Farber and the ECOG studies suggest that high-grade DCIS, even with small size and widely negative margins, may be associated with an ipsilateral breast tumor recurrence rate sufficient enough to recommend radiation after local excision.

The role of tamoxifen in the treatment of ductal carcinoma in situ has been evaluated in two prospective randomized trials (see Tables 58-8 and 58-9). In the NSABP B-24 trial, women with DCIS were randomized to radiation with or without tamoxifen. In contrast to the NSABP B-17 trial, the B-24 trial included patients with positive margins (16%) and diffuse disease.²⁹³ The United Kingdom Coordinating Committee on Cancer Research DCIS Working Group (UK/ANZ) trial randomized patients to conservative surgery with or without tamoxifen or conservative surgery and radiation with or without tamoxifen.²⁹⁴ The addition of tamoxifen to radiation significantly decreased ipsilateral breast tumor recurrence rates in the NSABP B-24 trial but not in the UK/ANZ trial. The addition of tamoxifen to conservative surgery alone had no significant impact on ipsilateral breast tumor recurrence rates in the UK/ANZ trial. Tamoxifen was associated with a reduction in contralateral breast cancers in both trials and in both trials women 50 years of age or younger appeared to have a greater benefit from tamoxifen. In the NSABP B-24 trial, 33% of the women were 50 years of age or younger compared with only 10% in the UK/ANZ trial. This difference in the age of the patient populations may account in part for the different results. More recently, the NSABP B-24 investigators have reported that the benefit of tamoxifen was limited to ER-positive DCIS.³²⁵ This finding has prompted the routine performance of estrogen and progesterone receptors in DCIS. The NSABP B-35 trial evaluates the role of aromatase inhibitors in the treatment of ductal carcinoma in situ. The trial randomizes postmenopausal women with DCIS and negative margins and estrogen and progesterone receptor-positive DCIS by immunohistochemistry to radiation and tamoxifen for 5 years or radiation and anastrozole for 5 years.

Several investigators have initiated trials of neoadjuvant therapy for DCIS. Hwang and Esserman³²⁶ at UCSF are evaluating neoadjuvant tamoxifen. Preliminary results suggest that response as measured by MRI and the subsequent excision pathology appears greater in ER-positive disease. Neoadjuvant lapatinib for HER-2/neu or epidermal growth factor receptor-positive DCIS is being evaluated at Baylor College of Medicine³²⁷ and neoadjuvant trastuzumab is under investigation at MD Anderson Cancer Center. An NSABP trial will randomize patients with HER-2/neu overexpressing DCIS with negative excision margins to radiation with or without concurrent trastuzumab. HER-2/neu overexpression is more common in high-grade DCIS or comedo DCIS and ER-negative DCIS.^328,329

Given the heterogeneity of DCIS, it is unlikely that a single treatment will be appropriate for all women. The management of DCIS requires a multidisciplinary approach with careful and detailed clinical and pathologic correlation. DCIS represents a spectrum of disease. Local treatment options include mastectomy with or without reconstruction, or breast-conserving surgery with or without radiation. For appropriately selected patients, differences in local treatment result in minor differences in breast cancer survival. For inappropriately selected patients, differences in local treatment significantly impact on recurrence and ultimately breast cancer survival. Mastectomy is the recommended treatment in patients with multicentric DCIS, (i.e., two or more separate quadrants of the breast) and those with extensive or diffuse malignant-appearing calcifications. Patients in whom negative margins of resection cannot be obtained after reasonable surgical attempts are advised to undergo mastectomy. Young women, (i.e., 40 years of age or younger) and those with multifocal DCIS have a significantly increased risk of ipsilateral breast tumor recurrence when treated with conservative surgery alone. Treatment options in these women include conservative surgery and radiation or mastectomy. Women with BRCA1– or BRCA2-linked DCIS may elect bilateral mastectomy given their increased risk of contralateral breast cancer and the increased risk of an ipsilateral breast tumor recurrence, which has been reported with invasive cancers treated with conservative surgery and radiation without tamoxifen or oophorectomy. For patients undergoing breast-conserving surgery, negative margins minimize the risk of an ipsilateral breast tumor recurrence, although the width of the negative margin that best achieves this goal remains controversial. Nuclear grade appears more important in predicting the risk of an ipsilateral breast tumor recurrence in patients undergoing conservative surgery alone. The presence of high nuclear grade or necrosis may prompt a recommendation for postoperative radiation. The addition of radiation to conservative surgery has not resulted in a significant increase in second non-breast cancer malignancies in the NSABP B-17 or B-24 trial, nor an increased risk in cardiovascular death in the EORTC or NSABP B-17 trials.^288,290,293 A Cochrane review of four randomized trials of breast-conserving surgery with or without radiation in 3925 women reported no significant long-term toxicity from radiotherapy.³³⁰

Paget disease of the nipple has traditionally been treated by mastectomy. However, the demonstrated success of breast-conserving surgery in the treatment of ductal carcinoma in situ prompted interest in this treatment for Paget disease. There are no randomized trials comparing breast-conserving surgery to mastectomy for Paget disease and because the overall incidence of Paget disease is low, the reported series have small numbers of patients. Ipsilateral breast tumor recurrence rates following conservative surgery alone for Paget disease were reported by Polgar et al.³³¹ in a series of 33 women with Paget disease, with or without an underlying ductal carcinoma in situ. With a median follow-up of 6 years, the crude local recurrence rate was 33% in this group of women. Sixty-one patients were treated with excision and radiation and the crude local recurrence rate was 7%. Bijker et al.³³² reported a 5-year local recurrence rate of 5% in 61 patients registered in the European Organization for Research and Treatment of Cancer Study 10873. Surgery consisted of complete excision of the nipple-areolar complex with negative margins, followed by radiation to the entire breast, consisting of 5000 cGy in 25 fractions. The authors concluded that breast-conserving surgery with radiation was associated with a low ipsilateral breast tumor recurrence rate at 5 years in patients with Paget disease with limited extent of underlying DCIS. Complete excision of the nipple-areolar complex was recommended. Marshall et al.³³³ reported results of 38 patients with Paget disease having a palpable mass or mammographic density in a multi-institution study of surgery consisting of complete or partial excision of the nipple-areolar complex followed by postoperative radiation to the entire breast, consisting of 5000 CGy, with a majority of patients receiving an additional boost to the tumor bed. The median follow-up was 9.4 years and the crude ipsilateral breast tumor recurrence rate was 11%. The local control rate at 10 and 15 years was 87%. These studies suggest that selected patients with Paget disease may be candidates for breast-conserving surgery with radiation and that excision of the nipple-areolar complex is recommended with negative microscopic margins. The removal of all malignant-appearing calcifications should be documented on a postsurgery mammogram before the initiation of radiation.

Breast-Conserving Surgery and Radiation versus Mastectomy

Six prospective randomized trials have compared conservative surgery and radiation with mastectomy for the treatment of stage I-II breast cancer. The published results of these trials are presented in Table 58-11. The trials have demonstrated no significant difference in terms of local regional recurrence, distant metastases, or long-term survival. In particular, patients with histologically positive axillary nodes have not been found to have an improved survival when treated with mastectomy. Both the NSABP B-06 and Milan I trials demonstrated a somewhat more favorable outcome in patients with positive nodes treated with breast-conserving surgery and radiation, although the differences were not statistically significant.^342,343

The Early Breast Cancer Trialists’ Collaborative Group (EBCTG)²⁶ reported a meta-analysis of seven prospective randomized trials comparing mastectomy and axillary dissection versus breast-conserving surgery, axillary dissection, and radiation. For axillary node-negative patients, the 5-year actuarial risk of an isolated local recurrence in mastectomy patients was 5.3% compared with 8.6% in the breast-conserving surgery arm. For node-positive patients, the 5-year actuarial risk of an isolated local recurrence was 7.9% in the mastectomy patients compared with 4.7% in patients undergoing breast-conserving surgery. Breast cancer mortality at 15 years for the node-negative patients was 26% with mastectomy and 27.3% for breast-conserving surgery and radiation. For the node-positive patients, breast cancer mortality was 47.4% with mastectomy and 49.3% with breast-conserving surgery and radiation. These differences were not statistically significant. Overall survival of the node-negative patients with mastectomy was 65.6% compared with 64.5% with breast-conserving surgery and radiation. For the node-positive patients, overall survival at 15 years was 46.7% with mastectomy and 46.2% for breast-conserving surgery and radiation. Therefore, the equivalence of breast-conserving surgery with radiation to mastectomy in terms of local-regional recurrence and overall survival in appropriately selected patients with stage I-II breast cancer has been demonstrated by the randomized trials with results now out to 20 years.

Patients may not be considered candidates for breast-conserving surgery and radiation for one or more of the following reasons: unacceptable cosmetic result, a significant risk of complications, or a substantial risk of ipsilateral breast tumor recurrence, which significantly exceeds the risk of a local recurrence following mastectomy.

Risk of Ipsilateral Breast Tumor Recurrence

In the majority of prospective randomized trials comparing conservative surgery and radiation to mastectomy, the incidence of local-regional recurrence has been almost identical for the two treatments (see Table 58-11).

A number of factors have been analyzed for their ability to predict for an ipsilateral breast tumor recurrence following conservative surgery and radiation, including clinical factors (patient age, tumor size, tumor location, gross multicentricity, or mammographic evidence of diffuse malignant-appearing calcifications), histopathologic factors (histologic and nuclear grade, tumor necrosis, vascular or lymphatic invasion, inflammatory infiltrate, the presence of DCIS in association with an invasive cancer, margins of resection, histology other than invasive ductal cancer, and pathologic nodal status), and treatment-related factors (the extent of surgical resection, radiation technique, and the use of adjuvant systemic therapy). Ipsilateral breast tumor recurrences following conservative surgery and radiation are most likely due to the presence of a significant residual tumor burden that cannot be controlled with moderate doses of radiation. While the magnitude of the risk of an ipsilateral breast tumor recurrence that would prompt a recommendation for a mastectomy may vary with the patient’s and/or physician’s preference or perspective, the overall risk at 5 to 10 years can be divided into three general categories: low risk (10%), intermediate risk (10% to 20%), and high risk (25% to 40%). For high-risk patients, mastectomy is generally the recommended treatment. For moderate-risk patients, modifications in treatment (wider surgical resections, chemotherapy) may potentially decrease the risk. For some patients with a moderate risk, there may be no identifiable way of decreasing the risk; therefore, these patients would be at greater risk for salvage mastectomy. A low-risk category implies no modification in treatment recommendations or follow-up.

IBTR and Molecular Subtypes and Genomic Profiles

Genomic profiling of breast cancers has been used to predict relapse (distant and local-regional) in addition to the standard clinical and pathologic factors. The 21-gene recurrence score (Oncotype DX, Genomics Health Inc., Redwood City, Calif.) includes 16 cancer-related genes (proliferation, estrogen, invasion, HER-2, GSTM 1, CD68, and BAG I and 5 reference genes). The score has been correlated with the risk of distant metastasis in node-negative, ER-positive women treated with tamoxifen in the NSABP trials.²²¹ More recently the score has been correlated with ipsilateral breast tumor recurrence in these women treated with conservative surgery and radiation. Ipsilateral breast tumor recurrence rates more than 25% have been reported in women less than 50 years of age with intermediate to high-risk score, (i.e., 18%).⁴²⁸ However, Nuyten et al.⁴²⁹ reported no correlation with the 70-gene prognosis profile (proliferation, invasion, metastases, stromal integrity, angiogenesis) and ipsilateral breast tumor recurrence rates in 161 women less than 53 years of age with invasive cancers treated with conservative surgery and radiation. The median follow-up was 7.7 years. A wound response signature (65 activated wound response and 96 quiescent wound response) correlated significantly with ipsilateral breast tumor recurrence. Patients with a high-risk wound response signature had a 29% ipsilateral breast tumor recurrence at 10 years compared with a 5% risk for a low-risk signature. In a subsequent study,⁴³⁰ the investigators compared gene expression profiles of 56 primary invasive cancer patients, who developed an IBTR within 10 years after breast-conserving surgery and radiation, to the profiles of 109 tumors from patients without a recurrence. Twenty-seven percent of the cases had positive margins compared with 17% of the controls. All patients were less than 51 years of age. Gene expression profiles examined included the 70-gene and the 76-gene profiles, wound response, hypoxia, 52-gene radioresistance, 45-gene proliferation, and the Sweden 81 gene. Young age was the only significant factor predicting IBTR in multivariate analysis. There was an association between IBTR and the 45-gene proliferation profile, which was not captured by histologic grading. Further studies are required to assess the role of gene expression signatures in predicting ipsilateral breast tumor recurrence.

Recent interest has also focused on breast cancer intrinsic subtypes identified by immuno-histochemistry with clinical approximation by receptor status. Triple negative breast cancers have not been found to have a significantly increased risk of ipsilateral breast tumor recurrence when treated with conservative surgery and radiation in three of five published studies.^{401,431–434} However, triple negative breast cancers have a shorter interval to local failure.⁴³³ HER-2⁺, estrogen receptor-negative patients have been noted to have an increased risk of local-regional failure in some series when treated with conservative surgery and radiation⁴³² or postmastectomy radiation without trastuzumab.⁴³⁵ In an analysis of the Danish Breast Cancer Group’s studies 82b and c, HER-2/neu positive, estrogen receptor-and progesterone receptor-negative patients had a 15-year local-regional failure rate of 21% with postmastectomy radiation compared with 3% in ER+ HER-2 negative or ER+, HER-2-positive patients. The 15-year local-regional failure rate was 15% for the triple negative patients. In contrast to these series, Harris et al.⁴³⁶ reported no correlation with HER-2 status and local-regional recurrence in women treated with conservative surgery and radiation. It should be noted that none of these patients received trastuzumab, and local-regional recurrence has not been correlated with HER-2 status in women treated with conservative surgery and radiation. Buchholz et al.⁴³⁷ reported a similar lack of correlation in patients treated with mastectomy and radiation.

Integration of Multiple Factors Associated With IBTR

Of the multiple factors that may influence ipsilateral breast tumor recurrence rates, age appears to be the most important. In a recursive partitioning analysis of ipsilateral breast tumor recurrence following conservative surgery and radiation for stage I-II breast cancer, Freedman et al.¹²² identified the first category as age, 55 versus more than 55. For women more than 55, ipsilateral breast tumor recurrence was influenced by the addition of tamoxifen. For women less than 55 years of age, EIC was the most important factor. For EIC-negative patients, age 35 and margin status were the most important factors. For age 36-55, estrogen receptor status and the addition of tamoxifen were most important. This analysis noted that margin status was less important in women more than 55 years of age and it is possible that an older woman with a focally positive margin and the absence of an extensive intraductal component or invasive lobular cancer may not require re-excision before the initiation of radiation.

IBTR! is a web-based tool developed by investigators at Tufts University to estimate the risk of an ipsilateral beast tumor recurrence following breast-conserving surgery with or without radiation.⁴³⁸ The nomogram includes seven prognostic factors (age, tumor size, tumor grade, lymphovascular invasion, margin status, and the use of chemotherapy and/or hormonal therapy) and information to calculate the risk was obtained from nine prospective randomized trials of breast-conserving surgery with or without radiation, the meta-analysis of the Early Breast Cancer Trialists’ Collaborative Group, and selected single institution retrospective studies. It should be noted that data for certain factors were not available in the randomized trials. Databases from the Massachusetts General Hospital (MGH) and British Columbia Cancer Agency have been used to validate the nomogram primarily in patients receiving radiation.^439,440 IBTR! was relatively accurate in predicting ipsilateral breast tumor recurrence rates in women with favorable factors associated with a low risk. However, in both data sets, the model overestimated the risk in women with unfavorable factors, (i.e., young age, positive margins). Its clinical usefulness will require further evaluation.

In summary, women with stage I-II breast cancer who are candidates for conservative surgery and radiation include those who have primary invasive cancers less than 4 to 5 cm, that are unicentric, and in whom negative margins of resection can be achieved without significant volume loss of the affected breast. Prior history of collagen-vascular disease, pregnancy, or prior radiation may be associated with a significant risk of complications. The role of radiation in patients with germline mutations requires further evaluation, especially in terms of the risk for radiation-induced second malignancies. Women who appear to be at greatest risk for ipsilateral breast tumor recurrence following conservative surgery and radiation include those who have gross multicentric disease, EIC-positive tumors in whom negative margins of resection are not obtained, and young women. Treatment modifications that may diminish the risk of an ipsilateral breast tumor recurrence include wider surgical excisions, the use of a boost dose of radiation in young women, and the addition of adjuvant systemic therapy, with chemotherapy and/or hormonal therapy. However, systemic therapy or higher doses of radiation do not appear ultimately to diminish the risk of an ipsilateral breast tumor recurrence in women with close or positive resection margins.