Radiation Therapy of cancer patients can be formulated as an optimization problem: choose the treatment parameters in such a way that one will achieve the best possible outcome for the patient. Unfortunately, reliable models to predict the outcome are not (yet) available, hence one uses the spatial radiation dose distribution as a surrogate of outcome. Since the invention of intensity modulated radiation therapy (IMRT) for the treatment of complex tumor target volumes using beams of non-uniform intensity, the optimization problem in radiation therapy is very similar to the problem of image reconstruction from projections, as in computed tomography.

In this talk, after some basic introductions into the field, I will focus on three recent research developments in radiotherapy optimization. First, I will discuss multi-criteria optimization approaches. These allow the clinicians to better understand the unavoidable tradeoffs in the treatment, and to interactively "navigate" until the most suitable plan has been found. Secondly, I will present the application of robust optimization concepts in radiation therapy. Conventionally, margins are being used to take uncertainties into account, but robust optimization can yield better solutions. Finally, I will briefly discuss optimization of a recent treatment modality, proton therapy, as opposed to standard radiation therapy with photons. Proton therapy opens up unique treatment opportunities through its ability to stop the beam in the tumor target volume, but this benefit also presents some specific challenges, which have to be taken into account in the optimization.