MRI Guided Radiotherapy and Radiobiological Data: the ISRAR Database (Irm Sequences for Radiobiological Adaptative Radiotherapy)

The MRI linac Unity is a major technological evolution in radiotherapy combining a linear accelerator with a 1.5T MRI (radiological quality). It allows to target the target volume more precisely and to adapt the daily dose distribution according to variations in the position and volume of the tumor, critical organs and the tumor response. In many studies conducted in radiology, the analysis of specific MRI sequences, particularly in radiomics, aims to characterize tumors and their sensitivity to treatment. Initial data show that in radiotherapy, it would eventually be possible to characterize the radiosensitivity of healthy and tumorous tissues. With linac 1.5T MRI, the performance of selected MRI sequences, at each session, could make it possible to identify different levels of radiosensitivity within the tumour. The reproduction of these sequences on a daily basis could make it possible to follow the variations in radiosensitivity during the treatment. The final objectives would be: 1- to adapt the doses of radiotherapy to each session with a modulation of the dose according to the daily level of intra-tumor radiosensitivity, 2- to develop Artificial Intelligence (AI) tools allowing an analysis sequences and the generation of 3D maps of intra-tumor radiosensitivity, fast and suitable for carrying out a radiotherapy session. A first work carried out in collaboration with the CREATIS lab of the University Claude Bernard Lyon 1 (UCBL1) made it possible to generate maps of tissue oxygenation from sequences produced on the MRI linac Unity of the Hospices Civils de Lyon (T2\* , IVIM, Carto T2 Multi Echo-Gradient). Hypoxia is known to be the first factor of tumor resistance to irradiation. A research program is structured in collaboration with UCBL1 in order to develop radiobiological adaptive radiotherapy approaches, based on 3D maps of intra-tumoral hypoxia and their variation during treatment. Several tumor locations were selected because of the preponderant place of MRI in tumor characterization: prostate, cervix, kidney, ENT and glioblastoma. Hypoxia is not the only factor of radioresistance. Changes in the microenvironment could also impact the sensitivity of tumor cells. The program will therefore also aim to optimize the maps initially based on hypoxia, by identifying other relevant factors to be taken into account to define intra-tumor sensitivity.