Abstract:
This thesis addresses one essential problem of radiation therapy: setup error, organ movement along with other inevitable sources of uncertainty can significantly degrade the tumor control probability or increase the risk of complications. Based on the workflow of the treatment planning process, this thesis introduces a new approach for dose optimization and evaluation.
The core of this work is an entirely new concept for uncertainty management. The key idea is to quantitatively incorporate the fact that different uncertainty realizations entail different treatment outcomes. In essence, the merit of a treatment plan is no longer quantified by a nominal value but by a treatment outcome distribution. This distribution, reflecting all possible outcomes, is optimized. This means that every scenario in conjunction with its likelihood of being realized is taken into consideration by the optimizer.
In this thesis, a novel framework for uncertainty management is presented. In contrast to earlier publications in the field, it explicitly accounts for the fact that the inherent uncertainties in radiotherapy render all associated quantities random variables. During optimization and evaluation, all random variables are rigorously treated as such. Any model that quantifies organ movement and deformation in terms of probability distributions can be used as basis for our algorithm. Thus, it can generate dose distributions that are robust against inter- and intrafraction motion alike, effectively removing the need for indiscriminate safety margins.