Epidemiologic effectors on the breast tumor microenvironment: Results from the Cancer Prevention Study-II and-3 cohorts.

10530 Background: Despite breast cancer being the leading cancer among women, we have a poor understanding of the impact of epidemiologic risk factors on its tumor microenvironment (TME). We examined risk factors for TME favorability using a novel aggregate marker: the Histomic Prognostic Score (HiPS). Methods: Data were from whole slide image (WSI) scans of hematoxylin & eosin slides from women in the Cancer Prevention Studies CPS-II and -3, diagnosed with nonmetastatic breast cancer. Diagnoses occurred between enrollment and the last administrative date (1992/1993 to 2011 in CPS-II and 2006–2013 to 2018 in CPS-3). Median year of diagnosis was 2004 in CPS-II and 2014 in CPS-3. Using the WSI, a deep-learning model was trained to extract features and summarize the morphology of the epithelium, stroma, and tumor-infiltrating lymphocytes within the TME. These features, along with demographic, tumor biomarkers, and survival data, were then used to generate the HiPS, a novel measurement of “ prognostic favorability” of TME, ranging from 0 (favorable) to 10 (unfavorable). Multivariable linear models estimated associations between the HiPS score and epidemiologic risk factors: hormone replacement therapy use (HRT; current vs. former/never), cigarette smoking (current vs. former/never), alcohol intake (>1 drink vs. <1 drink/day), aspirin intake (daily vs. not daily) controlled for, age at diagnosis, body mass index (BMI), race, and cancer detection method (screen- vs. self-detected). Results: We included 1264 CPS-II and 651 CPS-3 participants in the analysis. Compared to CPS-II, CPS-3 participants were younger (mean (± sd) age 54.7 ± 7.6 vs. 72.2 ± 7, p<0.001), heavier (mean (± sd) BMI 27.1 ± 5.8 vs. 26 ± 4.8, p<0.001), and more diverse (7.4% vs. 2.1% non-White, p<0.001). Most breast cancers (69.9% of CPS-II and 64.1% of CPS-3) were screen-detected, p=0.206. Increasing age and BMI were associated with higher HiPS score in both cohorts, Table. Daily aspirin use and current HRT use were associated with lower HiPS scores in CPS-3 only. The inverse association for HRT use was lost in the following subgroups age < 50, ER-, HER2+, TNM stage-1, and screen-detected cancers. Conclusions: Increasing age and BMI have independent adverse effects on breast TME. Daily aspirin intake and current HRT use were only associated with a favorable TME in the younger, more contemporary cohort. Additional studies are needed to confirm these findings and better understand the role of these epidemiologic factors on the TME. [Table: see text]