Evaluation of probabilistic radiotherapy planning in prostate cancer

External beam radiation therapy (RT) is one of the leading modalities for treating cancer. Current clinical RT planning guidelines were established in 1993 and account for uncertainties using target volume expansion by geometric margins. This approach is simple to use but has several limitations and oversimplifies complex processes that are not linear in nature.
In this work, we evaluate the performance of a probabilistic treatment planning approach using voxel-based tumor likelihood maps and robust optimization (RO), where the objective function is calculated with different scenarios of uncertainty realizations, which more realistically represents treatment planning and delivery than safety margins. Using in-house developed Tomotherapy treatment planning software we calculated and evaluated treatment plans for eleven prostate cancer patients with MRI-derived tumor likelihood maps. Plans were created using classical approach and RO with different scenario selection and optimization functions. The RO approach was able to create comparable plans to existing classical planning, as well as create dose plans based on tumor likelihood maps and plans with heterogeneous dose prescription.