Abstract Background: Malignant transformation involves progressive changes in cellular morphology. However, the point along the early carcinogenesis trajectory when these changes first signal cancer risk and etiologic origins remains unknown. Methods: In this prospective cohort study of 3,415 healthy women who donated normal breast tissue (NBT) to the US-based Susan G. Komen Tissue Bank (KTB) between 2009-2019, machine learning algorithms were used to characterize 72,237,858 cells in digitized histologic sections of NBTs. Occult alterations in epithelial nuclear area, perimeter, chromasia, nuclear contour irregularity, epithelial nuclear density, and stromal cellular density defined three cytomorphologic signatures that were denoted by their histogenesis i.e., epithelial, stromal, and mixed (epithelial and stromal). Associations of cytomorphologic signatures with subsequent breast cancer (BC) incidence were assessed using multivariable Cox proportional hazard regression with age as timescale and left truncated at study entry. Analyses were performed overall and by tumor subtype (estrogen receptor-positive (ER+) and ER-). Results: After a median follow-up of 6 years (range:1-16 years) post-donation, 104 previously healthy women developed BC. Women with the stromal (hazard ratio (HR) and 95% confidence interval (CI) = 3.01 1.56-5.80; P=0.001), epithelial (HR= 2.28 1.31-3.96; P=0.003), or mixed (HR= 2.18 1.18-4.03; P=0.010) signatures in their NBT had more than double the risk of BC than those without any signature. RNA sequencing analyses revealed the stromal signature to be characterized by heightened expression of inflammation, angiogenesis, and epithelial-to-mesenchymal transition genes. Etiologically, the stromal signature was associated with younger age, Black race, multiparity, contraceptive pill use, and BC family history, and in subtype-specific analysis, it was most strongly predictive of aggressive, ER- BC (HR= 9.04 3.22-25.39; P0.0001). The epithelial signature (reflecting age-related epithelial functional decline) demonstrated contrasting etiologic profile from the stromal signature, and was most strongly associated with future ER+ BC, particularly when detected in stroma-poor/adipose-rich microenvironments (HR= 7.04 1.49-33.19; P=0.01). The mixed signature, which was subtype-agnostic in its BC predictive value, was characterized by heightened expression of wound response genes, shorter (-3 years) latency to BC diagnosis than the epithelial and stromal signatures (P0.0001), and by its association with obesity and older age (25 years) at first childbirth. Conclusion: Oncogenic processes may imprint occult cytomorphologic signatures in breast tissues long before cancer diagnosis, with implications for BC risk prediction, targeted prevention, and early interception strategies. Citation Format: Mustapha Abubakar, Scott M. Lawrence, Jeri Danielle Hughes, Shaoqi Fan, Karun Mutreja, Ruth M. Pfeiffer, Gretchen L. Gierach, Brittany Jenkins, Jill E. Henry, Máire A. Duggan. Single-cell pathology of the normal breast uncovers occult cytomorphologic signatures that signal future cancer risk and etiologic origins abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 4072.
Abubakar et al. (Fri,) studied this question.