The study of magnetic resonance (MR) markers over the past decade has provided evidence that the tumor microenvironnement and hemodynamics play a major role in determining tumor response to therapy. The aim of the present work is to predict and monitor the efficacy of co-treatments to radio- and chemotherapy by noninvasive MR imaging. Ten different co-treatments were involved in this retrospective analysis of our previously published data, including NO-mediated co-treatments (insulin and isosorbide dinitrate), anti-inflammatory drugs (hydrocortisone, NS-398), anti-angiogenic agents (thalidomide, SU5416 and ZD6474), a vasoactive agent (xanthinol nicotinate), botulinum toxin and carbogen breathing. Dynamic contrast enhanced (DCE) MRI, intrinsic susceptibility-weighted (BOLD) MRI and electronic paramagnetic resonance (EPR) oximetry all reflect tumor microenvironment hemodynamic variables that are known to influence tumor response. Eight MR-derived parameters (markers) were tested for their ability to predict therapeutic outcome (factor of increase in regrowth delay) in experimental tumor models (TLT and FSaII) after radiation therapy and/or chemotherapy with cyclophosphamide, namely tumor pO₂ and O₂ consumption rate (using EPR oximetry); tumor blood flow and permeability, i.e. V(p), K(trans), K(ep) and percentage of perfused vessels (using DCE-MRI); and BOLD signal intensity and R₂* (using functional MRI). This multi-modal comparison of co-treatment efficacy points out the limitations of each MR marker and identifies in vivo pO₂ as a relevant endpoint for radiation therapy. DCE parameters (V(p) and K(ep)) were identified as a relevant endpoints for cyclophosphamide chemotherapy in our tumor models. This study helps qualify relevant imaging endpoints in the preclinical setting of cancer therapy.