Protective role of abundant healthy stroma against breast cancer: A prospective cohort study of healthy women in the Susan G. Komen Tissue Bank

Background Breast cancer (BC) remains a leading cause of morbidity and mortality in women worldwide. To mitigate this trend, future BC risk reduction strategies need to be grounded in BC etiologic histogenesis, i.e., the tissue pathways in the normal breast by which etiologic factors impact BC risk. Results from previous studies have indicated that the tissue composition of the breast may offer clues into the etiologic histogenesis of BC. However, many prior studies were conducted among women with benign breast disease (BBD), a tissue state characterized by underlying pathological changes that may obscure studies of etiologic histogenesis. Here, we investigated the associations of histologic tissue composition metrics of the normal breast with subsequent BC risk. Methods This prospective cohort study includes 3,435 healthy females (aged 18-75 years) who donated normal breast tissue to the Susan G. Komen Tissue Bank (KTB) between 2009-2019 and met study-specific inclusion criteria. Of these, 109 women developed BC after a median follow-up of 6 years (range: 1-16 years). Supervised machine-learning algorithms were used to analyze digitized histologic sections of normal breast tissues to generate quantitative data on percent epithelium, stroma, and adipose tissue for each woman. Associations between histologic metrics and breast cancer incidence were assessed in multivariable Cox proportional hazard regression models with age as the time-scale. Analyses were performed overall and following stratification by age at study entry and menopausal status. Models were adjusted for race, education, parity, body mass index, smoking status, alcohol use, and whole slide tissue area. Results Overall, abundant stroma was associated with decreased BC risk (hazard ratio (HR) per 1 standard deviation (SD) increase = 0.55, 95% confidence interval [CI] = 0.38 to 0.80, P < 0.01). The stromal protective effect was most profound among women younger than 55 years at tissue donation (HR/1-SD = 0.45, 95% CI = 0.26 to 0.79, P < 0.01) and in premenopausal women (HR/1-SD = 0.48, 95% CI = 0.26 to 0.87, P = 0.015), irrespective of age. Elevated epithelium and adipose tissue were associated with significantly increased BC risk in these same groups. However, when including stroma and adipose tissue in the same regression model, the mutual adjustment for stromal tissue mitigated adipose-related risk in women overall (HR/1-SD = 0.26, 95% CI = 0.06 to 1.16, P = 0.077), women younger than 55 years (HR/1-SD = 0.17, 95% CI = 0.04 to 0.71, P = 0.015), and in premenopausal women (HR/1-SD = 0.16, 95% CI = 0.04 to 0.65, P = 0.01). Conclusion Findings from this cohort of healthy women provide insights into the role of stroma in BC histogenesis during normal homeostasis, with potential implications for BC risk prediction and risk-reduction strategies.