Student Projects

Radiotherapy aims to kill tumor cells while minimizing the risk of complications to healthy tissues. The FLASH effect has been shown to increase the normal tissue sparing effect while achieving similar tumor control. To achieve this effect, however, an ultra-fast delivery is required. In proton therapy, this can be achieved by passively modulating the monoenergetic scanned beam with Energy Modulators (EM). EM currently in use presents a “hedgehog” shape and, due to its tiny structures, is fragile, requiring expensive industrial 3D printers that necessitate extra time, costs, and outsourcing. Therefore, a new EM concept utilizing simple FDM printers and generating “cube-like” structures to enhance mechanical robustness has been investigated, yielding promising results, as shown in Figure 1. In previous work, we developed an optimization framework for EM optimization, which was initially tested with geometrical shapes. However, we are still a step behind clinical implementation. Indeed, an efficient optimization for complex realistic targets is still missing, as well as a dosimetric evaluation of the EM performance.

The goal of the project is the development of an EM modulator optimization framework that covers clinically realistic application scenarios. A realistic scenario assumes input data in the form of a real CT with complex targets that incorporate density information. The framework should include EM design space modeling allowing for expedient generation and optimization of EM designs satisfying the EM production constraints (size, infill limits) and dosimetric requirements such as homogeneity of the target, dose to organ at risk, as well as an easy control of the design space parameters and variables in the pencil beam plan. Deliverables of this project include: - Development and implementation of an efficient EM optimization environment using established optimization approaches. - Determining the optimized input field by varying the number of protons per pencil beam for specific targets. - Generation of the optimized designs for clinical targets adhering to design and dosimetric requirements. - Experimental validation at the Paul Scherrer Institute of the optimized designs fabricated using additive manufacturing (in collaboration with the University of Applied Sciences Northwestern Switzerland). - Bonus: Development of an easy-to-use user interface to upload the CT and visualize target contours, decide the gantry angles, optimize the EM, visualize the dose distribution and target metrics, and output the Excel file for the 3D printing. Desirable student qualities: - Solid foundations in structural optimization and mathematical modeling. - Good Python and Ubuntu proficiency.

Isabella Colizzi isabella.colizzi@psi.ch Tino Stankovic tinos@ethz.ch