Abstract GS2-01: The breast pre-cancer atlas illustrates the molecular and micro-environmental diversity of ductal carcinoma in situ

Background: With the implementation of widespread breast cancer screening, the diagnosis of Ductal Carcinoma In Situ (DCIS) has increased 5 fold. Most cases are treated with combinations of surgery, radiation and endocrine therapy, reducing the risk of second events, including ipsilateral invasive cancer. Without standard markers to confidently identify the most indolent lesions, a subset of cases are likely over-treated. The mutational landscapes of DCIS and invasive ductal carcinoma (IDC) are similar and not sufficient to identify higher risk lesions with recent studies suggesting that clonal selection plays a limited role in progression. Histological analysis highlighted the role of the extracellular matrix and immune-surveillance to maintain duct integrity and limit progression. Due to the small size, limited availability and quality of research specimens, few biomarker studies investigated pure DCIS with adequate follow-up or genome-wide methods, let alone integrated more than one type of biomarker. Unlike breast cancer, there is no comprehensive, systematic, multi-modal atlas of DCIS, limiting the ability to test broad and novel hypotheses and characterize processes maintaining breast tissue homeostasis. Methods: Through sequential sectioning of pure DCIS archived specimens, a total of 70 histological regions from 40 cases were annotated and profiled using up to 3 platforms: multiplex immuno-histochemistry (mIHC), RNA-seq, and whole-exome sequencing. Stromal and epithelial spatial distribution of immune cells and states were quantified using mIHC. The epithelial compartments were microdissected and profiled using genome-wide gene expression, DNA mutations and copy number alterations. Results: Epithelial regions were classified according to expression subtypes consistent with histological markers, highlighting associations with the lesion architecture and grade. Compared to solid pattern, cribriform pattern DCIS has induced EMT processes and repressed proliferation processes, a trend reminiscent of low-recurrence-risk expression signatures measured in IDC. The DNA copy number burden increased with grade and the mutational burden was the highest in solid DCIS. Both were higher in Her2+ cases. The clustering of mutations at chromosome 17p - attributed to the APOBEC-driven Kataegis phenomenon - was observed in a subset of regions, suggesting this event can occur in pre-invasive lesions. All DCIS had somatic alterations of at least one known driver gene with some associated with grade (TP53) or expression subtype (ERBB2). Multi-region profiling available in a subset of samples revealed genetic heterogeneity of likely-driver events between proximal regions of similar histological characteristics. The density and proliferative states of selected immune cells - including T-cells, B-cells and Macrophages - highlights the diversity of the tumor immune environment with the highest densities observed in Her2+ ducts and stroma, minimal ductal infiltration in other lesions, fewer dividing B- and T-cells around the more proliferative areas and a small number of regions depleted from any adaptive immune cells. Conclusion: This first multi-modal profiling of pure DCIS reveals an unsuspected diversity of molecular and microenvironmental states and presents their association with progression risk factors. The observations support the need for stronger integration of molecular and clinicopathology features, especially at sub-histological levels, to ensure the findings can be interpreted in the correct clinical and phenotypic context. The compatibility of the approach with archived specimens supports the expansion to larger retrospective DCIS collections with outcomes. Citation Format: Olivier Harismendy, Daniela Nachmanson, Mark F. Evans, Hidetoshi Mori, Adam Officer, Christina Yau, Joseph Steward, Huazhen Yao, Thomas O9Keefe, Farnaz Hasteh, Gary S. Stein, Kristen Jepsen, Michael Campbell, Donald L. Weaver, Gillian L. Hirst, Brian L. Sprague, Laura J. Esserman, Jonathan A. Gordon, Alexander Borowsky, Janet L. Stein. The breast pre-cancer atlas illustrates the molecular and micro-environmental diversity of ductal carcinoma in situ [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr GS2-01.