J Clin Oncol 22: by American Society of Clinical Oncology INTRODUCTION. Women who carry a germline mutation in

VOLUME 22 NUMBER 12 JUNE 15 2004 JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Contralateral Breast Cancer in BRCA1 and BRCA2 Mutation Carriers Kelly Metcalfe, Henry T. Lynch, Parviz Ghadirian, Nadine Tung, Ivo Olivotto, Ellen Warner, Olufunmilayo I. Olopade, Andrea Eisen, Barbara Weber, Jane McLennan, Ping Sun, William D. Foulkes, and Steven A. Narod From the Faculty of Nursing and the Centre for Research on Women s Health, and Sunnybrook and Women s College Health Sciences Centre, University of Toronto; Toronto Sunnybrook Regional Cancer Centre, Toronto; Epidemiology Research Unit, Centre Hospitalier de Université de Montreal (CHUM); Program in Cancer Genetics, McGill University, Montreal; British Columbia Cancer Agency, Vancouver; Hamilton Regional Cancer Centre, Hamilton, Canada; Department of Preventive Medicine and Public Health, Creighton University School of Medicine, Omaha, NE; Beth Israel Deaconess Medical Center, Boston, MA; Department of Medicine, University of Chicago, Chicago, IL; Departments of Medicine and Genetics, Department of Hematology/ Oncology, University of Pennsylvania, Philadelphia, PA; Cancer Risk Program, UCSF Comprehensive Cancer Center, San Francisco, CA. Submitted April 3, 2003; accepted March 24, 2004. A B S T R A C T Purpose To estimate the risk of contralateral breast cancer in BRCA1 and BRCA2 carriers after diagnosis and to determine which factors are predictive of the risk of a second primary breast cancer. Patients and Methods Patients included 491 women with stage I or stage II breast cancer, for whom a BRCA1 or BRCA2 mutation had been identified in the family. Patients were followed from the initial diagnosis of cancer until contralateral mastectomy, contralateral breast cancer, death, or last follow-up. Results The actuarial risk of contralateral breast cancer was 29.5% at 10 years. Factors that were predictive of a reduced risk were the presence of a BRCA2 mutation (v BRCA1 mutation; hazard ratio [HR], 0.73; 95% CI, 0.47 to 1.15); age 50 years or older at first diagnosis (v 49 years; HR, 0.63; 95% CI, 0.36 to 1.10); use of tamoxifen (HR, 0.59; 95% CI, 0.35 to 1.01); and history of oophorectomy (HR, 0.44; 95% CI, 0.21 to 0.91). The effect of oophorectomy was particularly strong in women first diagnosed prior to age 49 years (HR, 0.24; 95% CI, 0.07 to 0.77). For women who did not have an oophorectomy or take tamoxifen, the 10-year risk of contralateral cancer was 43.4% for BRCA1 carriers and 34.6% for BRCA2 carriers. Conclusion The risk of contralateral breast cancer in women with a BRCA mutation is approximately 40% at 10 years, and is reduced in women who take tamoxifen or who undergo an oophorectomy. J Clin Oncol 22:2328-2335. 2004 by American Society of Clinical Oncology Authors disclosures of potential conflicts of interest are found at the end of this article. Address reprint requests to Steven Narod, MD, Centre for Research on Women s Health, Women s College Hospital, 790 Bay St, Room 750, Toronto, Ontario, Canada M5G 1N8; e-mail: steven.narod@sw.ca. 2004 by American Society of Clinical Oncology 0732-183X/04/2212-2328/$20.00 DOI: 10.1200/JCO.2004.04.033 INTRODUCTION Women who carry a germline mutation in either the BRCA1 or BRCA2 gene face a high lifetime risk of breast cancer, 1 and, once diagnosed with breast cancer, face a high risk of subsequently developing cancer in the contralateral breast. 2-7 As a result, women may opt for contralateral mastectomy as a preventive measure as part of their initial treatment. It is not clear to what extent the risk of contralateral breast cancer varies with age, and if the risk is equivalent for carriers of BRCA1 and BRCA2 mutations. Factors that modify the risk of contralateral breast cancer in BRCA carriers have not been studied prospectively. In a case-control study, 8 we found that the risk of contralateral breast cancer in BRCA carriers was reduced if the woman had taken tamoxifen (hazard ratio [HR], 0.50; 95% CI, 0.28 to 0.89), if she had had an oophorectomy (HR, 0.42; 95% CI, 0.22 to 0.83), or if she had received chemotherapy (HR, 0.40; 95% CI, 0.26 to 0.60). However, it was not possible to estimate the actuarial risk for contralateral breast cancer or to study the possible effect of age at diagnosis on contralateral breast cancer risk. Here we present data from a large-scale prospective study designed to generate estimates of the risk of contralateral breast cancer among women with BRCA1 or BRCA2 mutations, and to identify host- and treatment-related factors that might modify the risk. 2328

BRCA Mutations and Contralateral Breast Cancer PATIENTS AND METHODS Study Subjects Pedigrees of cancer families identified at 10 participating cancer genetics clinics were reviewed. A family was considered to be eligible for the study when a BRCA1 or a BRCA2 mutation was documented in the family, and at least one case of invasive breast cancer was recorded. Eligible study subjects included all women from these families who were diagnosed with stage I or stage II breast cancer at age 65 years or younger, between 1975 and 2000. Living and deceased women were eligible, but those with a prior diagnosis of cancer (including breast cancer) or those who resided outside of North America were excluded. It was not necessary to be a proven carrier of the mutation found in the family to be included in the study; however, affected women who were known to be noncarriers were excluded. All study procedures were approved by the institutional review boards at each of the participating centers. We identified a total of 1,139 breast cancer cases in 337 families. Of the total of 1,139 cases of breast cancer, 320 women were excluded because the date of diagnosis indicated on the pedigree was before 1975, and 50 women were excluded because the age of diagnosis was older than 65 years. An additional 19 women were not carriers of the familial mutation and were therefore excluded. Ten women were excluded because they had a diagnosis of other cancer before breast cancer, and 11 women were excluded because they were treated outside of North America. We were able to obtain identifying information for 587 of the remaining 729 women (80.5%). An attempt was made to contact each of these women or their respective next-of-kin to obtain permission to review the medical records. Thirteen women (or their next of kin) refused to provide consent for the release of their medical records. The medical record was requested from the hospital where treatment was received for the remaining 574 women. In 54 cases, the hospital was not able to locate the record or did not forward the requested documents. The medical record was obtained for the remaining 520 women (91%). After review of the medical records, an additional 29 women were excluded. Of these, 22 women were excluded because tumor stage was greater than 2, six women were excluded because the tumor was noninvasive (ductal carcinoma in situ or lobular carcinoma in situ), and one woman was excluded because she refused treatment. The remaining 491 women were included in the analysis. In summary, of the 1,139 breast cancers, 439 were ineligible, and we were able to enroll 491 of the remaining 700 eligible cases (70%). Ninety-two (18.7%) of the 491 women were deceased. Study Protocol The medical treatment records and pathology documents were reviewed. We established whether the tumor was unilateral (482 cases) or bilateral (nine cases) at initial diagnosis. We recorded tumor size (in centimeters), nodal status (positive or negative), and tumor grade (I to III). Where possible, we abstracted information on both mitotic grade and nuclear grade. Estrogen receptor status was recorded as positive, negative, or equivocal. We recorded the use of chemotherapy (yes or no), tamoxifen (yes or no), and radiotherapy (yes or no). We established whether or not the patient had undergone a bilateral oophorectomy, and if so, the date of the operation. We categorized the principal surgery as either breast-conserving surgery (ie, lumpectomy), unilateral mastectomy, or bilateral mastectomy. In some cases, the contralateral mastectomy and/or oophorectomy were performed several years after the initial surgical treatment; the dates of these late treatments were recorded. We reviewed the medical records of deceased women to establish the cause of death. Because of the relatively young ages of death of these women (mean age, 48.0 years; range, 31.5 to 84.1 years), the great majority of deaths in the cohort were due to breast or ovarian cancer. We reviewed the dates of diagnosis of all contralateral breast cancers reported in the cohort. Only invasive contralateral cancers were included. Contralateral breast cancers were assumed to be new primary breast cancers and not metastases of the original primary. No woman was diagnosed with contralateral breast cancer after she was diagnosed with distal metastases. Statistical Analysis A survival analysis was performed. We considered the woman to be at risk for contralateral breast cancer from the date of the first surgical procedure until either death from breast or ovarian cancer, death from another cause, or the date of last follow-up. Initially, we compared the risk of contralateral breast cancer in women who had and who did not have a contralateral mastectomy. In subsequent analyses, only women for whom the contralateral breast was intact were considered to be at risk for contralateral breast cancer. Survival curves were constructed and compared for subgroups of women defined by age ( 40 years; 41 to 50 years; 51 years) and by mutation status (BRCA1 v BRCA2). We also compared the risk of contralateral breast cancer for subgroups defined by each of the four treatments individually (ie, tamoxifen v no tamoxifen, oophorectomy v no oophorectomy, and so on) and in combination. Hazard ratios (HRs) were estimated using the Cox proportional hazards model, implemented in SAS (SAS Institute, Cary, NC). The HRs were adjusted for age, mutation status (BRCA1 v BRCA2), and other treatments received. RESULTS There were 491 women who were treated for stage I or II breast cancer between 1975 and 2000. There were 327 women from families with a BRCA1 mutation, and 152 women from families with a BRCA2 mutation. Twelve women carried a mutation in both genes. The characteristics of the 491 women with breast cancer included in the analysis are presented in Table 1, the characteristics of their tumors are detailed in Table 2, and the treatments they received are presented in Table 3. The majority of the subjects were treated with breastconserving surgery (39%) or with unilateral mastectomy (52%), but 44 women (9%) were initially treated with a bilateral mastectomy. One hundred six women had a contralateral preventive mastectomy at various times after the initial surgery (range, 1 month to 14 years). However, three women had had their contralateral breast removed before their first diagnosis of invasive breast cancer (two because of ductal carcinoma in situ, and one because of fibrocystic disease). These three women were observed for contralateral breast cancer after their diagnosis of invasive cancer, and were included with the women who had bilateral mastectomies. Of the nine women with bilateral synchronous www.jco.org 2329

Metcalfe et al Table 1. Characteristics of 491 Study Subjects Patients Characteristic No. % Year of birth Year 1947 Range 1916-1971 Age of diagnosis, years Mean 42.1 Range 24-65 Age of diagnosis, years 30 28 5.7 30-40 190 38.7 40-50 180 36.7 50 93 18.9 Mutation BRCA1 327 66.6 BRCA2 152 31.0 Both 12 2.4 Place of residence Canada 191 38.9 United States 300 61.1 Genetic testing Yes, positive 411 83.7 Not done 80 16.3 Vital status Living 399 81.3 Deceased 92 18.7 Table 2. Characteristics of 491 Breast Cancers in the Study Characteristic No. % Stage I 231 47.0 II 260 53.0 Size, cm 0-2 324 66.0 2-5 167 34.0 Grade I 13 2.6 II 93 18.9 III 214 43.6 Unknown 171 34.8 ER status Positive 146 29.7 Negative 213 43.4 Equivocal 12 2.4 Unknown 120 24.4 Lymph node status Positive 160 32.6 Negative 331 67.4 Laterality Unilateral 482 98.2 Bilateral 9 1.8 Histology Medullary 43 8.8 Ductal 410 83.5 Lobular 16 3.3 Other 19 3.9 Unknown 3 0.6 NOTE. In the case of the nine women with synchronous bilateral tumors, only the first of the tumors that was recorded in the database is included here. Abbreviation: ER, estrogen receptor. breast cancer, seven had bilateral mastectomy, and two had bilateral breast-conserving surgery. These nine women were not at risk for contralateral breast cancer and are excluded from the following analyses. There were 209 women who had an oophorectomy, either for treatment of breast cancer, for the prevention of ovarian cancer, or for another reason. The exact indication for the oophorectomy was not known. After a mean follow-up period of 9.2 years, one contralateral breast cancer (in the chest wall) occurred among the 146 women treated with bilateral, prior, or delayed contralateral mastectomy. In contrast, there were 97 contralateral breast cancers diagnosed among the 336 women who retained the contralateral breast (HR, 0.03; P.0005). Because the observed risk of contralateral breast cancer following mastectomy was so low, the following analyses are restricted to the 336 women for whom the contralateral breast was intact. The mean interval between the diagnosis of the first breast cancer and the diagnosis of the contralateral breast cancer was 5.5 years (range, 0.1 to 16.2 years). The 5-year actuarial risk of contralateral breast cancer was 16.9% (95% CI, 10.5% to 23.3%), and the 10-year risk of breast cancer was 29.5% (95% CI, 20.6% to 38.3%). The risk of contralateral breast cancer was estimated for patient subgroups defined by age, gene (BRCA1 v BRCA2), and by treatment received (surgery, chemotherapy, tamoxifen, ovarian ablation). The univariate and multivariate HRs associated with each of these factors are presented in Table 4. The 10-year risk was slightly, but not significantly, higher for women with BRCA1 mutations (32%) than for women with BRCA2 mutations (24.5%; Fig 1). The risk was also slightly higher for women younger than 50 years at first breast cancer diagnosis (31%) compared with those diagnosed after age 50 years (23.5%; Fig 2). Neither radiotherapy nor chemotherapy use modified the risk significantly, but there was a moderate and borderline-significant decrease in the risk of contralateral breast cancer associated with tamoxifen use (HR, 0.59; 95% CI, 0.35 to 1.01; P.05). After adjustment for age, mutation, and other treatments, the reduction associated with tamoxifen was no longer significant (HR, 0.62; P.12). One hundred seven of the 336 women with an intact contralateral breast had a bilateral oophorectomy at some time; 14 women had an oophorectomy before breast cancer (13%); 15 women had an oophorectomy within 1 year of the diagnosis of breast cancer (14%); and 78 women had an oophorectomy from 1 year to 25 years after their initial 2330 JOURNAL OF CLINICAL ONCOLOGY

BRCA Mutations and Contralateral Breast Cancer Table 3. Characteristics of Treatments Received Among 482 Women With Unilateral Breast Cancer Treatment No. % Surgical Lumpectomy 191 39.6 Unilateral mastectomy 254 52.7 Bilateral mastectomy 37 7.7 Chemotherapy Yes 296 61.4 No 176 36.5 Missing 10 2.1 Tamoxifen Yes 144 29.9 No 310 64.3 Missing 28 5.8 Oophorectomy Never 259 53.7 Ever 209 43.4 Missing 14 2.9 Timing of oophorectomy Prior to diagnosis 33 15.8 After diagnosis 174 83.3 Missing 2 1.0 Radiotherapy Yes 221 45.9 No 253 52.5 Missing 8 1.7 Fig 1. Actuarial risks of contralateral breast cancer in carriers of BRCA1 and BRCA2 mutations. Dx, diagnosis. breast cancer diagnosis. To adjust for the effects of the timing of oophorectomy, oophorectomy was treated as a time-dependent covariate in the survival analysis. Oophorectomy was associated with a 59% reduction in the risk of contralateral breast cancer (HR, 0.41; 95% CI, 0.18 to 0.90). The magnitude of the risk reduction observed with oophorectomy was greater for women younger than 50 years at diagnosis (HR, 0.24; 95%CI, 0.07 to 0.77; P.02) than for women 50 years or older (HR, 0.91; 95% CI, 0.26 to 3.21). There was only a modest difference in the risk of contralateral breast cancer for women who had their oophorectomy before or at the time of diagnosis (HR, 0.32; P.12) and those who had their oophorectomy after initial breast cancer surgery (HR, 0.50; P.15). All factors were considered Factor Table 4. Risks of Contralateral Breast Cancer Associated With Selected Factors Univariate Multivariate HR 95% CI P HR 95% CI P BRCA 0.47 to 1.15.17 0.39 to 1.09.10 BRCA1 1.0 1.0 BRCA2 0.73 0.65 Age, years 0.36 to 1.10.11 0.45 to 1.51.52 50 1.0 1.0 50 0.63 0.82 Oophorectomy 0.21 to 0.91.03 0.18 to 0.90.03 Yes 0.44 0.41 Chemotherapy 0.68 to 1.55.90 0.68 to 1.70.74 Yes 1.03 1.08 Radiotherapy 0.51 to 1.16.21 0.56 to 1.34.51 Yes 0.77 0.86 Tamoxifen 0.35 to 1.01.05 0.34 to 1.14.12 Yes 0.59 0.62 NOTE. Multivariate estimates are adjusted for age, mutation (BRCA1 or BRCA2), and other treatments. Analyses restricted to 336 women with intact contralateral breast. Abbreviation: HR, hazard ratio. www.jco.org 2331

Metcalfe et al Fig 2. Actuarial risks of contralateral breast cancer in BRCA carriers diagnosed prior to or after age 50 years. Dx, diagnosis. Fig 3. Actuarial risks of contralateral breast cancer in BRCA carriers with either tamoxifen or oophorectomy, compared with carriers receiving neither treatment. Dx, diagnosis. jointly in a multivariate analysis, but oophorectomy was found to be the only intervention protective against contralateral breast cancer at a statistical level of significance (HR, 0.41; P.03). Because women who took tamoxifen or had an oophorectomy experienced lower risks of contralateral breast cancer, the cumulative risk was re-estimated in women who did not receive either of these adjuvant hormonal treatments. This estimate accurately represents the underlying risk for contralateral breast cancer owing to the presence of the susceptibility gene in the absence of intervention. For BRCA1 carriers, the 5-year actuarial risk was 27.1%, and the 10-year risk was 43.4%. For BRCA2 carriers, the 5-year actuarial risk was 23.5%, and the 10-year risk was 34.6%. In comparison, the 10-year risks for women who used tamoxifen or who had an oophorectomy were 18.8% for BRCA1 carriers and 13.1% for BRCA2 carriers. The HR associated with either of these hormonal treatments, compared with neither treatment, was 0.41 (95% CI, 0.24 to 0.70; P.001; Fig 3). For women diagnosed before age 50 years, the HR associated with the combination of tamoxifen and oophorectomy was 0.09 (95% CI, 0.01 to 0.68; P.02). The risk reduction associated with tamoxifen treatment was similar for BRCA1 and BRCA2 carriers (Tables 5 and 6). Oophorectomy seemed to be more effective for BRCA1 carriers than for BRCA2 carriers, but the CIs were wide. Because the tumor estrogen receptor (ER) status was missing for many women, it was not possible to estimate the effects of hormonal therapies separately for the subgroup of ER-negative breast cancer patients. We did not see a reduction in contralateral breast cancer risk associated with chemotherapy (Tables 4, 5, and 6). Among women younger than 50 years, who may have undergone early menopause due to chemotherapy, the HR for contralateral breast cancer was 0.92 (95% CI, 0.58 to 1.45). Information was available on ipsilateral cancers for 188 of 191 women who had breast-conserving surgery. It was not possible to distinguish between local recurrences and new primary ipsilateral cancers, and therefore, the number of observed events represents the sum of these two categories. The 10-year risk of ipsilateral cancer was 11.5% for BRCA1 carriers and BRCA2 carriers combined (9.8% for BRCA1 carriers and 14.6% for BRCA2 carriers). The 10- year risk of ipsilateral cancer was 9.1% for 171 women who received radiotherapy, and was 33.8% for 17 women who did not receive radiotherapy (HR, 0.23; P.01; Fig 4). Tamoxifen use and chemotherapy were also associated with a reduction in the risk of ipsilateral breast cancer in the univariate analysis (Table 7). DISCUSSION We report a high risk of contralateral breast cancer among 491 women with breast cancer and for whom a BRCA1 or BRCA2 mutation was documented in the family. We estimate the 5-year actuarial risk of contralateral breast cancer to be 17%, and the 10-year risk to be 30%. Previous studies have been much smaller in size. Pierce et al 4 estimated the 5-year contralateral rate of breast cancer to be 20% in a historical cohort of 71 BRCA carriers. Chappuis et al 5 estimated the 5-year risk of contralateral breast cancer to be 10% in a study of 32 carriers. Haffty et al 7 estimated the 12-year risk of contralateral breast cancer to be 42% in a study of 22 women with BRCA mutations. We found the 10-year risk of contralateral breast cancer to be modestly greater for women diagnosed before age 50 years than after age 50 years. Verhoog et al 6 estimated the 10-year risk of contralateral breast cancers among 124 BRCA1 carriers diagnosed before age 50 years to be 40%, versus 12% for 40 patients diagnosed after age 50 years. Robson 2332 JOURNAL OF CLINICAL ONCOLOGY

BRCA Mutations and Contralateral Breast Cancer Factor Table 5. Risks of Contralateral Breast Cancer Associated With Selected Factors (BRCA1 carriers) Univariate Multivariate HR 95% CI P HR 95% CI P Age, years 0.33 to 1.35.25 0.44 to 1.94.83 50 1.0 1.0 50 0.66 0.92 Oophorectomy 0.16 to 0.86.021 0.13 to 0.84.020 Yes 0.37 0.33 Chemotherapy 0.60 to 1.62.96 0.61 to 1.76.89 Yes 0.99 1.04 Radiotherapy 0.56 to 1.50.74 0.59 to 1.62.98 Yes 0.92 0.98 Tamoxifen 0.26 to 1.27.17 0.26 to 1.33.59 Yes 0.57 0.59 NOTE. Multivariate estimates are adjusted for age, mutation, and other treatments. Analyses restricted to 224 women with an intact contralateral breast. Abbreviation: HR, hazard ratio. et al 3 estimated the 5-year risk of contralateral breast cancer to be 31% in a group of 30 BRCA carriers diagnosed before age 42 years. The current study has strengths in addition to its large sample size. All women with breast cancer in the BRCA mutation positive families were identified, and those who had been diagnosed with breast cancer from 1975 to 2000 at 65 years or younger were eligible to enter the study population. Eighty women with breast cancer who had not had genetic testing were included; however, women who had tested negative were excluded. We included untested and deceased women in the study in order to avoid the survivorship bias that would arise if only tested women were included. The majority of genetic tests were done several years after the diagnosis of breast cancer. Restricting the study population to living women who were available for testing would inevitably have introduced a survivorship bias. However, given the presence of a documented mutation in all of the study families, the inclusion of untested women should not introduce significant misclassification bias. Of the 430 women with breast cancer in known BRCA-positive families who received genetic testing, 411 (96%) were found to be mutation carriers, and 19 (4%) were found to be noncarriers. Forty-three of the 80 untested women were de- Factor Table 6. Risks of Contralateral Breast Cancer Associated With Selected Factors (BRCA2 carriers) Univariate Multivariate HR 95% CI P HR 95% CI P Age, years 0.27 to 1.66.38 0.22 to 1.93.45 50 1.0 1.0 50 0.67 0.66 Oophorectomy 0.13 to 2.45.47 0.16 to 3.48.72 Yes 0.58 0.75 Chemotherapy 0.51 to 2.34.81 0.45 to 3.08.75 Yes 1.10 1.17 Radiotherapy 0.22 to 1.04.06 0.23 to 1.44.24 Yes 0.47 0.58 Tamoxifen 0.33 to 1.65.45 0.25 to 1.70.37 Yes 0.73 0.65 NOTE. Multivariate estimates are adjusted for age, mutation, and other treatments. Analyses restricted to 112 women with an intact contralateral breast. Abbreviation: HR, hazard ratio. www.jco.org 2333

Metcalfe et al Fig 4. Actuarial risks of ipsilateral breast cancer in BRCA carriers with breast-conserving surgery, with and without radiotherapy (Radio.). Dx, diagnosis. ceased (54%). The average age of diagnoses of both the tested and untested women was 42.2 years. However, compared with the tested women, the untested women were more likely to have stage II tumors (64% v 51%). Bilateral breast cancer was not among the eligibility criteria for genetic testing in the participating centers. However, if families with members with bilateral breast cancer were more likely to seek and to obtain genetic testing than families with single primary cancers only, then our study sample would be enriched for women with bilateral cancers. This would have the effect of spuriously raising the estimate of the contralateral cancer risk. This is a potential bias among all studies reported to date, but this problem will be lessened for purely prospective studies. There was a clear and dramatic benefit of contralateral mastectomy in terms of cancer incidence. Future studies will be required to determine whether this confers a benefit in terms of mortality. We also found the risk of contralateral cancer to be reduced when there was a history of either oophorectomy or use of tamoxifen, in keeping with the results of our earlier case-control study. 8 The combination of the two was particularly effective among young women (HR, 0.09; 95% CI, 0.01 to 0.68). This degree of protection is perhaps surprising, given that the majority of these tumors (approximately 60%) will be ER-negative (Table 2). Unfortunately, we did not obtain data on the ER status of the contralateral tumors and data on ER status was missing for many of the tumors diagnosed before 1985. The observed protective effect of tamoxifen was slightly greater for the BRCA1 carriers (HR, 0.57) than for the BRCA2 carriers (HR, 0.73; Tables 5 and 6), but this difference was not statistically significant. Our findings have important implications for the primary prevention of hereditary breast cancer. It is now generally well accepted that oophorectomy protects against breast cancer in BRCA carriers, 9,10 but the role of tamoxifen remains controversial. King et al 11 found no primary protective role for tamoxifen against BRCA1-associated breast cancer, but the study included only eight carriers with breast cancer. In our study the observed magnitude of the protective effect was larger for oophorectomy (59%) than Factor Table 7. Risks of Ipsilateral Breast Cancer Associated With Selected Factors Univariate Multivariate HR 95% CI P HR 95% CI P BRCA 0.37 to 2.83.97 0.24 to 2.52.67 BRCA1 1.0 1.0 BRCA2 1.02 0.77 Age, years 0.65 to 6.54.22 0.74 to 13.2.67 50 1.0 1.0 50 2.07 3.13 Oophorectomy 0.11 to 1.80.26 0.07 to 1.59.17 Yes 0.45 0.33 Chemotherapy 0.07 to 0.73.013 0.10 to 1.07.06 Yes 0.23 0.32 Radiotherapy 0.09 to 0.97.045 0.10 to 1.25.11 Yes 0.29 0.35 Tamoxifen 0.02 to 1.26.084 0.02 to 1.51.11 No 1.0 0.11 Yes 0.17 0.17 NOTE. Multivariate estimates are adjusted for age, mutation (BRCA1 or BRCA2), and other treatments. Abbreviation: HR, hazard ratio. 2334 JOURNAL OF CLINICAL ONCOLOGY

BRCA Mutations and Contralateral Breast Cancer for tamoxifen (38%). Potential benefits of oophorectomy in the management of primary breast cancer include reduction in risk of contralateral breast cancer, reduction in risk of primary breast cancer recurrence, and in prevention of primary ovarian cancer. Ovarian ablation has been shown to be an effective treatment for premenopausal breast cancer. In a large meta-analysis of women treated for breast cancer younger than 50 years, 15-year survival after breast cancer improved from 46% to 52% with ovarian ablation. 12 We also observed a (nonsignificant) reduction in the risk of ipsilateral breast cancer with oophorectomy (OR, 0.33), but it has not yet been established whether oophorectomy impacts on total mortality in this situation. These studies are now underway. It has been suggested that BRCA1-associated tumors show enhanced chemosensitivity. This may be related to their rapid growth rate and the impaired ability of BRCAdeficient cells to repair double stranded DNA breaks. 13 In the present study, we saw a protective effect of chemotherapy on the risk of ipsilateral cancer (Table 7) but not on the risk of contralateral cancer (Table 4). In contrast, a strong protective effect of chemotherapy on contralateral breast cancer risk was seen in our case-control study. 8 The reason for the observed discrepancy between the two studies is unknown, but some of the variation may be due to chance. It has also been speculated that radiotherapy (or mammography) may be hazardous to BRCA carriers because of the potential for increased mutability due to impaired DNA repair. Despite this theoretical concern, radiotherapy treatment was associated with a significant decrease in the risk of ipsilateral cancer following breast-conserving surgery in BRCA mutation carriers. Among patients treated for breast cancer with breast-conserving surgery and radiotherapy, the 10-year risk of contralateral breast cancer (27.4%) greatly exceeded that of ipsilateral cancer (11.5%). This was largely due to the protective effect of radiotherapy on ipsilateral breast cancer (HR, 0.29) but not on contralateral breast cancer (HR, 1.02). Chappuis et al 5 also reported a higher 5-year rate of contralateral breast cancer (10%) than of ipsilateral cancers (6%) in a study of 32 BRCA carriers. In contrast Haffty et al 7 reported similar high rates of ipsilateral (49%) and contralateral (42%) cancers in BRCA carriers at 12 years, but their study was small (22 carriers). In the present study, the rate of contralateral breast cancer in subjects with breast tissue at risk was similar to the rate of ipsilateral cancer observed in the small group of women who did not receive radiotherapy as part of their initial breast cancer treatment (Fig 4). This suggests that radiotherapy following breast-conserving surgery is valuable for prevention of local recurrence in BRCA carriers. We saw no evidence of a hazardous effect of radiotherapy for contralateral or ipsilateral breast cancers throughout the course of this study, but it will be important to follow-up all patients in this cohort for a longer period before the longterm effects of radiotherapy can be properly assessed. Patients with hereditary susceptibility to breast cancer are generally treated according to protocols established for nonhereditary forms of cancer (ie, most premenopausal women with ER-negative, high-grade tumors of 2cmin diameter are offered breast-conserving surgery and adjuvant chemotherapy, but not ovarian ablation or tamoxifen). In this scenario, the risk of contralateral breast cancer is approximately 4% per year, or 40% at 10 years. The risk of contralateral cancer is almost zero for women who undergo bilateral mastectomy, and is decreased for women who undergo oophorectomy or are treated with tamoxifen. Authors Disclosures of Potential Conflicts of Interest The authors indicated no potential conflicts of interest. REFERENCES 1. Ford D, Easton DF, Stratton M, et al: Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families. Am J Hum Genet 62:676-689, 1998 2. Verhoog LC, Brekelmans CTM, Synaeve C, et al: Survival and tumor characteristics of breast cancer patients with germline mutations of BRCA1. Lancet 351:316-321, 1998 3. Robson M, Gilewki T, Haas B, et al: BRCAassociated breast cancer in young women. J Clin Oncol 16:1642-1649, 1998 4. Pierce LJ, Strawderman M, Narod SA, et al: Effect of radiotherapy after breast-conserving treatment in women with breast cancer and germline BRCA1/2 mutations. J Clin Oncol 18: 3360-3369, 2000 5. Chappuis PO, Kapusta L, Begin LR, et al: Germline BRCA1/2 mutations and p27kip protein levels independently predict outcome after breast cancer. J Clin Oncol 18:4045-4052, 2000 6. Verhoog LC, Brekelmans CTM, Seynaeve C, et al: Contralateral breast cancer risk is influenced by the age at onset in BRCA1-associated breast cancer. Br J Cancer 83:384-386, 2000 7. Haffty BG, Harrold E, Khan AJ, et al: Outcome of conservatively managed early-onset breast cancer by BRCA1/2 status. Lancet 359: 1471-1477, 2002 8. Narod SA, Brunet JS, Ghadirian P, et al: Tamoxifen and the risk of contralateral breast cancer in BRCA1 and BRCA2 mutation carriers: A case-control study. Lancet 356:1876-1881, 2000 9. Rebbeck TR, Levin AM, Eisen A, et al: Reduction in breast cancer risk after bilateral prophylactic oophorectomy in BRCA1 mutation carriers. JNCI 91:1475-1479, 1999 10. Rebbeck TR, Lynch HT, Neuhausen SL, et al: Prophylactic oophorectomy in carriers of BRCA1 and BRCA2 mutations. N Engl J Med 346:1616-1622, 2002 11. King MC, Wieand S, Hale K, et al: Tamoxifen and breast cancer incidence among women with inherited mutations in BRCA1 and BRCA2: National Surgical Adjuvant Breast and Bowel Project (NSABP) Breast Cancer Prevention Trial. JAMA 286:2251-2256, 2001 12. Early Breast Cancer Trialists Collaborative Group: Ovarian ablation in early breast cancer: Overview of the randomized trials. Lancet 348: 1189-1196, 1996 13. Hoeijmakers JH: Genomic maintenance mechanisms for preventing cancer. Nature 441: 366-374, 2001 www.jco.org 2335