Characterizing spatial tumor architecture and ECM recapitulation using an EO771 syngeneic TNBC model.

e13133 Background: Triple-negative breast cancer (TNBC) remains an aggressive malignancy with limited therapeutic options and high rates of treatment resistance. Increasing evidence indicates that therapeutic response is strongly influenced by tumor microenvironment (TME) architecture, including extracellular matrix (ECM) composition, spatial organization, and tumor-stroma interactions that regulate drug transport, mechanotransduction, and immune infiltration. While syngeneic models (such as EO771) are widely used to understand therapeutic response due to their immunocompetent context, the extent to which these systems recapitulate the spatial tumor architecture and ECM organization observed in patient TNBC remains poorly defined. Quantitative, spatially resolved assessments of ECM recapitulation across human and syngeneic TNBC models are limited, representing a critical step toward better understanding the tumor biology and optimizing clinical treatment strategies. Methods: Primary human TNBC specimens (n = 14) and EO771 (also referred to as E0771 in literature) syngeneic TNBC tumors (n = 3) were analyzed using quantitative spatial tumor mapping. Human specimens were formalin-fixed, paraffin-embedded tissues; syngeneic tumors were generated by orthotopic implantation of EO771 cells into the mammary fat pad of C57BL/6 mice and harvested at matched tumor burden. Dual-stain histology was performed to visualize collagen (Picrosirius Red) and glycosaminoglycans (Alcian Blue). Whole-slide brightfield images were acquired (10x resolution) and analyzed using an automated, blinded computational pipeline to extract ECM features including collagen area fraction, peripheral versus intratumoral localization, GAG content and distribution, and collagen:GAG ratio. Identical analytical parameters were applied across all specimens. Results: Human TNBC specimens exhibited heterogeneous ECM distribution with notable intratumoral collagen and GAG presence, whereas EO771 tumors showed more uniform ECM and collagen was concentrated at the peripheral tumor margin. Quantitative metrics, including collagen area fraction and localization indices, highlighted these spatial differences, with human TNBC demonstrating higher intratumoral ECM heterogeneity compared with the syngeneic model. Conclusions: These data characterize distinct patterns of spatial tumor architecture and ECM organization in human TNBC versus a syngeneic mouse model of the disease, supporting further evaluation of a recapitulation gap and its implications for therapeutic response modeling. Quantitative spatial mapping may inform selection and refinement of preclinical models to better reflect human TME features relevant to therapy resistance phenotypes, and support the integration of human-relevant new approach methods (NAMs) to complement established translational research frameworks.