Abstract Introduction: Although cancer is a heterogeneous disease, current chemotherapy is based on the use of anti-cancer agents that broadly affect cell proliferation. As a result of this non-specific tumor therapy there is often high toxicity and therapeutic resistance, which increases the risk of disease progression. There is thus a critical need to find novel anti-cancer compounds, however, as a result of preliminary screens often performed in 2D systems, and the use of mouse models for secondary screens, many drugs fail during full clinical trials. Replicating a complex, 3D microenvironment incorporating human cells is therefore likely to improve drug screening efficiency. Methods: Here we have developed a system for high throughput screening of drug efficacy. Our system features 3D tumor tissues incorporating human cells co-cultured with stromal cells and connected by human microvessels in a naturally occurring 3D matrix in a PDMS microdevice. Transduced colorectal cancer cells (CRC) such as HCT116, SW480 and SW620s that express green fluorescent protein (GFP) constitutively are introduced into the device in co-culture with human endothelial colony-forming cell-derived EC (ECFC-EC) and fibroblasts. ECFC-EC or fibroblasts that express longer wavelength fluorescent proteins are also used to track both tumor cells and vasculature/fibroblast growth. The tri-culture is resuspended in fibrinogen (10 mg/ml) and mixed with thrombin before injecting into the microchambers. Cells are fed with endothelial growth media-2 and drugs are delivered through the microfluidic channels using a hydrostatic pressure gradient. Results: A complete vascular network was formed in the chambers by 7-10 days and this carried medium into the tissue. CRC showed continuous growth, often forming spheroids. After 7-10 days of culture, cells were exposed to various FDA-approved drugs at concentrations reported for plasma levels in patients. Tumor growing in the device responded differently to most drugs compared to cells growing in 2D. We identified cytotoxic and cytostatic drug actions in the device with IC50s close to the in vivo pharmacologic plasma concentration, but higher than those obtained in parallel 2D experiments. Importantly, we also identified drugs that were active in our 3D-device but not in 2D cultures. Conclusions: Preliminary results using our 3D microphysiological system, in which we have created a more complex microenvironment than used in 2D cultures, support an improved strategy for drug validation. In addition, future experiments will include the use of non-invasive optical imaging techniques to better understand the human 3D tumor microenvironment, including the study of different tumor behaviors, tumor metabolism and pathological angiogenesis. This proof-of-concept study validates the use of 3D microphysiological systems in place of more simplistic 2D systems. Citation Format: Agua Sobrino, Dúc Phan, Steven C George, Christopher C.W Hughes. A 3D tumor tissue microphysiological system for realistic tumor microenvironment mimicry and therapeutic modeling. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 5500. doi:10.1158/1538-7445.AM2015-5500