Optimisation of Radiotherapy: Inverse Problems and Artificial Intelligence for Quality Assurance, Treatment Planning and Dosimetry

Xavier Franceries

Portal imaging (EPID) has been used for several years in external radiotherapy. Even if some solutions exist for a quasi-homogeneous environment, they are difficult to implement and cannot give satisfactory results in heterogeneous environments. Inverse problems coupled with neural networks will increase and simplify the use of EPID. These new possibilities also allow for improved treatment planning and possible adaptive radiotherapy.

This project is structured around 2 research axes:

• Improving quality assurance and in-vivo dosimetry through the use of portal imaging.
• Implementation and development of new methods for adaptive radiotherapy.

Partners :

• LAAS
• Dosisoft