The current paradigm for clinical diagnostic for the treatment of vascular disease relies exclusively on diagnostic imaging data to define the present state of the patient, empirical data to evaluate the efficacy of prior treatments for similar patients. These techniques are insufficient to predict the outcome of a given treatment for an individual patient. We here propose a new paradigm of predictive medicine where physician could use computational simulation to construct and evaluate a specific geometrical/anatomical model to predict the outcome for an individual patient. For this purpose it is necessary to develop a complex software system which combines user friendly interface, automatic solid modeling, automatic finite mesh generation, computational fluid dynamics and post-processing visualization. The flow dynamics is defined according to the incompressible Navier-Stokes equations for Newtonian and non-Newtonian fluids. Mass transport of oxygen and macromolecules is modeled by the convection diffusion equation and coupled with flow dynamics. The computer simulations are based upon finite element analysis where the new computer methods for coupling oxygen transport and fluid flow are described. The comparison results shows a good agreement between clinical observation for critical zones of flow separation, flow recirculation, low wall shear stresses which may contribute to the development of atherosclerotic diseases.