ESTRO 2024 - Abstract Book

S94 ESTRO 2024 prediction of the uncertainty of the converted images, which can speed up manual inspection or can directly be incorporated into the treatment adaptation. Another crucial step is contouring of targets and organs-at-risk. Two PhD students have analyzed and developed several contouring methods, including registration and deep learning segmentation, for different imaging modalities. Again, because of their use in a time-sensitive adaptation, special emphasis was given to quantifying the uncertainties of these algorithms and how to include them in the adaptation process. When moving towards real-time adaptation, 4D imaging enters the scope. Most research done in the past used 4D CT imaging, but this has several drawbacks. Therefore, one PhD project entirely focuses on the incorporation of 4D MR imaging into the adaptation. On the one hand, it aims to incorporate 4D MRI for treatment planning in addition to the standard 4D CT. On the other hand, it works towards 4D motion modeling to improve correlation between surrogate signals and internal motion. Finally, online adaptation can benefit from taking into account the total dose delivered to a patient up to a certain time. Both inter- and intrafractional (4D) dose accumulation require deformable image registration, which introduces uncertainties. To enable its use, one PhD project is focused on quantifying these uncertainties so that an informed decision can be taken when taking into account the accumulated dose in the adaptation. Invited Speaker

3483

RAPTOR project: Intervention

Nadine Vatterodt

Aarhus University Hospital, Danish Centre for Particle Therapy, Aarhus, Denmark. Aarhus University, Department for Clinical Medicine, Aarhus, Denmark

Abstract: