Differences in Genome-wide DNA Methylation Profiles in Breast Milk by Race and Lactation Duration

Black women in the United States are disproportionately affected by early-onset, triple-negative breast cancer. DNA methylation has shown differences by race in healthy and tumor breast tissues. We examined associations between genome-wide DNA methylation levels in breast milk and breast cancer risk factors, including race, to explain how this reproductive stage influences a woman's risk for, and potentially contributes to racial disparities in, breast cancer. Breast milk samples and demographic, behavioral, and reproductive data, were obtained from cancer-free, uniparous, and lactating U.S. black (n = 57) and white (n = 82) women, ages 19-44. Genome-wide DNA methylation analysis was performed on extracted breast milk DNA using the Infinium HumanMethylation450 BeadChip. Statistically significant associations between breast cancer risk factors and DNA methylation beta values, adjusting for potential confounders, were determined using linear regression followed by Bonferroni Correction (P < 1.63 × 10^-7). Epigenetic analysis in breast milk revealed statistically significant associations with race and lactation duration. Of the 284 CpG sites associated with race, 242 were hypermethylated in black women. All 227 CpG sites associated with lactation duration were hypomethylated in women who lactated longer. Ingenuity Pathway Analysis of differentially methylated promoter region CpGs by race and lactation duration revealed enrichment for networks implicated in carcinogenesis. Associations between DNA methylation and lactation duration may offer insight on its role in lowering breast cancer risk. Epigenetic associations with race may mediate social, behavioral, or other factors related to breast cancer and may provide insight into potential mechanisms underlying racial disparities in breast cancer incidence.

Reported Incidence and Survival of Fallopian Tube Carcinomas: A Population-Based Analysis From the North American Association of Central Cancer Registries

Background

Recognition that serous tubal intraepithelial carcinoma (STIC) may represent the first manifestation of many high-grade cancers that were once considered ovarian primary tumors has led to changes in diagnostic practices that could dramatically increase the reporting of tubal carcinomas in US population-based cancer registries. Further, increased detection of early-stage tubal carcinomas through increased recognition coupled with meticulous pathology processing protocols raises important unanswered questions about the clinical behavior of such lesions, which can only be answered using large data sets. However, rates of tubal carcinomas have not been recently analyzed. Accordingly, we analyzed population-based incidence and survival data for fallopian tube carcinoma in situ (CIS; an imperfect surrogate of STIC), tubal carcinomas, and for comparison, ovarian carcinomas, in the North American Association of Central Cancer Registries (NAACCR) registries.

Methods

Total counts, standardized incidence rates, and stage-specific survival were computed using 30 NAACCR registries (1999-2012). Temporal incidence rate patterns were analyzed by joinpoint regression with estimates of annual percentage change (APC). All statistical tests were two-sided.

Results

Fallopian tube CIS incidence rates were stable from 1999 to 2002, then increased from 2002 to 2012 (APC = 16.2%, 95% confidence interval [CI] = 10.9% to 21.7%, P < .001). Rates of early- and late-stage tubal carcinomas showed similar patterns, whereas high-grade serous ovarian carcinoma rates were relatively stable. Five-year cause-specific survival was 97.9% (95% CI = 93.7% to 99.3%) for tubal CIS and 83.2% (95% CI = 77.3% to 87.7%) for early-stage high-grade serous tubal carcinoma.

Conclusions

Reporting of tubal CIS and tubal carcinoma have increased in recent years, likely reflecting changes in pathology processing of specimens and diagnosis. Developing standardized reporting for tubal neoplasms is needed to enable analysis of outcomes for these comparatively uncommon but increasingly recognized tumors.

Relationship of circulating insulin-like growth factor-I and binding proteins 1-7 with mammographic density among women undergoing image-guided diagnostic breast biopsy

Background

Mammographic density (MD) is a strong breast cancer risk factor that reflects fibroglandular and adipose tissue composition, but its biologic underpinnings are poorly understood. Insulin-like growth factor binding proteins (IGFBPs) are markers that may be associated with MD given their hypothesized role in breast carcinogenesis. IGFBPs sequester IGF-I, limiting its bioavailability. Prior studies have found positive associations between circulating IGF-I and the IGF-I:IGFBP-3 ratio and breast cancer risk. We evaluated the associations of IGF-I, IGFBP-3, and six other IGFBPs with MD.

Methods

Serum IGF measures were quantified in 296 women, ages 40–65, undergoing diagnostic image-guided breast biopsy. Volumetric density measures (MD-V) were assessed in pre-biopsy digital mammograms using single X-ray absorptiometry. Area density measures (MD-A) were estimated by computer-assisted thresholding software. Age, body mass index (BMI), and BMI²-adjusted linear regression models were used to examine associations of serum IGF measures with MD. Effect modification by BMI was also assessed.

Results

IGF-I and IGFBP-3 were not strongly associated with MD after BMI adjustment. In multivariable analyses among premenopausal women, IGFBP-2 was positively associated with both percent MD-V (β = 1.49, p value = 0.02) and MD-A (β = 1.55, p value = 0.05). Among postmenopausal women, positive relationships between IGFBP-2 and percent MD-V (β = 2.04, p = 0.003) were observed; the positive associations between IGFBP-2 and percent MD-V were stronger among lean women (BMI < 25 kg/m²) (β = 5.32, p = 0.0002; p interaction = 0.0003).

Conclusions

In this comprehensive study of IGFBPs and MD, we observed a novel positive association between IGFBP-2 and MD, particularly among women with lower BMI. In concert with in vitro studies suggesting a dual role of IGFBP-2 on breast tissue, promoting cell proliferation as well as inhibiting tumorigenesis, our findings suggest that further studies assessing the role of IGFBP-2 in breast tissue composition, in addition to IGF-1 and IGFBP-3, are warranted.

Electronic supplementary material

The online version of this article (10.1186/s13058-019-1162-8) contains supplementary material, which is available to authorized users.

Abstract 3331: Association of lifestyle and clinical characteristics with receipt of radiotherapy treatment among women diagnosed with DCIS in the National Institutes of Health-AARP Diet and Health Cohort Study

Introduction: The long-term risks and benefits of breast radiotherapy for the treatment of ductal carcinoma in situ (DCIS) remain unclear. A recent study using data from the Surveillance, Epidemiology and End Results (SEER) registries showed that DCIS-associated radiotherapy treatment significantly increased risks of second non-breast cancers. To help disentangle those observations and understand whether breast cancer risk factors are related to radiotherapy treatment decision making, we aimed to identify factors associated with receipt of radiotherapy treatment for DCIS.

Methods: Accordingly, among 1,628 women with DCIS who participated in the large prospective National Institutes of Health (NIH)-AARP (formerly American Association of Retired Persons) Diet and Health Cohort Study, that were diagnosed between study entry and follow up (through 31st December 2011) and eligible for inclusion, we examined associations between demographic, lifestyle and clinical factors and receipt of radiotherapy treatment. Odds ratios and 95% confidence intervals (CIs) were estimated from multivariable logistic regression models, to determine associations.

Results: Among women diagnosed with DCIS, 45% (n=730) received radiotherapy treatment. No relationships were observed between receipt of radiotherapy and lifestyle factors including smoking status. The strongest associations were observed for clinical factors. Receipt of radiotherapy was associated with a more recent diagnoses period (2005-2011 versus 1995-1999; OR=1.60, 95%CI: 1.14, 2.25), diagnosis within eastern (Michigan/New Jersey) versus western US states, (OR=1.44, 95%CI: 1.07, 1.93), poorly than well-differentiated tumours (OR=1.69, 95%CI: 1.16, 2.46) and receipt of endocrine therapy (OR=3.37, 95%CI: 2.56, 4.44).

Discussion: Clinical characteristics of the DCIS and temporal trends in radiotherapy use were the strongest determinants of radiotherapy treatment within this population. Receipt of radiotherapy was largely unrelated to lifestyle factors suggesting that the previously observed associations in SEER between radiotherapy treatment and risk of non-breast cancers were likely not confounded by lifestyle factors including smoking status. Thus, these findings are relevant for understanding the aetiology of subsequent secondary cancers among the increasing population of DCIS survivors. The lack of association observed for breast cancer family history and prior breast biopsy suggest that these risk factors do not influence radiotherapy clinical decision making.

Citation Format: Maeve Mullooly, Diana R. Withrow, Rochelle E. Curtis, Shaoqi Fan, Linda M. Liao, Ruth M. Pfeiffer, Amy Berrington de Gonzalez, Gretchen L. Gierach. Association of lifestyle and clinical characteristics with receipt of radiotherapy treatment among women diagnosed with DCIS in the National Institutes of Health-AARP Diet and Health Cohort Study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3331.

Abstract 589: Circulating progesterone is associated with increased postmenopausal breast cancer risk: B~FIT cohort

BACKGROUND: Epidemiologic data provide compelling evidence of the association between elevated endogenous estrogens and androgens and increased postmenopausal breast cancer risk. However, the role of progesterone remains largely unexplored, primarily due to limitations in assay sensitivity and precision of progesterone measurements at low concentrations in postmenopausal women. Recently identified progesterone metabolites may provide etiologic insights as experimental data suggest that relative changes in concentrations of 5-α dihydroprogesterone (5αP) and 3-α dihydroprogesterone (3αHP) reflect cancer promoting and cancer inhibiting properties, respectively.

METHODS: We developed a sensitive and reliable liquid chromatography-tandem mass spectrometry (LC-MS/MS) assay and quantified prediagnostic levels of progesterone/progesterone metabolites in a case-cohort study within the Breast and Bone Follow-up to the Fracture Intervention Trial (B~FIT) including 405 breast cancer cases diagnosed during follow-up and a subcohort of 495 postmenopausal women not using exogenous hormones at blood draw. Multivariable adjusted hazard ratios (HR) and 95% confidence intervals (CIs) were estimated using Cox regression and linearity was assessed using splines.

RESULTS: Hormone concentrations among women in the subcohort were on average 4.6 pg/mL (range 2.3-21.6) for progesterone; all measured values were above the assay detection limit. Women with higher circulating progesterone levels had an elevated postmenopausal breast cancer risk [HR (95% CI) per 10 pg/mL increase in progesterone: 1.18 (0.99-1.41)]. This association was linear in a 5-knot spline and strengthened [1.24 (1.03-1.49)] in models excluding women who reported current use of tamoxifen/raloxifene. Higher levels of 5αP relative to 3αHP were associated with a 4% increased postmenopausal breast cancer risk [per unit increase in ratio: 1.04 (1.00-1.07)]. For the individual metabolites, higher levels of both 5αP [per 10 pg/mL increase: 1.04 (0.93-1.18) and 3αHP [per 3 pg/mL increase: 1.11 (0.999-1.24)] were associated with elevated risk. All associations remained after adjustment for circulating estrogen levels.

CONCLUSIONS: Our prospective data suggest that postmenopausal women with increased serum progesterone concentrations, measured using a highly sensitive LC-MS/MS assay, are at increased risk of breast cancer. Consistent with experimental studies, higher levels of 5αP relative to 3αHP were indicative of increased breast cancer risk. Unlike experimental studies, our data do not suggest that endogenous concentrations of 3αHP are associated with reduced breast cancer risk; instead both metabolites were associated with elevated risk. The identification of these risk-related progesterone metabolites supports the need for additional research regarding their role in the etiology of breast cancer.

Citation Format: Britton Trabert, Doug C. Bauer, Louise A. Brinton, Diane S. Buist, Jane A. Cauley, Cher M. Dallal, Gretchen L. Gierach, Roni T. Falk, Trisha F. Hue, James V. Lacey, Andrea Z. LaCroix, Jeffrey A. Tice, Xia Xu. Circulating progesterone is associated with increased postmenopausal breast cancer risk: B~FIT cohort [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 589.

Abstract 588: Relationship of serum progesterone and progesterone metabolites with mammographic density

Background:

Mammographic breast density (MBD) is a strong breast cancer (BC) risk factor, but its biologic underpinnings are poorly understood. Use of estrogen plus progestin menopausal hormone therapy is linked to increased MBD and BC risk. Experimental data suggest that ratios of tumor promoting (5α-dihydroprogesterone [5αP]) and anti-tumorigenic (3α-dihydroprogesterone [3αP]) progesterone metabolite levels may be related to BC risk. However, relationships of endogenous progesterone and its metabolites with MBD and BC risk have not been established. Accordingly, we assessed levels of circulating progesterone and its metabolites with MBD.

Methods:

In this cross-sectional study, serum progesterone and its metabolites were quantified using a novel liquid chromatography-tandem mass spectrometry assay in 103 postmenopausal and 52 premenopausal (luteal menstrual cycle phase) women, ages 40-65, undergoing diagnostic image-guided ipsilateral breast biopsy. MBD was measured as percent fibroglandular volume (MBD-V) on pre-biopsy digital mammograms using single X-ray absorptiometry. Square-root transformed MBD-V was examined across tertile categories of progesterone/progesterone metabolites using age and body mass index (BMI)-adjusted linear regression models.

Results:

Concentrations of the hormones were as follows among postmenopausal women: progesterone [mean: 12.6 pmol/L (range: 5.2-45.8)], 3αP [5.6 pmol/L (1.4-18.8)], 5αP [100 pmol/L (16.7-388)], and 5αP/3αP ratio [26.1 (2.1-150)]; and among luteal phase premenopausal women: progesterone [2063 pmol/L (13.6-7098), 3αP [12.7 pmol/L (2.4-64.4)], 5αP [243 pmol/L (25.3-774)], 5αP/3αP ratio [25.9 (2.3-73.7)]. Among postmenopausal women, progesterone was positively associated with MBD-V (Tertile 3 vs. 1: β=0.68, p-trend=0.02). A similar borderline positive association was observed among premenopausal women (β=0.74, p-trend=0.10). Additional adjustment for circulating estradiol did not substantively alter observed associations. Levels of 3αP, 5αP and the 5αP/3αP ratio were not associated with MBD-V among pre- or postmenopausal women.

Conclusions:

Concentrations of progesterone and it metabolites show substantial inter-woman variation. We observed a positive association between endogenous progesterone and MBD-V among both postmenopausal and premenopausal luteal phase women. We did not observe an association with the ratio of 5αP to 3αP levels and MBD-V. These findings suggest the need for additional studies to understand the biological basis of the role of progesterone and its metabolites in MBD and BC risk.

Citation Format: Manila Hada, Hannah Oh, Sharon Fan, Roni T. Falk, Berta Geller, Pamela Vacek, Donald Weaver, John Shepherd, Jeff Wang, Bo Fan, Amir P. Mahmoudzadeh, Serghei Malkov, Sally Herschorn, Louise A. Brinton, Xia Xu, Mark E. Sherman, Britton Trabert, Gretchen L. Gierach. Relationship of serum progesterone and progesterone metabolites with mammographic density [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 588.

Associations between mammographic density and tumor characteristics in Chinese women with breast cancer

PURPOSE

Mammographic density (MD) is a strong risk factor for breast cancer, yet its relationship with tumor characteristics is not well established, particularly in Asian populations.

METHODS

MD was assessed from a total of 2001 Chinese breast cancer patients using Breast Imaging Reporting and Data System (BI-RADS) categories. Molecular subtypes were defined using immunohistochemical status on ER, PR, HER2, and Ki-67, as well as tumor grade. Multinomial logistic regression was used to test associations between MD and molecular subtype (luminal A = reference) adjusting for age, body mass index (BMI), menopausal status, parity, and nodal status.

RESULTS

The mean age at diagnosis was 51.7 years (SD = 10.7) and the average BMI was 24.7 kg/m² (SD = 3.8). The distribution of BI-RADS categories was 7.4% A = almost entirely fat, 24.2% B = scattered fibroglandular dense, 49.4% C = heterogeneously dense, and 19.0% D = extremely dense. Compared to women with BI-RADS = A/B, women with BI-RADS = D were more likely to have HER2-enriched tumors (OR = 1.81, 95% CI 1.08-3.06, p = 0.03), regardless of menopausal status. The association was only observed in women with normal (< 25 kg/m²) BMI (OR = 2.43, 95% CI 1.24-4.76, p < 0.01), but not among overweight/obese women (OR: 0.98, 95% CI 0.38-2.52, p = 0.96).

CONCLUSIONS

Among Chinese women with normal BMI, higher breast density was associated with HER2-enriched tumors. The results may partially explain the higher proportion of HER2+ tumors previously reported in Asian women.

Abstract P1-08-04: Withdrawn

This abstract was withdrawn by the authors.

Citation Format: Trabert B, Bauer DC, Brinton LA, Buist DS, Cauley JA, Dallal CM, Gierach GL, Falk RT, Hue TF, Lacey, Jr. JV, LaCroix AZ, Tice JA, Xu X. Withdrawn [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P1-08-04.

The relationship between terminal duct lobular unit features and mammographic density among Chinese breast cancer patients

Extensive mammographic density (MD), a well-established breast cancer risk factor, is a radiological representation of stromal and epithelial breast tissue content. In studies conducted predominantly among Caucasian women, histologic measures of reduced terminal duct lobular unit (TDLU) involution have been correlated with extensive MD, but independently associated with breast cancer risk. We therefore examined associations between TDLU measures and MD among Chinese women, a low-risk population but with high prevalence of dense breasts. Diagnostic pre-treatment digital mammograms were obtained from 144 breast cancer cases at a tertiary hospital in Beijing and scored using the Breast Imaging Reporting and Data System (BI-RADS) density classification. TDLU features were assessed using three standardized measures (count/100 mm² , span [μm], and acini count/TDLU) in benign tissues. Associations between each of TDLU measures and MD were examined using generalized linear models for TDLU count and span and polytomous logistic regression for acini count with adjustment for potential confounders stratified by age. Among women ≥50 years, 63% had dense breasts; cases with dense breasts (BI-RADS, c-d) had greater TDLU count (21.1 [SE = 2.70] vs. 9.0 [SE = 1.83]; p = 0.0004), longer span (480.6 μm [SE = 24.6] vs. 393.8 μm [SE = 31.8]; p = 0.03), and greater acini count (OR_trend = 16.1; 95%CI = 4.08-63.1; p_trend < 0.0001) compared to those with non-dense breasts (BI-RADS, a-b). Among women <50 years, 91% had dense breasts, precluding our ability to detect associations. Our findings are consistent with previously reported associations between extensive MD and reduced TDLU involution, supporting the hypothesis that breast cancer risk associated with extensive MD may be related to the amount of "at-risk" epithelium.

Pooled Analysis of Nine Cohorts Reveals Breast Cancer Risk Factors by Tumor Molecular Subtype

Various subtypes of breast cancer defined by estrogen receptor (ER), progesterone receptor (PR), and HER2 exhibit etiologic differences in reproductive factors, but associations with other risk factors are inconsistent. To clarify etiologic heterogeneity, we pooled data from nine cohort studies. Multivariable, joint Cox proportional hazards regression models were used to estimate HRs and 95% confidence intervals (CI) for molecular subtypes. Of 606,025 women, 11,741 invasive breast cancers with complete tissue markers developed during follow-up: 8,700 luminal A-like (ER+ or PR+/HER2-), 1,368 luminal B-like (ER+ or PR+/HER2+), 521 HER2-enriched (ER-/PR-/HER2+), and 1,152 triple-negative (ER-/PR-/HER2-) disease. Ever parous compared with never was associated with lower risk of luminal A-like (HR, 0.78; 95% CI, 0.73-0.83) and luminal B-like (HR, 0.74; 95% CI, 0.64-0.87) as well as a higher risk of triple-negative disease (HR, 1.23; 95% CI, 1.02-1.50; P value for overall tumor heterogeneity < 0.001). Direct associations with luminal-like, but not HER2-enriched or triple-negative, tumors were found for age at first birth, years between menarche and first birth, and age at menopause (P value for overall tumor heterogeneity < 0.001). Age-specific associations with baseline body mass index differed for risk of luminal A-like and triple-negative breast cancer (P value for tumor heterogeneity = 0.02). These results provide the strongest evidence for etiologic heterogeneity of breast cancer to date from prospective studies.Significance: These findings comprise the largest study of prospective data to date and contribute to the accumulating evidence that etiological heterogeneity exists in breast carcinogenesis. Cancer Res; 78(20); 6011-21. ©2018 AACR.

Using deep convolutional neural networks to identify and classify tumor-associated stroma in diagnostic breast biopsies

The breast stromal microenvironment is a pivotal factor in breast cancer development, growth and metastases. Although pathologists often detect morphologic changes in stroma by light microscopy, visual classification of such changes is subjective and non-quantitative, limiting its diagnostic utility. To gain insights into stromal changes associated with breast cancer, we applied automated machine learning techniques to digital images of 2387 hematoxylin and eosin stained tissue sections of benign and malignant image-guided breast biopsies performed to investigate mammographic abnormalities among 882 patients, ages 40-65 years, that were enrolled in the Breast Radiology Evaluation and Study of Tissues (BREAST) Stamp Project. Using deep convolutional neural networks, we trained an algorithm to discriminate between stroma surrounding invasive cancer and stroma from benign biopsies. In test sets (928 whole-slide images from 330 patients), this algorithm could distinguish biopsies diagnosed as invasive cancer from benign biopsies solely based on the stromal characteristics (area under the receiver operator characteristics curve = 0.962). Furthermore, without being trained specifically using ductal carcinoma in situ as an outcome, the algorithm detected tumor-associated stroma in greater amounts and at larger distances from grade 3 versus grade 1 ductal carcinoma in situ. Collectively, these results suggest that algorithms based on deep convolutional neural networks that evaluate only stroma may prove useful to classify breast biopsies and aid in understanding and evaluating the biology of breast lesions.

Pro-inflammatory cytokines and growth factors in human milk: an exploratory analysis of racial differences to inform breast cancer etiology

BACKGROUND

Analysis of cytokines and growth factors in human milk offers a noninvasive approach for studying the microenvironment of the postpartum breast, which may better reflect tissue levels than testing blood samples. Given that Black women have a higher incidence of early-onset breast cancers than White women, we hypothesized that milk of the former contains higher levels of pro-inflammatory cytokines, adipokines, and growth factors.

METHODS

Participants included 130 Black and 162 White women without a history of a breast biopsy who completed a health assessment questionnaire and donated milk for research. Concentrations of 15 analytes in milk were examined using two multiplex and 4 single-analyte electrochemiluminescent sandwich assays to measure pro-inflammatory cytokines, angiogenesis factors, and adipokines. Mixed-effects ordinal logistic regression was used to identify determinants of analyte levels and to compare results by race, with adjustment for confounders. Factor analysis was used to examine covariation among analytes.

RESULTS

Thirteen of 15 analytes were detected in ≥ 25% of the human milk specimens. In multivariable models, elevated BMI was significantly associated with increased concentrations of 5 cytokines: IL-1β, bFGF, FASL, EGF, and leptin (all p-trend < 0.05). Black women had significantly higher levels of leptin and IL-1β, controlling for BMI. Factor analysis of analyte levels identified two factors related to inflammation and growth factor pathways.

CONCLUSION

This exploratory study demonstrated the feasibility of measuring pro-inflammatory cytokines, adipokines, and angiogenesis factors in human milk, and revealed higher levels of some pro-inflammatory factors, as well as increased leptin levels, among Black as compared with White women.

Serum insulin-like growth factor (IGF)-I and IGF binding protein-3 in relation to terminal duct lobular unit involution of the normal breast in Caucasian and African American women: The Susan G. Komen Tissue Bank

Lesser degrees of terminal duct lobular unit (TDLU) involution, as reflected by higher numbers of TDLUs and acini/TDLU, are associated with elevated breast cancer risk. In rodent models, the insulin-like growth factor (IGF) system regulates involution of the mammary gland. We examined associations of circulating IGF measures with TDLU involution in normal breast tissues among women without precancerous lesions. Among 715 Caucasian and 283 African American (AA) women who donated normal breast tissue samples to the Komen Tissue Bank between 2009 and 2012 (75% premenopausal), serum concentrations of IGF-I and binding protein (IGFBP)-3 were quantified using enzyme-linked immunosorbent assay. Hematoxilyn and eosin-stained tissue sections were assessed for numbers of TDLUs ("TDLU count"). Zero-inflated Poisson regression models with a robust variance estimator were used to estimate relative risks (RRs) for association of IGF measures (tertiles) with TDLU count by race and menopausal status, adjusting for potential confounders. AA (vs. Caucasian) women had higher age-adjusted mean levels of serum IGF-I (137 vs. 131 ng/mL, p = 0.07) and lower levels of IGFBP-3 (4165 vs. 4684 ng/mL, p < 0.0001). Postmenopausal IGFBP-3 was inversely associated with TDLU count among AA (RR_T3vs.T1  = 0.49, 95% CI = 0.28-0.84, p-trend = 0.04) and Caucasian (RR_T3vs.T1 =0.64, 95% CI = 0.42-0.98, p-trend = 0.04) women. In premenopausal women, higher IGF-I:IGFBP-3 ratios were associated with higher TDLU count in Caucasian (RR_T3vs.T1 =1.33, 95% CI = 1.02-1.75, p-trend = 0.04), but not in AA (RR_T3vs.T1 =0.65, 95% CI = 0.42-1.00, p-trend = 0.05), women. Our data suggest a role of the IGF system, particularly IGFBP-3, in TDLU involution of the normal breast, a breast cancer risk factor, among Caucasian and AA women.

Abstract 3260: Utilizing digital pathology to understand breast epithelial characteristics of benign breast disease among women undergoing diagnostic image-guided breast biopsy

Introduction: Defining the histologic correlates of mammographic breast density (MD) may provide insights into why elevated MD is related to increased breast cancer risk. Studies suggest that reduced terminal ductal lobular (TDLU) involution is associated with elevated MD, and both are independent breast cancer risk factors among women who have undergone a biopsy for benign breast disease (BBD). Prior digital histologic analyses of normal breast tissues revealed epithelial nuclear density (END) and TDLU involution are correlated. Accordingly, we examined associations of END with TDLU involution and MD in clinical biopsies. Methods: We analyzed 262 image-guided breast biopsies diagnosed as BBD from 224 women. TDLU involution was visually assessed as TDLU count/mm2 and TDLU span (inversely related to level of involution) in background normal tissue, evaluated using digitized images. The Genie Classifier (Aperio) was applied to images to estimate nuclei count per unit epithelial area, (END). Single X-ray Absorptiometry of pre-biopsy craniocaudal digital mammograms was applied to measure global MD (percent fibroglandular volume (%FGV)). Analysis of covariance, adjusted for age and body mass index, examined mean END differences across tertiles of TDLU/MD measures. Analyses were conducted at the biopsy level using SAS PROC GENMOD to account for within-woman correlations. All tests were two-tailed. Results: Overall, 67% of BBD biopsies were proliferative. Higher END was observed among proliferative than non-proliferative BBD (median END: 10,187 vs. 9,953 respectively; p=0.04). Among all women, END significantly increased with increasing tertiles of TDLU measures (p-trends: TDLU count/100mm2=0.0001, TDLU span=0.046). Whereas TDLU metrics were positively associated with %FGV, no relationship was observed between END and %FGV. In analyses stratified by BBD severity, however, END and %FGV were positively associated among women with non-proliferative disease (p-trend=0.04), findings not observed with proliferative disease. Conclusions: Automated END and visually assessed TDLU involution metrics were positively associated with each other and with MD. However, associations were diluted for proliferative lesions, suggesting that applying automated digital pathology tools to unsegmented digital images of whole sections of BBD biopsies does not demonstrate the same associations with MD as visual assessment of TDLU involution.

Citation Format: Maeve Mullooly, Samantha Puvanesarajah, Shaoqi Fan, Ruth M. Pfeiffer, Linnea Olsson, Manila Hada, Erin L. Kirk, Pamela M. Vacek, Donald L. Weaver, John A. Shepherd, Amir P. Mahmoudzadeh, Jeff Wang, Stephen M. Hewitt, Sally D. Herschorn, Mark E. Sherman, Melissa A. Troester, Gretchen L. Gierach. Utilizing digital pathology to understand breast epithelial characteristics of benign breast disease among women undergoing diagnostic image-guided breast biopsy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3260.

Abstract A36: Serum insulin-like growth factor (IGF)-I and IGF binding protein-3 in relation to terminal duct lobular unit involution in Caucasian and African American women: The Susan G. Komen Tissue Bank

Background: Insulin-like growth factor (IGF)-I and IGF binding protein (IGFBP)-3 play important roles in carcinogenesis, particularly for breast cancer. However, little is known about whether the IGF system influences histologic characteristics of normal glandular tissue and whether relationships vary by race. Lesser degrees of age-related terminal duct lobular unit (TDLU) involution, as reflected by higher numbers of TDLUs and acini per TDLU, have been associated with higher breast cancer risk. We examined the associations of IGF measures with TDLU involution of normal breast using standardized TDLU measures.

Methods: Among 715 Caucasian and 283 African American (AA) women with normal breast tissue samples from the Komen Tissue Bank, serum concentrations of IGF-I and IGFBP-3 were quantified using enzyme-linked immunosorbent assay (ELISA). Hematoxilyn and eosin-stained tissue sections were assessed for numbers of TDLUs (“TDLU count”) and acini/TDLU. Zero-inflated Poisson regression models with a robust variance estimator were used to estimate associations of IGF-I, IGFBP-3, and IGF-I:IGFBP-3 molar ratio (tertiles) with TDLU count by race and menopausal status, adjusting for potential confounders. We also tested for interactions by race using likelihood ratio tests.

Results: AA (vs. Caucasian) women had higher age-adjusted mean levels of serum IGF-I (137 vs. 131 ng/mL, p=0.07) and lower levels of IGFBP-3 (4165 vs. 4684 ng/mL, p&lt;0.0001); the differences persisted after adjustment for additional covariates including BMI and parity/age at first birth. Postmenopausal IGFBP-3 was inversely associated with TDLU count among both AA (RR T3vs.T1=0.49, 95% CI=0.28-0.84, p-trend=0.04) and Caucasian (RR T3vs.T1=0.63, 95% CI=0.41-0.99, p-trend=0.04) women. In premenopausal women, higher IGF-I:IGFBP-3 ratios were associated with higher TDLU count in Caucasians (OR T3vs.T1=1.33, 95% CI=1.01-1.31, p-trend=0.04) but not in AA (OR T3vs.T1=0.65, 95% CI=0.42-1.00, p-trend=0.05) women. There was no statistically significant interaction by race (p-interaction≥0.10).

Conclusions: Our data suggest the potential role of the IGF system, particularly IGFBP-3, in TDLU involution of the normal breast among both Caucasian and AA women.

Citation Format: Hannah OH, Ruth M. Pfeiffer, Roni T. Falk, Hisani N. Horne, Jackie Xiang, Michael Pollak, Louise A. Brinton, Anna Maria V. Storniolo, Mark E. Sherman, Gretchen L. Gierach, Jonine D. Figueroa. Serum insulin-like growth factor (IGF)-I and IGF binding protein-3 in relation to terminal duct lobular unit involution in Caucasian and African American women: The Susan G. Komen Tissue Bank [abstract]. In: Proceedings of the Tenth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2017 Sep 25-28; Atlanta, GA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2018;27(7 Suppl):Abstract nr A36.

Stroma modifies relationships between risk factor exposure and age-related epithelial involution in benign breast

Delayed age-related lobular involution has been previously associated with elevated breast cancer risk. However, intraindividual variability in epithelial involution status within a woman is undefined. We developed a novel measure of age-related epithelial involution, density of epithelial nuclei in epithelial areas using digital image analysis in combination with stromal characteristics (percentage of section area comprising stroma). Approximately 1800 hematoxylin and eosin stained sections of benign breast tissue were evaluated from 416 participants having breast surgery for cancer or benign conditions. Two to sixteen slides per woman from different regions of the breast were studied. Epithelial involution status varied within a woman and as a function of stromal area. Percentage stromal area varied between samples from the same woman (median difference between highest and lowest stromal area within a woman was 7.5%, but ranged from 0.01 to 86.7%). Restricting to women with at least 10% stromal area (N = 317), epithelial nuclear density decreased with age (-637.1 cells/mm² per decade of life after age 40, p < 0.0001), increased with mammographic density (457.8 cells/mm² per increasing BI-RADs density category p = 0.002), and increased non-significantly with recent parity, later age at first pregnancy, and longer and more recent oral contraceptive use. These associations were attenuated in women with mostly fat samples (<10% stroma (N = 99)). Thirty-one percent of women evaluated had both adequate stroma (≥10%) and mostly fat (<10% stroma) regions of breast tissue, with the probability of having both types increasing with the number breast tissue samplings. Several breast cancer risk factors are associated with elevated age-related epithelial content, but associations depend upon stromal context. Stromal characteristics appear to modify relationships between risk factor exposures and breast epithelial involution.

Intra-individual Gene Expression Variability of Histologically Normal Breast Tissue

Several studies have sought to identify novel transcriptional biomarkers in normal breast or breast microenvironment to predict tumor risk and prognosis. However, systematic efforts to evaluate intra-individual variability of gene expression within normal breast have not been reported. This study analyzed the microarray gene expression data of 288 samples from 170 women in the Normal Breast Study (NBS), wherein multiple histologically normal breast samples were collected from different block regions and different sections at a given region. Intra-individual differences in global gene expression and selected gene expression signatures were quantified and evaluated in association with other patient-level factors. We found that intra-individual reliability was relatively high in global gene expression, but differed by signatures, with composition-related signatures (i.e., stroma) having higher intra-individual variability and tumorigenesis-related signatures (i.e., proliferation) having lower intra-individual variability. Histological stroma composition was the only factor significantly associated with heterogeneous breast tissue (defined as > median intra-individual variation; high nuclear density, odds ratio [OR] = 3.42, 95% confidence interval [CI] = 1.15–10.15; low area, OR = 0.29, 95% CI = 0.10–0.86). Other factors suggestively influencing the variability included age, BMI, and adipose nuclear density. Our results underscore the importance of considering intra-individual variability in tissue-based biomarker development, and have important implications for normal breast research.

Abstract P6-10-07: Differences in genome-wide DNA methylation levels in breast milk by race and lactation duration

Introduction: Identifying biomarkers of breast cancer risk among young women would have value in developing effective screening and prevention strategies at early ages. We have proposed that DNA methylation analysis of breast milk may provide breast cancer risk information among young women, and could possibly provide etiologic clues related to the higher rates of early onset cancers among African American as compared with White women in the US.

Objective: The purpose of this project was to identify associations between genome-wide DNA methylation levels in breast milk and race adjusted for other breast cancer risk factors.

Study Population: Cancer-free, lactating U.S. Black (n=57) and White (n=82) women, ages 19 to 44, provided frozen breast milk samples, as well as demographic, behavioral, and reproductive data, to the Breastmilk Laboratory at University of Massachusetts Amherst. Women were uniparous and did not have a personal history of breast cancer at the time of milk donation.

Methods: DNA was extracted from breast milk samples using the phenol-chloroform method. Genome-wide methylation analysis was performed on breast milk samples using the Infinium HumanMethylation450 BeadChip. Probes with 50% or more missing data, cross-reactive probes, as well as probes with minor allelic frequency greater than 0.05 in European- or African-Americans were removed, leaving 379,042 CpG sites for analysis. Multivariate generalized linear regression models were used to examine associations between race and other breast cancer risk factors and methylation beta values, adjusting for potential confounding factors. P-values less than 1E-7 were considered statistically significant.

Results: Black women in this study were more likely to be never smokers, to not have used over-the-counter pain medication in the past week, and to breastfeed longer. After adjustment by age, BMI, smoking status, and batch number, race was significantly associated with differential methylation at 1143 CpG sites, including 1024 at which Black women demonstrated increased methylation levels. Additionally, breastfeeding duration was associated with 269 CpG sites, with 268 showing a significant inverse relationship with methylation. Methylation sites significantly associated with Black race and lactation duration were located within tumor suppressor and promoter genes as well as in genes implicated in obesity and diabetes.

Conclusion: This preliminarily analysis of DNA methylation in breast milk suggests that Black women have increased methylation and longer breastfeeding is associated with reduced methylation. Further research to understand how etiologic factors related to breast cancer may alter DNA methylation patterns in normal breast may lead to improved understanding of breast cancer risk at a young age and potentially causes of racial disparities in breast cancer incidence between White and Black women.

Citation Format: Davis Lynn BC, Bodelon C, Pfeiffer RM, Yang HP, Yang H, Lee M, Laird PW, Campan M, Weisenberger DJ, Murphy J, Sherman ME, Browne EP, Anderton DL, Arcaro KF, Gierach GL. Differences in genome-wide DNA methylation levels in breast milk by race and lactation duration [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P6-10-07.

Epidemiologic Risk Factors for In Situ and Invasive Breast Cancers Among Postmenopausal Women in the National Institutes of Health-AARP Diet and Health Study

Comparing risk factor associations between invasive breast cancers and possible precursors may further our understanding of factors related to initiation versus progression. Accordingly, among 190,325 postmenopausal participants in the National Institutes of Health-AARP Diet and Health Study (1995-2011), we compared the association between risk factors and incident ductal carcinoma in situ (DCIS; n = 1,453) with that of risk factors and invasive ductal carcinomas (n = 7,525); in addition, we compared the association between risk factors and lobular carcinoma in situ (LCIS; n = 186) with that of risk factors and invasive lobular carcinomas (n = 1,191). Hazard ratios and 95% confidence intervals were estimated from multivariable Cox proportional hazards regression models. We used case-only multivariable logistic regression to test for heterogeneity in associations. Younger age at menopause was associated with a higher risk of DCIS but lower risks of LCIS and invasive ductal carcinomas (P for heterogeneity < 0.01). Prior breast biopsy was more strongly associated with the risk of LCIS than the risk of DCIS (P for heterogeneity = 0.04). Increased risks associated with use of menopausal hormone therapy were stronger for LCIS than DCIS (P for heterogeneity = 0.03) and invasive lobular carcinomas (P for heterogeneity < 0.01). Associations were similar for race, age at menarche, age at first birth, family history, alcohol consumption, and smoking status, which suggests that most risk factor associations are similar for in situ and invasive cancers and may influence early stages of tumorigenesis. The differential associations observed for various factors may provide important clues for understanding the etiology of certain breast cancers.

Divergent oestrogen receptor-specific breast cancer trends in Ireland (2004-2013): Amassing data from independent Western populations provide etiologic clues

The aetiology and clinical behaviour of breast cancers vary by oestrogen receptor (ER) expression, HER2 expression and over time. Data from the United States and Denmark show rising incidence rates for ER+ and falling incidence rates for ER- breast cancers. Given that Ireland is a somewhat similar Western population but with distinctive risk exposures (especially for lactation), we analysed breast cancer trends by ER status; and for the first time, by the joint expression of ER±/HER2±. We assessed invasive breast cancers (n = 24,845; 2004-2013) within the population-based National Cancer Registry of Ireland. The population at risk was obtained from the Irish Central Statistics Office (n = 10,401,986). After accounting for missing ER and HER2 data, we assessed receptor-specific secular trends in age-standardised incidence rates (ASRs) with the estimated annual percentage change (EAPC) and corresponding 95% confidence intervals (95% CI). Age-period-cohort models were also fitted to further characterise trends accounting for age, calendar-period and birth-cohort interactions. ASRs increased for ER+ (EAPC: 2.2% per year [95% CI: 0.97, 3.45%/year]) and decreased for ER- cancers (EAPC: -3.43% per year [95% CI: -5.05, -1.78%/year]), as well as for specific age groups at diagnosis (<30-49, 50-64 and ≥65 years). ER+/HER2- cancers rose, ER+/HER2+ cancers were statistically flat and ER-/HER± cancers declined. Secular trends for ER± cancers in Ireland were like those previously observed. Stratification by HER2± expression did not substantively alter ER± trends. The divergence of ER± incidence rates among independent Western populations likely reflects calendar-period and/or risk factor changes with differential effects for ER+ and ER- breast cancers.

Abstract 4235: Application of convolutional neural networks to breast biopsies to uncover tissue correlates of mammographic breast density

Background: High percent mammographic density (MD), which reflects the relative fibroglandular tissue content of the breast, is one of the strongest breast cancer risk factors; however, the pathologic mediators of this risk are unknown. We hypothesize that analysis of breast tissue sections using deep learning approaches may characterize histologic features that underpin risk associated with high MD.

Methods: Non-targeted H&E stained breast tissue sections of diagnostic image-guided breast biopsies were evaluated among 588 women enrolled following an abnormal mammogram in the Breast Radiology Evaluation and Study of Tissues (BREAST) Stamp Project (2007-2010). Overall volumetric percent MD for the biopsied breast and localized volumetric percent MD surrounding the biopsy site were determined for each participant. A deep convolutional neural network (CNN) model was trained to identify and quantitatively assess breast epithelial, stroma and fat tissue and their organizational and spatial arrangements. Least absolute shrinkage and selection operator (Lasso) regression was used to determine relationships between MD measures and pathological features. To ensure reliability of the model, a cross-validation strategy was employed to build and assess the performance of the fitted model. Finally, Spearman correlation coefficients were estimated to test the association between the predicted density values by each model (predicting overall or localized MD) and the actual MD measurements. We report the average and standard deviation (SD) of the correlation coefficients.

Results: In an independent validation set, the CNN model was 95.5% accurate in classifying epithelial, stromal and fat tissue. The mean (SD) correlations between the predicted model and the actual measurements for overall and localized MD were 0.70 (0.06) and 0.65 (0.06) respectively. The amount of stroma identified (normalized to tissue area) had the highest selection probability (P-value) by the Lasso model and thus the strongest positive relationship with MD (P-value &gt;0.9 for each MD measurement). In contrast, the amount of normalized epithelial tissue was not related to MD (P-value=0.01 for each MD measurement). No association was observed for the total normalized fat area with MD (P-value &lt;0.31 for each MD measurement). In addition, the number of distributed epithelial regions was positively associated, whereas the distance between epithelial regions was inversely associated with overall MD (P-value &lt;0.87 and 0.62, respectively).

Conclusions: These results show that greater stromal tissue amount and spatial distribution patterns of epithelial regions, rather than total epithelial amounts, had the strongest relationships with elevated MD. Future work will determine the relationship of these MD features with biopsy diagnosis.

Citation Format: Maeve Mullooly, Babak Ehteshami Bejnordi, Maya Palakal, Pamela M. Vacek, Donald L. Weaver, John A. Shepherd, Bo Fan, Amir Pasha Mahmoudzadeh, Jeff Wang, Jason M. Johnson, Sally D. Herschorn, Brian L. Sprague, Ruth M. Pfeiffer, Louise A. Brinton, Mark E. Sherman, Andrew Beck, Gretchen L. Gierach. Application of convolutional neural networks to breast biopsies to uncover tissue correlates of mammographic breast density [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4235. doi:10.1158/1538-7445.AM2017-4235

Abstract 257: Estrogen receptor positive and negative breast cancer secular trends in Ireland (2004-2013)

Background: Estrogen receptor (ER) status in breast cancer is strongly associated with etiologic factors, prognosis and treatment response. Data suggest that population-based incidence rates of ER-positive (ER+) tumors are increasing and ER-negative (ER-) cancers have fallen over time in the United States (US), Denmark and Scotland. Whether these trends are consistent in other populations with different etiologic exposures and across ER±/HER2± subtypes is unknown. We, therefore, analyzed breast cancer secular trends in Ireland by ER expression; and for the first time, by joint expression of ER±/HER2±, a well-established clinical marker of uncertain etiological significance.

Methods: This study included invasive breast cancers (n=24,845; 2004-2013) among women aged 20-84 years within the population-based National Cancer Registry of Ireland (97% coverage). The general population at risk (n=10,401,986) was obtained from the Irish Central Statistics Office. Using a validated approach to account for missing ER/HER2-status, we assessed ER± and ER±/HER2± secular trends with age-standardized incidence rates (ASR) and estimated annual percentage change (EAPC) with 95% confidence intervals (95%CI).

Results: Over the 10-year follow-up period, 19,264 ER+ and 4,161 ER- breast cancers were diagnosed (1,420 were ER-unknown). ASRs increased significantly for ER+ (EAPC: 2.20%/year (95%CI: 0.97, 3.45%/year)) and decreased significantly for ER- cancers (EAPC: -3.43%/year (95%CI: -5.05, -1.78%/year)). The ER-specific trends were qualitatively similar among three age groups (&lt;50, 50-64 and ≥65 years). With further stratification by HER2± expression, ER+/HER2- cancers rose (EAPC: 2.87%/year (95%CI: 1.33, 4.44%/year)), ER+/HER2+ cancers were statistically flat, while ER-/HER± cancers significantly declined.

Conclusion: Secular trends for ER± cancers in Ireland were similar to those previously observed in the US and Denmark. Further, trends for ER/HER2 followed those for ER, although rates were constant for ER+/HER+ tumors. The divergence of cancer incidence patterns by ER among independent populations over time could reflect comparable changes in risk factor exposures with dual or opposite effects for risk of ER+ and ER- cancers (e.g., obesity, age at first birth/parity). Additionally, although HER2 is a well-established clinical predictive and prognostic marker for breast cancer outcome, it did not appear to substantively impact breast cancer incidence trends.

Citation Format: Maeve Mullooly, Jeanne Murphy, Gretchen L. Gierach, Brittny Davis, Paul Walsh, Sandra Deady, Thomas I. Barron, Mark E. Sherman, Philip S. Rosenberg, William F. Anderson. Estrogen receptor positive and negative breast cancer secular trends in Ireland (2004-2013) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 257. doi:10.1158/1538-7445.AM2017-257

DEEP LEARNING-BASED ASSESSMENT OF TUMOR-ASSOCIATED STROMA FOR DIAGNOSING BREAST CANCER IN HISTOPATHOLOGY IMAGES

Diagnosis of breast carcinomas has so far been limited to the morphological interpretation of epithelial cells and the assessment of epithelial tissue architecture. Consequently, most of the automated systems have focused on characterizing the epithelial regions of the breast to detect cancer. In this paper, we propose a system for classification of hematoxylin and eosin (H&E) stained breast specimens based on convolutional neural networks that primarily targets the assessment of tumor-associated stroma to diagnose breast cancer patients. We evaluate the performance of our proposed system using a large cohort containing 646 breast tissue biopsies. Our evaluations show that the proposed system achieves an area under ROC of 0.92, demonstrating the discriminative power of previously neglected tumor associated stroma as a diagnostic biomarker.