Thesis: Despite the development of radiotherapy (RT), the biological effects on healthy tissue remain poorly studied. To predict radiation-induced biological effects, radiobiologists use the Relative Biological Effectiveness (RBE) concept to compare doses between two ionizing radiations given the same biological effect. RBE is essentially based on clonogenic assay. According to several studies, this assay is insufficient to predict effects on healthy tissues after radiation exposure. Based on the various effects known after irradiation (IR), the aim of this work is therefore to acquire multiparametric biological measures to be integrated in a predictive model to foresee the biological effects of emerging radiation therapy modalities and/or protocols. For single dose, in vitro data show a deleterious effect at the highest dose rate on clonogenic survival, cell morphology, viability, cell cycle, senescence and gene expression signing cell dysfunction. These results were confirmed in vivo on a preclinical mice model of radiation-induced enteropathy. In contrary to ICRP statement, our results show an RBE of photon different from 1 and depending on the dose rate. On the other hand, different fractionated IR protocols highlight in vitro an impact of the dose rate based on a continuum of “effective biological dose” (BED). Different protocols with equivalent BED show different radio-induced response both in vitro and in vivo. This results exhibit certain limits of BED measures for clinical use to compare different fractionated IR protocols. The use of multiparametric biological measures could ultimately allow a better risk prediction related to current and future radiotherapy practices.