Abstract Introduction: Paget's classical seed and soil hypothesis states that specific interactions between the metastatic cell (the seed) and the recipient tissue microenvironment (the soil) mediate the non-random patterns of spread observed in breast cancer (tropism). We hypothesize that the ability of metastatic cells to survive dormancy, exit quiescence, and colonize a specific tissue depends upon the ability of the soil to first sustain survival, and subsequently trigger outgrowth. To this end, we created a biomaterial platform that varies extracellular matrix (ECM) density and composition, and used integrin-mediated phenotyping to study how highly metastatic cells differentiate between secondary sites via integrin binding. With this relationship in hand, we have found that both the ECM and growth factors also mediate dormancy in cells that display metastatic latency in vivo. Methods: ECM microenvironments were created by linking combinations of ECM proteins to glass coverslips using silane chemistry. Human breast cancer cell lines were seeded onto the surfaces, and we quantified initial cell adhesion and long-term migration via microscopy. Using cell lines that only metastasize to the bone, brain, or lung (provided by J. Massagué), we created fingerprints of behavior characteristic of each tropism, and compared these with the phenotypes of other cell lines with known metastasis. Quiescence was induced on ECMs via culture in serum free medium for up to 14 days, and subsequent serum-stimulation for allowed for recovery and growth. Proliferation was assayed via immunofluorescent staining for Ki67, and cell number was determined via DAPI staining. Results: We identified unique phenotypes of bone, brain, and lung tropic subpopulations, and confirmed these in seven of nine other known, more heterogeneous, metastatic cell lines, establishing a connection between in vitro phenotype and in vivo fate. Function-affecting antibodies to integrins each shift a cell line's predicted tropism in vitro, and these in vitro phenotyping results mimic the clinical metastasis patterns dictated by both α2 and α6 integrin gene expression. Interestingly, however, our functional predictions regarding β1 integrin actually provide insights that are largely independent of clinical gene expression, potentially identifying new metastatic therapeutic targets, and emphasizing the need for this type of functional analysis. However, when we applied this approach to cell lines that display latency in vivo (e.g., ZR-75-1), our predictions were not correct. These cells were generally insensitive to ECM and growth factors in adhesion and motility, but instead, we found that these factors regulate entrance into dormancy. We uniformly observe that a laminin-coupled ECM is unable to permit survival of dormant cells, whereas binding to collagen I maintains cell survival for at least 14 days. Epidermal growth factor supplementation improves the entrance into dormancy on collagen alone. We have found that this collagen-specific, EGF-enhanced survival ability is mediated by β1 integrin binding and activation of ERK1/2, suggesting that β1 integrin function is critical both for aggressive metastasis and for permitting dormancy. Conclusions: We are the first to report that bone, brain, and lung tropism can be distinguished via simple integrin-mediated phenotyping, and that this approach can be used to predict in vivo tropism. We suggest that functional β1 integrin binding mediates both aggressive and latent metastasis. In sum, we propose that this functional screen of cell-matrix interactions can predict in vivo outcomes, and thus, is a valuable tool to provide insight toward integrins as druggable targets for metastatic disease. Citation Format: Lauren E. Barney, Ari Gilman, Arthur Mercurio, Shelly R. Peyton. Extracellular matrix control of metastasis and dormancy. [abstract]. In: Proceedings of the AACR Special Conference on Tumor Metastasis; 2015 Nov 30-Dec 3; Austin, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(7 Suppl):Abstract nr PR14.