The objective of the thesis was to explore the value of different noninvasive imaging technologies to assess tumor metabolic profiles in vivo and their pharmacological modulations. Adaptive behaviors supporting a deregulated metabolism in cancer represent major challenges for therapies. Therefore, a better understanding of these processes associated with the characterization of metabolic patterns in tumors would lead to the development of more efficient anticancer strategies. In parallel to in vitro characterizations, advanced imaging techniques including EPR oximetry, 13C-hyperpolarized NMR, 17O MRS and 18F-FDG PET, were used to identify the metabolic signature in well-established tumor models in vivo. Our study reveals major discordances between in vitro and in vivo metabolic profiles, highlighting the limitation of in vitro studies to truly reflect the complex tumor behavior. Using imaging modalities, we confirmed the key role of local microenvironment to shape tumor features.