With cost effective distributed memory computer systems reaching high performances, it may become feasible in the near future to provide routinely reliable blood flow simulations during angiographic procedures to enhance standard medical imaging techniques. The long term goal of our work is to produce a fast parallel image base Navier-Stokes flow solver for these kinds of clinical procedures. To achieve such performance, our method relies on the proper combination of three techniques that are the L_2 penalty method to deal with complex geometry, a level set method to extract the geometry of the arteries, and a fast scalable domain decomposition solver for the parallel processing of the discrete equations. We developed a web application with a database back-end to store both the original medical images and the results of these simulations. This will assist the endovascular surgeon by providing quickly blood flow indicators such as pressure or shear stresses at the wall of the artery, in addition to existing image analysis technique aimed at improving therapy.