Résumé:
Vectorized internal radiotherapy is based on the administration of radiopharmaceuticals to patients
by the oral or intravenous route. Unlike external radiotherapy, this modality has no standard dosimetric methods that are easily applicable in clinical practice. However, the polymer gels have proven to be good dosimeters with a lot of potential, it allows a three-dimensional dose reading in particular. Our work focuses on the study and modeling of the response of MAGIC gel (Methacrylic and Ascorbic Acid in Gelatine Initiated by Copper) which is a last generation normoxic gel. The aim is to implement mathematical tools that facilitate the use of MAGIC
polymer gel as a dosimeter in internalized radiotherapy. In this study, and for practical reasons,
we used Technetium-99m (radioelement dedicated to nuclear imaging) as an internal source homogeneously distributed in the volume of the dosimeter. The radiation dose from the different activities of the radionuclide was measured by MRI by analyzing the radiation-induced change in the gels. We have developed, based on compartmental analysis, a mathematical model to describe the response of the MAGIC gel. Equations governing this model allowed us to establish the relationships that link the MRI signal to the proton densities that we classified into three categories. The intensity of this signal depends on the relaxation time of the proton considered, which varies according to the environmental situation of the proton. To establish these equations describing the variations of these densities, we classified the different protons contributing to the formation of the magnetic resonance signal, and thus the value of R2 of the gel, in several categories, each
forming a compartment. The MRI images of the MAGIC gel samples in different experimental conditions validated the theoretical approach adopted and confirmed all the hypotheses emitted during the construction of the mathematical model. The obtained results open the way for many research perspectives. This work should be able to be integrated with the numerous dosimetry simulation software platforms using Monté Carlo methods, such as GEANT4 or PENELOPE.
