Vachon, Celine M.

Abstract The incidence of multiple myeloma (MM) is twice as high in blacks compared with whites. MM consists of several different cytogenetic subtypes, with differing presentations, response to therapy, and prognosis. The studies that have characterized the distribution of the primary cytogenetic abnormalities by fluorescence in-situ hybridization (FISH) testing have been primarily conducted in white populations, and more data are needed on whether the distribution of these cytogenetic abnormalities differs in black compared with white MM patients, and to determine if the increased prevalence of MM in blacks is related to specific subtypes of the disease. In this multi-center study, we examined four common clinically assessed cytogenetic subtypes (t(11;14), t(4;14), del13q, and del17p) at time of MM diagnosis in 245 black patients from 3 clinics who had prior FISH testing available in medical records. We compared the prevalence of these subtypes with our previously characterized Mayo Clinic cohort which was all Caucasian (n=484). The t(11;14) translocation was seen in 9 of 245 (4%) black patients with MM; this was significantly lower than the rate of t(11;14) translocations seen in white patients from Mayo Clinic (18%), p < 0.0001). In contrast to t(11;14), the prevalence of the t(4;14) translocation across the two groups was similar (p= 0.92). We conclude that the prevalence of t(11;14) translocation is significantly lower in blacks with myeloma compared with whites. Our findings suggest that the distribution of cytogenetic subtypes found in MM may differ significantly by race, and further studies assessing differences in the prevalence of other primary cytogenetic abnormalities such as trisomies, t(6;14), t(14;16), and t(14;20) are needed. Furthermore, these results support the hypothesis previously put forth that there are several different subtypes of multiple myeloma that may be defined by the major cytogenetic translocations. Since these results differ from those previously reported (Baker et al 2013, Blood), we are conducting further analyses of these data to examine possible factors that could account for the differential association of translocation by race, including probes used. Citation Format: Alexandra Joelle Greenberg, Sunita Philip, Agne Paner, Sylvia Velinova, Ashraf Badros, Rosalind Catchatorian, Dirk Larson, Rhett Ketterling, Robert A. Kyle, Shaji Kumar, Celine M. Vachon, S. Vincent Rajkumar. Evidence of decreased prevalence of the t(11;14) translocation in black patients with multiple myeloma: A multi-center study. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 272. doi:10.1158/1538-7445.AM2014-272

Abstract Background: Mammographic breast density (MBD) reflects the proportion of fibroglandular tissue in the breast, and is an established risk factor for breast cancer. Benign breast disease (BBD) is also a known risk factor for breast cancer. We previously showed an association between lobular involution (physiologic atrophy) and MBD in a large cohort study of women with BBD. We performed a comprehensive evaluation of the association of histologic features of BBD with MBD. Methods: The Mayo BBD cohort includes women, 18 to 85 years, who were diagnosed with BBD at Mayo Clinic in Rochester, MN between 01/01/1967 and 12/31/1991. For this study, we included women who had a mammogram within 6 months of BBD diagnosis, and diagnosed with BBD between 1985 and 1991 when parenchymal pattern of breast density was clinically recorded. Risk factor information such as age, and body mass index (BMI) was collected from medical records. Parenchymal pattern, assessed clinically and previously used in multiple studies of MBD and risk, classified breast density based on extent and type of density, into four categories: N1 (non-dense, no ducts visible); P1 (ductal prominence occupying <25% breast); P2 (prominent ductal pattern occupying >25% of the breast; DY (homogenous plaque-like areas of density). All assessments of benign breast tissue were performed by expert breast pathologists blinded to both MBD and BBD reports. Histologic characteristics included overall impression (non-proliferative disease (NPD), proliferative disease without atypia (PDWA), and atypical hyperplasia (AH)), proportion of normal lobules that were involuted (no (0%), partial (1 to 74%), or complete (≥75%)), type of AH (ADH or ALH), ductal and lobular hyperplasia, calcifications, sclerosing adenosis, columnar alteration, cyst, fibroadenoma, marked fibrosis, intra-ductal papilloma, radial scar (number and size), duct ectasia, and mucocele like tumors. Associations of parenchymal pattern with BBD characteristics were examined using multicategorical nominal logistic regression models. We first examined associations after adjustment for age and BMI. All variables statistically significant (p<0.05) in these models were then simultaneously included in a fully-adjusted model. Results: Of 2,257 women in the study, 14% were <40 years old, 52.7% between 40- 59 years, and 32.9% were ≥60 years. In this sample, 55.2% of women had NPD, 36% had PDWA, and AH in 8.8%. MBD was classified as N1 in 21.8%, P1 in 12.6%, P2 in 23.7% and DY in 41.9%. Age- and BMI-adjusted analyses showed that there was an association of parenchymal pattern with overall impression, lobular involution, ductal hyperplasia, sclerosing adenosis, columnar alteration, cyst, duct ectasia and fibrosis. Multivariate analyses found that women of younger age (p<0.001) and with lower BMI (p<0.001), presence of fibrosis (p<0.001) and no involution (p=0.016) were more likely to have P2 or DY parenchymal pattern than women without those characteristics. Conclusion: Among women with benign breast disease, higher MBD is associated with younger age, lower BMI, fibrosis, and lack of lobular involution. These findings, concordant with other studies of MBD, provide insight into underlying mechanisms by which MBD and BBD contribute to breast cancer risk. Citation Format: Karthik Ghosh, Robert A Vierkant, Ryan D Frank, Daniel W Visscher, Vernon S Pankratz, Christopher G Scott, Derek C Radisky, Marlene H Frost, Lynn C Hartmann, Amy C Degnim, Celine M Vachon. Association between mammographic breast density and histologic features of benign breast disease [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P6-10-14.

Abstract Introduction: Current models to predict breast cancer risk do not differentiate risk for estrogen receptor (ER) positive and negative breast cancer (BC), despite growing evidence that these tumors are biologically very different. We hypothesized that women with ER+ BC cancers have different clinical risk factors and histologic findings on prior benign breast biopsies than those with ER- BC. Methods: After IRB approval, we examined associations of age at benign biopsy and histologic features of the benign biopsy with ER status of incident BCs within the Mayo Benign Breast Disease Cohort. Benign biopsy slides were reviewed for extent of lobular involution and degree of epithelial proliferation by a single breast pathologist blinded to BC events. Invasive BCs occurring within 15 years after benign biopsy were classified as ER+ if ER staining was >1%. BC case-only associations with ER status were evaluated using multivariate logistic regression. Full-cohort hazard ratios (HR) and 95% confidence intervals (CI) for risk of ER-specific subtypes were estimated using Cox proportion hazards regression. Results: Among 13,410 women undergoing a benign breast biopsy from 1967-2001, 656 invasive BCs (459 ER+, 106 ER, 106 unknown) occurred within 15 years. Women who developed ER+ and ER- BCs were similar in age at the time of their prior benign breast biopsy (p=0.34). Although benign biopsies in women who later developed ER+ BC were more likely to show complete involution (23% vs 15% for ER- BC), this was not statistically significant (p=0.06). However, the degree of epithelial proliferation was significantly associated with ER status of later BCs (p=0.001), with ER+ BCs more likely than ER- BCs to have had a prior biopsy with atypical hyperplasia (16% vs 8%), and ER+ BCs less likely than ER- BCs to have had a prior biopsy with proliferative disease without atypia (33% vs 52%); this association remained after multivariate adjustment (p=0.003). We further pursued the association of epithelial proliferation with differential risk of ER+ and ER- BC in our overall cohort of 13,410 women (Table 1). Compared to women with non-proliferative disease, women with proliferative disease +/- atypia had ∼2-fold hazard ratios for ER- BC, whereas hazard ratios for ER+ BC were higher in women with atypical hyperplasia (∼4-fold) compared to proliferative disease. Hazard Ratio of ER+ and ER- BC within 15 years of BBD based upon degree of epithelial proliferation at benign biopsyCancer TypeGroupNN EventsHazard Ratio (95% CI)p-valueInvasive ER- ≤15 yearsNP8478421.00 (ref)<0.0001 PD4229542.63 (1.76, 3.93) AH70382.60 (1.22, 5.54) Invasive ER+ ≤15 yearsNP84782321.00 (ref)<0.0001 PD42291511.33 (1.09, 1.64) AH703744.41 (3.39, 5.73) NP=nonproliferative disease, PD=proliferative disease without atypia; AH=atypical hyperplasia Conclusion: ER+ and ER- breast cancers appear to have different features on prior benign breast biopsy, with atypical hyperplasia showing increased risk for both types of breast cancer, but a greater risk for ER+ tumors. Citation Format: Amy C Degnim, Derek C Radisky, Robert A Vierkant, Ryan D Frank, Marlene H Frost, Vernon S Pankratz, Celine M Vachon, Tanya L Hoskin, Julie M Cunningham, Chen Wang, Jean-Pierre Kocher, Teresa M Allers, Joanne L Johnson, Tina J Hieken, Karthik Ghosh, Lynn C Hartmann, Daniel W Visscher. Histologic features of benign breast biopsy tissue and association with ER positive and ER negative breast cancer in the Mayo BBD cohort study [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P6-10-06.

Abstract Purpose: Tamoxifen and raloxifene, are primary prevention strategies for women at high risk of breast cancer (Visvanathan, 2013). Recent advances in genetic studies of breast cancer risk have identified common susceptibility loci that explain 14% of familial risk for breast cancer in the general population (Michailidou, 2013). However, it is not known if these loci are risk factors for breast cancer among high-risk women treated with SERMs for breast cancer prevention. We hypothesized that the large risk reduction associated with SERMs, coupled with the fact that several breast cancer loci correlate with family history, may limit the contribution of these common genetic loci to breast cancer in this high risk population. We present the first report to evaluate 75 established breast cancer susceptibility loci, in the context of a polygenic risk score (PRS), as a risk factor for breast cancer among high risk women taking raloxifene and tamoxifen for breast cancer prevention. Methods: We conducted a matched case-control study of 594 cases (i.e., participants who developed breast cancer while on SERM therapy) and 1,171 matched controls selected from the 33,000 participants enrolled in the NSABP P-1 and P-2 breast cancer prevention trials. Genotypes of 75 single nucleotide polymorphisms (SNPs) were available from a genome-wide association study conducted at the RIKEN Center for Genomic Medicine. We formed a quantitative PRS from reported per-SNP odds ratios (OR) for the 75 susceptibility loci. Conditional logistic regression was used to examine the PRS as a risk factor for breast cancer and to assess whether the PRS and breast cancer association differed by treatment type, family history, or other clinical characteristics. Analyses also examined associations of PRS with invasive vs. in situ cancer and ER-positive vs. ER-negative cancer. Results: The PRS ranged from 3.98 to 7.74, and a one unit change in PRS was associated with a 42% increase in breast cancer (OR=1.42; 95% CI: 1.18-1.70; P = 0.0002). There was evidence of a stronger association of PRS with breast cancer among women with no first-degree family history (OR=1.62, 95% CI: 1.18-2.21) compared to those with a positive family history (OR=1.32, 95% CI: 1.06-1.66) (Pintx<0.05). The PRS also appeared a stronger risk factor for ER-positive (OR=1.59, 95% CI: 1.25-2.02, P < 0.0002) vs. ER-negative (OR=1.05, 95% CI: 0.68-1.62, P=0.84) breast cancer, although differences did not reach statistical significance (Pintx=0.10). PRS and breast cancer associations were similar across tamoxifen and raloxifene treatments, age at trial entry, 5-year predicted Gail model risk, hysterectomy status, BMI, presence of atypical hyperplasia and invasive vs. in situ cancer. Conclusion: A polygenic risk score composed of 75 loci was a risk factor for ER-positive breast cancer, especially in the absence of a first-degree family history of breast cancer. Further, the PRS associations with breast cancer were similar for women taking tamoxifen or raloxifene for prevention. These data suggest that common genetic variation adds information on risk of ER-positive breast cancer in a high-risk population receiving SERMs. Citation Format: Celine M Vachon, Daniel J Schaid, James N Ingle, Matthew P Goetz, Donald L Wickerham, Michiaki Kubo, Erin E Carlson, Soonmyung Paik, Norman Wolmark, Yusuke Nakamura, Liewei Wang, Richard M Weinshilboum, Fergus J Couch. The contribution of common genetic variation to breast cancer risk among women receiving tamoxifen or raloxifene within the National Surgical Adjuvant Breast and Bowel Project (NSABP) P-1 and P-2 trials [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P6-10-03.

Abstract Background: Guidelines recommend germline mutation testing of breast cancer predisposition genes in triple negative (TN) breast cancer cases with a family history of breast or ovarian cancer or when diagnosed under age 60. However, the prevalence of mutations in these genes among TN cases unselected for family history of breast or ovarian cancer is not known. Methods: To assess the frequency of mutations in 16 predisposition genes in TN cases we screened a large cohort of TN patients (n=1824) unselected for family history of breast or ovarian cancer from 12 centers and 824 study matched unaffected controls for mutations using a panel-based sequencing approach. Results: Deleterious mutations were identified in 15% of TN patients: 8.5% had BRCA1, 2.7% had BRCA2, and 3.6% had mutations in 12 other genes. Mutations in non-BRCA1/2 genes encoding proteins implicated in homologous recombination repair of DNA double strand breaks were detected at the same frequency as in breast cancer families. TN cases with mutations had high-grade tumors and were diagnosed at an earlier age than non-mutated cases. However, 10% of TN cases diagnosed at ≥60 years and 5% with no family history of cancer were also found to carry mutations. Inactivating mutations in non-BRCA1/2 predisposition genes were associated with moderate to high risks of TN breast cancer. Conclusions: National Comprehensive Cancer Network (NCCN) guidelines support clinical genetic testing of breast cancer predisposition genes in 95% of TN breast cancer patients carrying mutations in susceptibility genes. In contrast, National Institute of Health and Care Excellence (NICE) guidelines in the U.K. do not support genetic testing of a substantial proportion of TN patients with predisposing alleles. Frequency tables for inherited mutations in known predisposition genes based on age of diagnosis and family history of cancer will allow for selection of TN patients most likely to carry mutations in the predisposition genes. Citation Format: Fergus J Couch, Steven N Hart, Priyanka Sharma, Amanda Ewart Toland, Penelope Miron, Janet E Olson, Andrew Godwin, Vernon S Pankratz, Curtis Olswold, Seth Slettedahl, Lucia Guidugli, Matthias W Beckmann, Brigitte Rack, Arif B Ekici, Irene Konstantopoulou, Florentia Fostira, George Fountzilas, Liisa M Pelttari, Song Yao, Judy Garber, Angela Cox, Hiltrud Brauch, Christine Ambrosone, Heli Nevanlinna, Drakoulis Yannoukakos, Susan L Slager, Celine M Vachon, Diana M Eccles, Peter A Fasching. Triple-negative breast cancer: Frequency of inherited mutations in breast cancer susceptibility genes [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P4-12-03.

Abstract Introduction: We recently reported the percentage of mammographic density (MD) to be a breast cancer risk factor across tumor characteristics and age groups. Yet, we noted stronger associations for tumors of large size and positive lymph nodes across all ages, and ER-negative status among women ages <55 years, suggesting high MD may play an important role in tumor aggressiveness, especially in younger women. We extend this initial study to investigate whether these associations with aggressive disease are due to the underlying associations of absolute dense or non-dense area, or both, with breast cancer risk. Methods: Data were pooled from six studies including 3517 women with invasive breast cancer and 8558 without. Dense area and non-dense area were assessed from digitized film-screen mammograms using a computer-assisted threshold technique. We standardized dense and non-dense area measurements made within each study and categorized into quartiles based on the control distribution for pooled analyses. We used polytomous logistic regression to calculate breast cancer odds according to tumor type, histopathological characteristics, and receptor (ER, PR) status by age (<55, 55-64, and ≥65 years), adjusting for continuous age, body mass index, and study site. Results: Absolute dense area was associated with increased breast cancer risk [odds ratios (ORs) Q1 vs Q2: 0.67, Q3 vs Q2: 1.25, Q4 vs Q2: 1.62] and non-dense area was associated with decreased breast cancer risk (ORs: Q1 vs Q2: 1.44, Q3 vs Q2: 0.89, Q4 vs Q2: 0.71) overall and across all age groups. Positive associations for absolute dense area were observed across all invasive tumor characteristics. Stronger associations were noted for larger tumors across all ages (p-trend <0.01). Among women <55 years, stronger associations were observed for ER+ vs ER- [p-heterogeneity (het) = 0.02] and PR+ vs PR- tumors (p-het = 0.02). Non-dense area was inversely associated with all tumor characteristics evaluated except for tumors <1.1 cm, for which no association was present. The inverse association was stronger for larger tumors, particularly for women <55 years (p-trend <0.01). Also among women <55 years, a stronger inverse association with non-dense area was observed for ER- vs. ER+ tumors (p-het = 0.02). Conclusion: These results suggest the strong association we previously observed of high percent MD and ER-negative disease in women <55 years could be explained by the low amount of non-dense area in these women. Further research is warranted to clarify the possible differential associations of absolute dense and non-dense area on breast cancer risk according to tumor characteristics. Citation Format: Kimberly A. Bertrand, Christopher G. Scott, Rulla M. Tamimi, Matthew R. Jensen, V. Shane Pankratz, Aaron Norman, John Shepherd, Yunn-Yi Chen, Karla Kerlikowske, Celine M. Vachon. Dense and non-dense mammographic area and risk of breast cancer by age and tumor characteristics. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3279. doi:10.1158/1538-7445.AM2014-3279

Abstract OBJECTIVE: To determine the role of grape extract as a potential aromatase inhibitor by testing of plasma hormone levels. PROCEDURE: A six-week study was conducted involving postmenopausal women during which 94 grams of grape extract was consumed in addition to their usual diet. Plasma hormones were measured before and after treatment. RESULTS: Of the 18 women involved in the study, average age and BMI were 61.4 years and 24.4, respectively. For the hormone levels studied, the following median (IQR) percent change from baseline to 6 weeks values were found: estradiol +11.8% (-34.4%,+44.2%), p = 0.42; estrone +3.4% (-15.7%,+12.9%), p = 0.64; estrone sulfate +5.3% (-19.9%,+56.3%), p = 0.35; testosterone -1.5% (-14.7%,+10.7%), p = 0.97; androstenedione +12.6% (-17.1%, +49.1%), p = 0.15. The hormone levels did not significantly change between baseline and six weeks. Further, the changes that were observed did not tend to go in the hypothesized direction (estrogens and conjugates increased slightly, testosterone decreased slightly). Only androstenedione showed a trend toward change in the hypothesized direction. CONCLUSION: In this study, there was no evidence that plasma hormone levels are altered by 6 weeks of daily consumption of 94 grams of grape extract powder. Citation Format: Janet E. Olson, Summer V. Allen, Vera J. Suman, Tanya L. Hoskin, Celine M. Vachon, James N. Ingle. Evaluation of the aromatase inhibition potential of standardized grape extract. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3259. doi:10.1158/1538-7445.AM2014-3259

Abstract Mammographic density measures adjusted for age and body mass index (BMI) are heritable predictors of breast cancer risk, but few mammographic density-associated genetic variants have been identified. Using data for 10,727 women from two international consortia, we estimated associations between 77 common breast cancer susceptibility variants and absolute dense area, percent dense area and absolute nondense area adjusted for study, age, and BMI using mixed linear modeling. We found strong support for established associations between rs10995190 (in the region of ZNF365), rs2046210 (ESR1), and rs3817198 (LSP1) and adjusted absolute and percent dense areas (all P < 10−5). Of 41 recently discovered breast cancer susceptibility variants, associations were found between rs1432679 (EBF1), rs17817449 (MIR1972-2: FTO), rs12710696 (2p24.1), and rs3757318 (ESR1) and adjusted absolute and percent dense areas, respectively. There were associations between rs6001930 (MKL1) and both adjusted absolute dense and nondense areas, and between rs17356907 (NTN4) and adjusted absolute nondense area. Trends in all but two associations were consistent with those for breast cancer risk. Results suggested that 18% of breast cancer susceptibility variants were associated with at least one mammographic density measure. Genetic variants at multiple loci were associated with both breast cancer risk and the mammographic density measures. Further understanding of the underlying mechanisms at these loci could help identify etiologic pathways implicated in how mammographic density predicts breast cancer risk. Cancer Res; 75(12); 2457–67. ©2015 AACR.

Purpose Legislation mandating disclosure of breast density (BD) information has passed in 21 states; however, actual awareness of BD and knowledge of its impact on breast cancer detection and risk are unknown. Methods We conducted a national cross-sectional survey administered in English and Spanish using a probability-based sample of screening-age women, with oversampling of Connecticut, the only state with BD legislation in effect for > 1 year before the survey. Results Of 2,311 women surveyed, 65% responded. Overall, 58% of women had heard of BD, 49% knew that BD affects breast cancer detection, and 53% knew that BD affects cancer risk. After multivariable adjustment, increased BD awareness was associated with white non-Hispanic race/ethnicity (Hispanic v white non-Hispanic: odds ratio [OR], 0.23; P < .001), household income (OR, 1.07 per category increase; P < .001), education (OR, 1.19 per category increase; P < .001), diagnostic evaluation after a mammogram (OR, 2.64; P < .001), and postmenopausal hormone therapy (OR, 1.69; P = .002). Knowledge of the masking effect of BD was associated with higher household income (OR, 1.10; P < .001), education (OR, 1.22; P = .01), prior breast biopsy (OR, 2.16; P < .001), and residing in Connecticut (Connecticut v other states: OR, 3.82; P = .003). Connecticut residents were also more likely to have discussed their BD with a health care provider (67% v 43% for residents of other US states; P = .001). Conclusion Disparities in BD awareness and knowledge exist by race/ethnicity, education, and income. BD legislation seems to be effective in increasing knowledge of BD impact on breast cancer detection. These findings support continued and targeted efforts to improve BD awareness and knowledge among women eligible for screening mammography.

Abstract Background: Mammographic density (MD) is a strong breast cancer risk factor. We previously reported associations of percent mammographic density (PMD) with larger and node-positive tumors across all ages, and estrogen receptor (ER)–negative status among women ages &lt;55 years. To provide insight into these associations, we examined the components of PMD [dense area (DA) and nondense area (NDA)] with breast cancer subtypes. Methods: Data were pooled from six studies including 4,095 breast cancers and 8,558 controls. DA and NDA were assessed from digitized film-screen mammograms and standardized across studies. Breast cancer odds by density phenotypes and age according to histopathologic characteristics and receptor status were calculated using polytomous logistic regression. Results: DA was associated with increased breast cancer risk [OR for quartiles: 0.65, 1.00 (Ref), 1.22, 1.55; Ptrend &lt;0.001] and NDA was associated with decreased risk [ORs for quartiles: 1.39, 1.00 (Ref), 0.88, 0.72; Ptrend &lt;0.001] across all ages and invasive tumor characteristics. There were significant trends in the magnitude of associations of both DA and NDA with breast cancer by increasing tumor size (Ptrend &lt; 0.001) but no differences by nodal status. Among women &lt;55 years, DA was more strongly associated with increased risk of ER+ versus ER− tumors (Phet = 0.02), while NDA was more strongly associated with decreased risk of ER− versus ER+ tumors (Phet = 0.03). Conclusions: DA and NDA have differential associations with ER+ versus ER− tumors that vary by age. Impact: DA and NDA are important to consider when developing age- and subtype-specific risk models. Cancer Epidemiol Biomarkers Prev; 24(5); 798–809. ©2015 AACR.