The in vivo tumor microenvironment is a complex niche that includes heterogeneous physical structures,unique bio-chemical gradients and multiple cell interactions.Its high-fidelity in vitro reconstruction is of fundamental importance to improve current understandings of cell behavior,efficacy predictions and drug safety.In this study,we have developed a high-throughput biochip with hundreds of composite extracellular matrix (ECM) microchambers to co-culture invasive breast cancer cells (MDA-MB-231-RFP) and normal breast epithelial cells (MCF-10A-GFP).The composite ECM is com-posed of type Ⅰ collagen and Matrigel which provides a heterogeneous microenvironment that is similar to that of in vivo cell growth.Additionally,the growth factors and drug gradients that involve human epidermal growth factor (EGF),discoidin domain receptor 1 (DDR1) inhibitor 7rh and matrix metalloproteinase inhibitor batimastat allow for the mimicking of the complex in vivo biochemical microenvironment to investigate their effect on the spatial-temporal dynamics of cell growth.Our results demonstrate that the MDA-MB-231-RFP cells and MCF-10A-GFP cells exhibit different spatial proliferation behaviors under the combination of growth factors and drugs.Basing on the experimental data,we have also developed a cellular automata (CA) model that incorporated drug diffusion to describe the experimental phenomenon,as well as em-ployed Shannon entropy (SE) to explore the effect of the drug diffusion coefficient on the spatial-temporal dynamics of cell growth.The results indicate that the uniform cell growth is related to the drug diffusion coefficient,which reveals that the pore size of the ECM plays a key role in the formation of complex biochemical gradients.Therefore,our integrated,biomimetic and high-throughput co-culture platforms,as well as the computational model can be used as an effective tool for investigating cancer pathogenesis and drug development.