Description

Proton arc therapy (PAT) is a novel modality to treat cancer, involving continuous rotation of the irradiation device (gantry) around the patient while delivering a proton beam to the tumor. Compared to conventional fixed-angle proton beam delivery, PAT paves a promising treatment avenue, with numerous possible irradi-ation directions, to improve plan quality and streamline the treatment workflow. Nevertheless, this increased degree of freedom also becomes a burden for its clin-ical implementation, in particular in the treatment planning optimization step. The cornucopia of irradiation directions and energy layer candidates increases the complexity of the associated treatment planning optimization problem. It requires substantial computational power as well as the development of new algorithms and objectives to yield optimal treatment plans that are both high in quality and efficient to deliver. To this purpose, one must solve the energy layer selection problem, which constitutes the major challenge of PAT treatment planning op-timization. This thesis is dedicated to addressing this challenge and providing potential solutions to it. The chapters unfold as a journey of deeper and deeper exploration of PAT, maturation of the problem statement, and refinement of the proposed solutions. Initially, we tackled the problem with all available degrees of freedom, but practical constraints arose. The computational demand of im-plemented solvers proved to be too prohibitive. Consequently, we transitioned to a second approach that selectively incorporates a subset of degrees of free-dom, aligning with technical constraints. Ultimately, this thesis concludes with an in-depth theoretical exploration of the PAT problem, confirming its inherent com-plexity. This work contributes to advancing the field of PAT treatment planning optimization, addressing the delicate balance between plan quality and clinical feasibility.