Jelte de Jong

Proton therapy is a relatively novel type of cancer treatment that is especially useful for the irradiation of tumors that lie in the vicinity of organs at risk. Protons deposit their energy mostly at the end of their range in a so-called Bragg peak, therefore their dose can be delivered more accurately in the tumor region than with traditional photon radiotherapy. This also means however that proton therapy is more sensitive to uncertainties during the treatment compared to conventional radiotherapy. Deviations in - amongst others - patient alignment, proton range, anatomical changes between treatment days or intra-fractional motion during irradiation can lead to uncertainties in the final dose, which in turn can degrade plan quality. 

Robust optimization takes these uncertainties into account by optimizing the treatment plan for multiple different error scenarios simultaneously. The resulting plans are robust for the considered scenarios, meaning that the desired tumor coverage is achieved in all scenarios. However, robust optimization neither takes into account the full range of potential error scenarios that can occur, nor their probability of occurrence. Consequently, robust plans can only be realized by compromising healthy tissues to higher doses than necessary, meaning more tissue is potentially damaged. 

In my research I will develop methods for probabilistic optimization. In this approach, the probability for a certain error scenario to occur is also considered. As a result, the high dose margins around the tumor can be decreased, without sacrificing robustness. In the end, this means that more tissue will be spared.