Twenty-One Gene Recurrence Score Assay in BRCA-Associated Versus Sporadic Breast Cancers: Differences Based on Germline Mutation Status

Twenty-One Gene Recurrence Score Assay in BRCA-Associated Versus Sporadic Breast Cancers: Differences Based on Germline Mutation Status Payal D. Shah, MD 1 ; Sujata Patil, PhD 2 ; Maura N. Dickler, MD 3 ; Kenneth Offit, MD 4 ; Clifford A. Hudis, MD 3 ; and Mark E. Robson, MD 3,4 BACKGROUND: Biological differences between BRCA-associated breast cancer and sporadic breast cancer may warrant different adjuvant chemotherapy (ACRx) recommendations despite similar phenotypic features. A 21-gene expression profile (Oncotype DX) generates a prognostic recurrence score (RS) that predicts the ACRx benefit in patients with hormone receptor positive breast cancer. No reports describe assay results for BRCA-associated breast cancer. METHODS: A review of Memorial Sloan Kettering Cancer Center databases identified 4908 patients with hormone receptor positive, node-negative breast cancer who underwent Oncotype DX testing between July 2006 and March 2014. / carriers (cases) were identified and matched (1:2) by age at diagnosis and tumor size to noncarrier controls. Two-sample nonparametric tests were used to compare the baseline characteristics, RSs, and risk stratification between and patients. Conditional logistic regression was used to assess these differences by mutational status. RESULTS: Fifty mutation-associated cases (19 cases and 31 cases) and 100 controls who were well matched for age (P 5.9) and tumor size (P 5.6) were included. and carriers had similar median RSs (P 5.6) and risk category stratification (P 5.3). The median RS was higher for cases versus controls (24 vs 16; P <.0001). Risk stratification also differed by mutational status (P 5.0002). had more high-risk disease (28% vs 7%) and intermediate-risk disease (56% vs 36%) and less low-risk disease (16% vs 57%). were more likely than controls to receive ACRx (74% vs 46%; P 5.002). CONCLUSIONS: Germline BRCA-associated hormone receptor positive breast cancer may be associated with intrinsically less favorable biology. Few affected carriers have RS indicating a clear absence of benefit from ACRx. The increased use of ACRx and benefit from ACRx in BRCA carriers may mitigate otherwise inferior outcomes. Cancer 2016;122:1178-84. VC 2016 American Cancer Society. KEYWORDS: adjuvant chemotherapy, BRCA, breast cancer, Oncotype DX, recurrence. INTRODUCTION Deleterious germline mutations in and confer an increased risk of breast cancer. Risk reduction strategies 1,2 and novel therapeutic approaches for advanced disease 3-5 have been described for these patients; however, significantly fewer data exist for the optimal management of affected BRCA mutation carriers with early-stage breast cancers. The Oncotype DX assay (Genomic Health, Inc, Redwood City, Calif) has clinical validity in predicting the risk of breast cancer recurrence as well as clinical utility in quantifying the benefit of adjuvant systemic therapy in patients with early-stage, hormone-responsive breast cancer. 6,7 The output of this 21-gene expression profile is a recurrence score (RS), which translates into a therapeutically meaningful categorization of risk as low, intermediate, or high. A tumor s RS complements its histopathologic description and, when placed in context with other risk stratification elements, helps to inform the selection of patients appropriate for adjuvant systemic therapy. The BRCA mutation status may be one such relevant risk stratification element. 8-10 BRCA-associated cancers have distinct pathological features 8,11,12 and gene expression profiles 13 in comparison with sporadic cancers. These biological differences may translate into treatment responses different from those for sporadic cancers. Indeed, multiple reports indicate that BRCA-associated tumors demonstrate an enhanced response to chemotherapy, 9,10,14-16 which may, in these patients, abrogate otherwise inferior outcomes. 8,10 Accordingly, different recommendations for adjuvant treatment may be warranted for BRCA mutation carriers and noncarriers with phenotypically similar cancers. Corresponding author: Mark E. Robson, MD, Breast Medicine Service/Clinical Genetics Service, Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York, NY 10065; Fax: (646) 888-4081; robsonm@mskcc.org 1 Abramson Cancer Center, University of Pennsylvania, Philadelphia, Pennsylvania; 2 Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, New York; 3 Breast Medicine Service, Memorial Sloan Kettering Cancer Center, New York, New York; 4 Clinical Genetics Service, Memorial Sloan Kettering Cancer Center, New York, New York. DOI: 10.1002/cncr.29903, Received: October 12, 2015; Revised: December 10, 2015; Accepted: December 28, 2015, Published online February 9, 2016 in Wiley Online Library (wileyonlinelibrary.com) 1178 Cancer April 15, 2016

Oncotype DX in BRCA Mutation Carriers/Shah et al Intrinsic differences between BRCA-associated and sporadic cancers may be reflected in the results of an assay that quantifies recurrence risk and adjuvant systemic therapy benefit. Although Oncotype DX has been well validated for prognosis and for the prediction of chemotherapy benefit, to our knowledge, it has not been described specifically for breast cancers associated with deleterious BRCA mutations. Here we report what we believe is the first description of assay results for breast cancer patients with germline BRCA mutations versus matched cases without germline alterations. MATERIALS AND METHODS Study Population Institutional review board/privacy board approval was obtained to permit subject identification through a review of electronic medical records. A database query identified all patients with hormone receptor positive, human epidermal growth factor receptor 2 (HER2) unamplified, stage I or II breast cancer diagnosed at Memorial Sloan Kettering between July 2006 and March 2014 whose tumors underwent gene expression profiling with the commercially available 21-gene assay. This query yielded 4908 patients. Among these patients, 613 were noted in the Clinical Genetics Service database or medical record as having been tested for alterations in /, and 50 patients harbored deleterious germline mutations (cases). Controls matched for patient age and tumor size were selected from initial query results in a 2:1 fashion by a single investigator blinded to RSs. The matching parameters were as follows: age within 5 years of diagnosis and tumor size within 0.6 cm. If more than 2 controls were available, the selected controls were closest to the case by year of diagnosis. Data Collection Clinical and pathological data were abstracted from the electronic medical record and included the following: patient age, BRCA testing date and results, date of diagnosis, TNM stage, morphological subtype, histologic grade, estrogen receptor status, progesterone receptor status, human epidermal growth factor receptor 2 amplification status, tumor multifocality, RS, and administration of adjuvant chemotherapy (ACRx; yes or no). All data that were initially obtained via the automated query were confirmed through a manual chart review. Tumor specimens were reviewed by surgical pathology at Memorial Sloan Kettering Cancer Center, and these data, rather than data from local pathology reports, were recorded. The tumor size recorded and used for matching was the diameter of the largest focus of invasive carcinoma. For reports without numerical histological grading, tumors described as poorly differentiated, moderately differentiated, and well differentiated were given grades of 3, 2, and 1, respectfully. When 2 descriptors were used (ie, moderately to poorly differentiated), the recorded grade was the higher of the 2. Gene Expression Profiling In all patients, gene expression profiling was performed in fixed paraffin-embedded tumor tissue by Genomic Health, Inc with the Oncotype DX assay as part of routine clinical care to aid in decisions regarding adjuvant systemic therapy. The assay has been previously described. 6,17 Statistical Methods Patient and clinicopathologic characteristics at diagnosis were compared in 2 sets: 1) a comparison of and with 2-sample nonparametric tests and 2) a comparison of cases and controls with differences assessed with a conditional logistic regression model. This model accounted for the matched pairs design. In sensitivity analyses, 3 subset analyses were performed: 1) excluding a mutation carrier with micrometastatic nodal involvement and her controls, 2) excluding a carrier with low (1%-10%) hormone receptor positivity and her controls, and 3) excluding both patients and their respective controls. To examine whether the relation between RS and case/control status (BRCA carrier/noncarrier status) persisted independently of grade, we conducted 2 secondary analyses. In the first, we included Oncotype DX risk stratification (low, intermediate, or high) and grade (1 1 2 vs 3) as covariates in a multivariate conditional logistic regression. In the second, we analyzed the subset of patients with grade 3 tumors. Two-sided P values less than 5% were considered significant. All analyses were conducted with SAS 9.4. RESULTS Study Population Characteristics The study population included 50 patients with BRCA mutations (19 with mutations and 31 with mutations) as well as 100 matched control patients without known deleterious germline BRCA mutations. One hundred forty-six of the 150 patients included in this analysis underwent testing for deleterious / mutations either as part of their routine clinical care (142 patients) or as part of research protocols (4 patients) at Memorial Sloan Kettering Cancer Center Cancer April 15, 2016 1179

TABLE 1. Study Population Characteristics All Patients (n 5 19) (n 5 31) P ( vs ) (n 5 50) Controls (n 5 100) P ( vs Controls) Age, median (range), y 46 (24-62) 43 (31-69).858 44.5 (24-69) 44 (22-70).936 Tumor size, median (range), cm 1.3 (0.4-2.4) 1.4 (0.3-2.8).351 1.4 (0.3-2.8) 1.2 (0.3-3.2).600 Morphology, No. (%).119.800 Invasive ductal 18 (95) 27 (87) 45 (90) 87 (87) Invasive lobular 0 4 (13) 4 (8) 9 (9) Other 1 (5) 0 1 (2) 4 (4) Estrogen receptor, median (range), 90 (5-100) 90 (12.5-100).876 90 (5-100) 95 (30-100).051 quantitative % a Progesterone receptor, median (range), 60 (0-100) 55 (0-98).976 60 (0-100) 85 (0-100).011 quantitative % a Histologic grade, No. (%) b.215.005 I 0 0 0 7 (7) II 1 (5) 6 (19) 7 (14) 28 (28) III 18 (95) 22 (71) 40 (80) 54 (54) Multifocal, No. (%).452.077 Yes 2 (11) 7 (23) 9 (18) 8 (8) No 17 (89) 24 (77) 41 (82) 92 (92) Adjuvant chemotherapy, No. (%).054.002 Yes 11 (58) 26 (84) 37 (74) 46 (46) No 8 (42) 5 (16) 13 (26) 54 (54) a If hormone receptor positivity was reported in the pathology report as a range, the mean was taken. b For the histologic grade, grades 1 and 2 were combined for the analysis; the histologic grade was unavailable for 3 -associated tumors and 11 sporadic tumors. or other institutions. Ninety-six of the 100 control patients had BRCA testing performed with negative results. Four additional control patients had no testing performed but had a low pretest probability for carrying a germline mutation on the basis of their personal and family histories. This included a postmenopausal diagnosis at an age 59 years for all 4 patients; an absence of a family history of breast, ovarian, or other BRCA-associated cancers for 3 patients; and only a mother with breast cancer diagnosed at the age of 70 years for the fourth patient. These patients were not known to be Ashkenazi Jews or members of other known founder populations. Clinical and pathological characteristics are shown in Table 1. The median age of the cases was similar to that of the controls (44.5 and 44; P 5.9), with no significant differences between and mutation carriers. Tumor sizes were also similar between cases and controls (1.4 cm vs 1.2 cm; P 5.6) and between and mutation carriers. Ninety percent of the cases and 87% of the controls had tumors with invasive ductal histology. The median estrogen receptor positivity was 90% for carriers, 90% for carriers, and 95% for controls (P for cases vs controls 5.05). The median progesterone receptor expression was 60% for cases and 85% for controls (P 5.01). were significantly more likely to have histologic grade 3 cancers (P 5.005). The histologic grade was unavailable for 3 -associated tumors and 11 sporadic tumors. All patients had node-negative disease with the exception of 1 patient with a deleterious mutation and micrometastatic lymph node involvement (patient B16). One patient with a deleterious mutation and low (1%-10%) expression of estrogen and progesterone receptors according to immunohistochemistry (patient B33) was included. Comparison of RSs and Risk Stratification in Mutation Carriers Versus Mutation Carriers RS medians and distributions are shown in Table 2. and carriers had similar median RSs (23 vs 24; P 5.6) and risk category stratification (P 5.3). More low-risk tumors were seen in patients with mutations (26%) versus patients with mutations (10%). In both groups of patients, most tumors were intermediate-risk. The 2 groups had similar proportions of tumors categorized as high-risk (26.3% in the group vs 29% in the group). Distribution of RS in Controls The median RS for controls was 16 (range, 3-46). Most RSs for controls were categorized as low-risk (57%). 1180 Cancer April 15, 2016

Oncotype DX in BRCA Mutation Carriers/Shah et al TABLE 2. Recurrence Scores, Medians, and Distributions All Patients (n 5 19) (n 5 31) P ( vs ) (n 5 50) Controls (n 5 100) P ( vs Controls) Recurrence score, median (range) 23 (12-61) 24 (10-71).619 24 (10-71) 16 (3-46) <.0001 Low risk (recurrence score < 18), No. (%) 5 (26.3) 3 (10).289 8 (16) 57 (57).0002 Intermediate risk (recurrence score 18 9 (47.4) 19 (61) 28 (56) 36 (36) and 30), No. (%) High risk (recurrence score 31), No. (%) 5 (26.3) 9 (29) 14 (28) 7 (7) remained statistically significant (data not shown). RS remained significantly correlated with the BRCA mutation status in analyses adjusted for grade and hormone receptor status. Among only grade 3 tumors, the RS distribution also remained significantly associated with the BRCA mutation status (P 5.0106). Figure 1. Risk stratification by the germline mutation status. The percentages of cases and controls with low-, intermediate-, and high-risk disease are shown. The relative proportions of and mutation carriers in each risk category are also indicated. A minority of controls had high-risk disease (7%), and the remainder had intermediate-risk disease (36%). Comparison of RSs in BRCA Mutation Carriers Versus Noncarriers The median RS was higher for mutation carriers (cases) versus controls (24 vs 16; P <.0001). There was also a statistically significant difference in the distribution of risk stratification based on the BRCA mutation status (P 5.0002), as shown in Figure 1. had less low-risk disease (16% vs 57%) than controls. had more intermediate-risk (56% vs 36%) and high-risk disease (28% vs 7%) than controls. Sensitivity Analyses The RS medians and distributions were assessed with the inclusion and exclusion of case B16 and/or case B33 as well as their respective controls individually and together. The difference between carriers and noncarriers in the median RSs and in the risk stratification distributions Administration of ACRx Fifty-eight percent of mutation carriers and 84% of mutation carriers received adjuvant systemic therapy (P 5.054). were significantly more likely to receive ACRx than controls (74% vs 46%; P 5.002). Within risk categories, however, there were no significant differences between cases and controls in proportions receiving chemotherapy (Table 3). DISCUSSION The current report is, to our knowledge, the first description of RSs from gene expression profiles for patients with breast cancers associated with deleterious germline mutations in and. This study also compares these RSs with those from a control group of patients with sporadic breast cancer who were matched to cases by patient age and tumor size. Finally, the study describes rates of chemotherapy administration in the 2 groups overall and stratified by risk category. In this data set, there is a statistically and clinically significant difference in the median RS between mutation carriers and noncarriers and in the risk stratification distributions generated from these scores. These RS differences were reflected in significantly different rates of adjuvant systemic therapy administration in carriers and noncarriers. Because the Oncotype DX assay has demonstrated prognostic validity and predictive utility resulting in its incorporation into clinical decision making, these findings have potentially meaningful implications. In this analysis, BRCA mutation carriers were more likely to have intermediate- or high-risk disease and less likely to have low-risk disease in comparison with Cancer April 15, 2016 1181

TABLE 3. Administration of Adjuvant Chemotherapy by Risk Stratification and Germline Mutational Status Low Risk Intermediate Risk High Risk Noncarriers (n 5 7) All BRCA 1 (n 5 14) (n 5 9) (n 5 5) Noncarriers (n 5 36) All BRCA 1 (n 5 28) (n 5 19) (n 5 9) Noncarriers (n 5 57) All BRCA 1 (n 5 8) (n 5 3) (n 5 5) Adjuvant Chemotherapy Yes, No. (%) 1 (20) 1 (33) 2 (25) 14 (25) 6 (67) 16 (84) 22 (79) 25 (69) 4 (80) 9 (100) 13 (93) 7 (100) No, No. (%) 4 (80) 2 (67) 6 (75) 43 (75) 3 (33) 3 (16) 6 (21) 11 (31) 1 (20) 0 (0) 1 (7) 0 (0).9999.1338 Not applicable P for cases vs controls controls. Mutation carriers also had a significantly higher median RS than controls (24 vs 16; P <.0001). These patterns remained unchanged with statistical significance in subset analyses after the exclusion of the 2 borderline cases (patient B16 with micrometastatic lymph node involvement and patient B33 with low positive hormone receptor expression) individually and together. Because this multigene assay is prognostic, these data indicate that BRCA-associated tumors may have intrinsically less favorable biology than sporadic cancers. Furthermore, because this assay is also predictive of the benefit of ACRx, these findings suggest that the presence of a germline BRCA mutation may be a marker for response to ACRx in patients with early-stage, hormone receptor positive breast cancers, and few BRCA mutation carriers have RSs indicative of a clear absence of benefit of adjuvant therapy. These inferences rely on the assumed clinical validity and clinical utility of the assay in BRCA-associated breast cancer, which have not specifically been examined in this subset of patients. Despite the inferior prognoses of BRCA-associated tumors suggested by the multigene assay scores, literaturereported outcomes are similar among affected carriers and noncarriers. 8 It has previously been suggested that the increased chemosensitivity of BRCA-related breast cancers might explain an observed prognosis similar to that of patients with sporadic cancers despite more aggressive tumor phenotypic features. 8,10,14 Goodwin et al 8 found that outcomes were similar for mutation carriers and noncarriers with stage I breast cancer. In their study, as in the current report, a significantly higher proportion of BRCA mutation carriers received adjuvant systemic therapy in comparison with noncarriers, and the authors concluded that BRCA mutation carriers who receive standard adjuvant treatment may fare as well as noncarriers of the same age who present with similar cancers. A 2004 analysis by Robson et al 10 also noted a poorer prognosis for Ashkenazi Jewish women with mutations that was abrogated by ACRx administration. Narod et al 9 noted improved survival in patients with BRCA mutations treated for stage I breast cancer with chemotherapy versus those who did not receive chemotherapy. The mortality difference was not attributable to distinct baseline risk profiles because women who received chemotherapy had larger tumors that were more likely to be estrogen receptor negative. That BRCA carriers may derive a particular benefit from chemotherapy has also been suggested by groups exploring treatment responses in BRCA carriers in the neoadjuvant 14,16,18,19 and metastatic settings. 15,20 Arun et al 14 1182 Cancer April 15, 2016

Oncotype DX in BRCA Mutation Carriers/Shah et al found that the status was independently associated with higher pathologic complete response rates after preoperative systemic therapy, although the mutation status was not. Another, smaller report noted an increased rate of pathologic complete response in / mutation carriers versus noncarriers. 16 Kriege et al 15 examined chemosensitivity to first-line metastatic chemotherapy for BRCA-associated breast cancer and found significantly increased chemosensitivity in mutation carriers on the basis of the overall response, progression-free survival, and overall survival as well as a trend toward a better response rate to and clinical benefit from therapy in mutation carriers. Consistent with this, in this series, BRCA-associated breast cancers had higher RSs indicative of unfavorable biology and chemotherapy benefit, and significantly more BRCA mutation carriers than affected noncarriers received ACRx. When it was stratified by risk category, ACRx administration was similar for carriers and noncarriers. That is, a mutation carrier with intermediate-risk disease was not more likely to receive ACRx than a noncarrier with intermediate-risk disease. Taken together, the findings are concordant with the literature in suggesting that increased use of, and benefit from, adjuvant systemic therapy in BRCA mutation carriers may mitigate otherwise inferior outcomes of intrinsically less favorable disease; this accounts for the similar prognoses for carriers and noncarriers reported in the literature. The study population included patients with hormone receptor positive, -associated breast cancers, an important subset on which to report because -associated tumors are more commonly hormone receptor negative. 11 Interestingly, - and - associated tumors have been noted to have different gene expression profiles 13 and disparate chemosensitivity. 15 This study did not find statistically significant differences in terms of RSs in -associated tumors versus -associated tumors. However, a higher proportion of -associated tumors than -associated tumors were low-risk (26% vs 10%). This raises the question whether some of these low-risk tumors in mutation carriers occurred sporadically despite the presence of a germline mutation. Tumor mutational profile and loss-of-heterozygosity data are unavailable to confirm the inactivation of the wild-type allele; therefore, it is not possible to establish with certainty that all tumors in patients with germline BRCA mutations arose because of the underlying predisposition. This study has limitations. First, the distribution of RSs in the control sample is slightly different from that reported in the literature. The proportion of low-risk RSs seen in the control subset is similar to an initial validation of the 21-gene assay. 6 However, the control patient population had more intermediate-risk disease than has been reported in the literature (36% in this series vs 20%- 22% 6,7 and fewer high-risk cancers (7% vs 25%-27% 6,7. In comparison with literature-reported RS distributions, the relative shift of BRCA-associated breast cancers toward intermediate-risk RSs becomes more prominent, whereas the relative shift of BRCA-associated breast cancers toward high-risk RSs becomes less prominent. Second, the sample size was too small to fully allow controlling for the tumor grade, hormone receptor status, and multifocality. The histologic grade in particular is known to correlate with the Oncotype DX RS and, therefore, may be viewed as a potential confounding factor in the current analysis. To address this, a grade-adjusted analysis was performed, and the RS distribution remained significantly associated with the BRCA mutation status. In addition, an examination of only grade 3 tumors similarly yielded a persistent, significant association of the RS with the BRCA mutation status. Furthermore, insofar as these downstream phenotypic features are a reflection of the genomic composition of a tumor, controlling for these factors may not be necessary or beneficial in assessing a biological estimation of recurrence risk and therapeutic benefit. Finally, the analysis included breast tumors as small as 0.3 cm. The National Comprehensive Cancer Network guidelines advocate the use of this multigene assay as a category 2A recommendation when evaluating patients with primary tumors characterized as 0.6-1.0 cm with unfavorable features or >1 cm, and node-negative, hormone receptor positive, and HER2-negative. 21 However, because cases and controls were well matched for tumor size, this is unlikely to significantly affect our overall findings. Because most data concerning the care of BRCA mutation carriers focus on risk management and treatment of advanced disease, these data supplement a relatively limited but clinically pertinent body of literature informing adjuvant systemic therapy decisions in affected mutation carriers. The findings are consistent with prior reports noting a benefit of systemic therapy for BRCAassociated breast cancer. That a relatively small proportion of affected BRCA carriers have low-risk breast cancer indicates that few BRCA carriers with breast cancer have a clear absence of benefit from ACRx. Importantly, the tumor size, estrogen receptor status, and age of diagnosis are not predictive of mortality, 8-10,22 so caution should be Cancer April 15, 2016 1183

exercised before one foregoes chemotherapy for BRCAassociated early-stage breast cancers because of reassuring phenotypic features. In sum, these data suggest that, in the absence of gene expression profiling through this validated assay, adjuvant systemic therapy should be strongly considered for BRCA mutation carriers with early-stage, hormone receptor positive breast cancer, as indicated by conventional factors. Although higher RSs for carriers indicate biologically unfavorable disease, higher rates of systemic therapy administration and potentially increased benefit from chemotherapy may account for similar prognoses among affected carriers and noncarriers. Given the available data, we concur with the current recommendations of the National Comprehensive Cancer Network in favor of the use of this multigene assay as an adjunct to other factors to inform ACRx recommendations in patients with early-stage, hormone receptor positive breast cancer, 21 regardless of the germline BRCA mutation status. However, because of the observed differences in RSs and risk stratification between cases and controls and between cases and literature-reported RSs, validation of this 21-gene assay in patients with BRCA-associated breast cancers may be warranted. FUNDING SUPPORT No specific funding was disclosed. CONFLICT OF INTEREST DISCLOSURES The authors made no disclosures. REFERENCES 1. Domchek SM, Friebel TM, Singer CF, et al. Association of riskreducing surgery in or mutation carriers with cancer risk and mortality. JAMA. 2010;304:967-975. 2. Robson M, Offit K. Clinical practice. Management of an inherited predisposition to breast cancer. N Engl J Med. 2007;357:154-162. 3. Nathanson KL, Domchek SM. Therapeutic approaches for women predisposed to breast cancer. Annu Rev Med. 2011;62:295-306. 4. Tutt A, Robson M, Garber JE, et al. Oral poly(adp-ribose) polymerase inhibitor olaparib in patients with or mutations and advanced breast cancer: a proof-of-concept trial. Lancet. 2010;376:235-244. 5. Kaufman B, Shapira-Frommer R, Schmutzler RK, et al. Olaparib monotherapy in patients with advanced cancer and a germline /2 mutation. J Clin Oncol. 2015;33:244-250. 6. Paik S, Shak S, Tang G, et al. A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med. 2004;351:2817-2826. 7. Paik S, Tang G, Shak S, et al. Gene expression and benefit of chemotherapy in women with node-negative, estrogen receptor positive breast cancer. J Clin Oncol. 2006;24:3726-3734. 8. Goodwin PJ, Phillips KA, West DW, et al. Breast cancer prognosis in and mutation carriers: an International Prospective Breast Cancer Family Registry population-based cohort study. J Clin Oncol. 2012;30:19-26. 9. Narod SA, Metcalfe K, Lynch HT, et al. Should all mutation carriers with stage I breast cancer receive chemotherapy? Breast Cancer Res Treat. 2013;138:273-279. 10. Robson ME, Chappuis PO, Satagopan J, et al. A combined analysis of outcome following breast cancer: differences in survival based on / mutation status and administration of adjuvant treatment. Breast Cancer Res. 2004;6:R8-R17. 11. Mavaddat N, Barrowdale D, Andrulis IL, et al. Pathology of breast and ovarian cancers among and mutation carriers: results from the Consortium of Investigators of Modifiers of /2 (CIMBA). Cancer Epidemiol Biomarkers Prev. 2012;21: 134-147. 12. Lakhani SR, Jacquemier J, Sloane JP, et al. Multifactorial analysis of differences between sporadic breast cancers and cancers involving and mutations. J Natl Cancer Inst. 1998;90:1138-1145. 13. Hedenfalk I, Duggan D, Chen Y, et al. Gene-expression profiles in hereditary breast cancer. N Engl J Med. 2001;344:539-548. 14. Arun B, Bayraktar S, Liu DD, et al. Response to neoadjuvant systemic therapy for breast cancer in BRCA mutation carriers and noncarriers: a single-institution experience. J Clin Oncol. 2011;29:3739-3746. 15. Kriege M, Seynaeve C, Meijers-Heijboer H, et al. Sensitivity to firstline chemotherapy for metastatic breast cancer in and mutation carriers. J Clin Oncol. 2009;27:3764-3771. 16. Chappuis PO, Goffin J, Wong N, et al. A significant response to neoadjuvant chemotherapy in /2 related breast cancer. J Med Genet. 2002;39:608-610. 17. Cronin M, Pho M, Dutta D, et al. Measurement of gene expression in archival paraffin-embedded tissues: development and performance of a 92-gene reverse transcriptase-polymerase chain reaction assay. Am J Pathol. 2004;164:35-42. 18. Byrski T, Gronwald J, Huzarski T, et al. Pathologic complete response rates in young women with -positive breast cancers after neoadjuvant chemotherapy. J Clin Oncol. 2010;28:375-379. 19. Byrski T, Huzarski T, Dent R, et al. Pathologic complete response to neoadjuvant cisplatin in -positive breast cancer patients. Breast Cancer Res Treat. 2014;147:401-405. 20. Kriege M, Jager A, Hooning MJ, et al. The efficacy of taxane chemotherapy for metastatic breast cancer in and mutation carriers. Cancer. 2012;118:899-907. 21. National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology: breast cancer. Version 3.2014. http://www. nccn.org/professionals/physician_gls/pdf/breast.pdf. Accessed January 5, 2015. 22. Rennert G, Bisland-Naggan S, Barnett-Griness O, et al. Clinical outcomes of breast cancer in carriers of and mutations. N Engl J Med. 2007;357:115-123. 1184 Cancer April 15, 2016