Cancer is the leading cause of death worldwide. The complex and disorganized tumor microenvironment makes it very difficult to treat this disease. The most common in vitro drug screening method now is based on 2D culture models which poorly represent actual tumors. Therefore, many 3D tumor models which are more physiologically relevant have been developed to conduct in vitro drug screening and alleviate this situation. Among all these models, the microfluidic tumor model has the unique advantage of recapitulating the tumor microenvironment in a comparatively easier and representative fashion. While there are many review papers available on the related topic of microfluidic tumor models, in this review we aim to focus more on the possibility of generating "clinically actionable information" from these microfluidic systems, besides scientific insight. Our topics cover the tumor microenvironment, conventional 2D and 3D cultures, animal models, and microfluidic tumor models, emphasizing their link to anti-cancer drug discovery and personalized medicine.