Thesis: MR-LINAC are radiotherapy devices that combine a linear accelerator (LINAC) and a magnetic resonance imager (MRI), allowing an improved tumor tracking. Dosimetric issues associated with the presence of the static magnetic field have been discussed in the literature. The aim of this thesis is to develop dosimetric tools and robust protocols for use on MRI-LINACs to enhance the knowledge of the doses delivered to patients treated with these devices. The first part of the work focuses on determining output factors (OF) in small fields on this device and studying detectors' responses. No significant variation in response was observed for EBT3 films exposed to different durations in the presence of the magnetic field (0.35 T). This high-resolution 2D detector was then used to perform several series of OF measurements on an MRidian MR-LINAC, which were then compared with measurements using active detectors (ionization chambers, diodes, microdiamond) as well as with data from the treatment planning system (TPS). A good agreement was observed between the measured OF and those calculated by the TPS for field sizes larger than or equal to 2.5×2.5 cm²; an underestimation of the TPS OF was observed for smaller field sizes (0.83×0.83 cm²: 6% for films and f 4% on average for solid active detectors). After applying TRS483 correction factors, measurements with active detectors converge with those obtained with films. These discrepancies with the TPS suggest the need for a more robust adjustment of the TPS algorithm for small fields. Measurements were complemented by Monte Carlo simulations using the Geant4 code to establish correction factors in the presence of a magnetic field for active detectors. The second part focuses on the feasibility of MRI reading of TruView dosimetric gels (ModusQA) manufactured in the laboratory to evaluate dose distributions. These gels, characterized by MRI reading (measuring T2 relaxation time), exhibited a dose-response linearity up to 7 Gy, along with relatively low sensitivity compared to the literature. Significant thermal sensitivity and gel inhomogeneity between the gel surface and deeper layers were observed in non-irradiated gels, and protocols were implemented to address these issues. The feasibility of using dosimetric gels for patient quality control in MR-Linac has been demonstrated, but an enhancement of gel sensitivity is required to achieve reliable dosimetry with this protocol.