Abstract Introduction: The tumor microenvironment is known to play an influential role in the angiogenic and metastatic progression of cancer and is regulated by different factors (stromal fibroblasts, extracellular matrix (ECM) proteins and endothelial cells) present in the complex milieu. Recapitulation of this complexity in three-dimensional (3D) tumor models is critical in understanding the processes involved in cancer progression and to provide clinically relevant efficacy data for cancer drugs. To address this challenge, we developed a microfluidic oncomimetic platform where breast cancer cells are co-cultured with fibroblasts along with a complex, intricate vascular network. We further investigated the effect of mechanical stiffness of the tumor stroma on the growth and morphology of cancer cells, migration of cancer cells into surrounding vasculature and the ability of standard cancer drugs to perfuse through the vasculature and target cancer cells. Materials and Methods: Poly(dimethyl siloxane) (PDMS)-based microfluidic devices, containing vascular networks in communication with tumor chamber, were fabricated using photolithography as described earlier1. Poly(ethylene glycol)-fibrinogen (PEG-Fb), used for 3D co-culture of cancer cells and fibroblasts, was prepared using established protocols2. Human breast tumor associated endothelial cells (hBTECs) were seeded within fibronectin-coated vascular channels and maintained under flow to develop endothelial networks. To obtain 3D cancer-fibroblast co-culture system, MCF7 or MDA-MB-231 breast cancer cells were co-encapsulated with BJ-5ta human fibroblasts in PEG-Fb hydrogel in the tumor chamber. Immunostaining with standard endothelial and cancer markers was conducted to confirm maturity and functionality of seeded cells. GFP-labelled cancer cells were co-cultured with hBTECs to visualize tumor cell migration. Cancer cells were also maintained in culture over several weeks within the devices to demonstrate applicability of this system to perform long-term drug dosing experiments. Finally, the cytotoxic effects of doxorubicin and paclitaxel at two different concentrations on cancer cells were evaluated by perfusing the drugs through the endothelial channels and cell viability quantified via Live/Dead staining. Conclusions: We have developed a novel 3D microfluidic, vascularized cancer-fibroblast co-culture platform, with the ability to predict drug efficacy for breast cancer. This platform can also be extended in future for cancer-immunotherapy based investigations and screening of novel nano-carrier based anti-cancer drugs. Acknowledgements: We gratefully acknowledge financial support from NIH (#HHSN261201400037C) and AURIC.