One way of gaining insight into what can be observed in medical images is through physiological modeling. For instance, anatomical and functional modifications occur in the organ during the appearance and the growth of a tumor. Some of these changes concern the vascularization. We propose a computational model of tumor-affected renal circulation that represents the local heterogeneity of different parts of the kidney (cortex, medulla). We present a simulation of vascular modifications related to vessel structure, geometry, density, and blood flow in case of renal cell carcinoma. We also use our model to simulate computed tomography scans of a kidney affected by the renal cell carcinoma, at two acquisition times after injection of a contrast product. This framework, based on a physiological model of the organ and physical model of medical image acquisition, offers an opportunity to help radiologists in their diagnostic tasks. This includes the possibility of linking image descriptors with physiological perturbations and markers of pathological processes.