IMRT beam angle optimization using DDS with a cross-validation approach for configuration selection

Abstract

Radiation incidences (angles) that are used in Intensity Modulated Radiation Therapy (IMRT) treatments have a significant influence in the treatment clinical outcome. In clinical practice, the angles are usually chosen after a lengthy trial and error procedure that is significantly dependent on the planner's experience and time availability. The use of optimization models and algorithms can be an important contribution to the treatment planning, improving the quality of the solution reached and decreasing the time spent on the process. This paper describes a Dynamically Dimensioned Search (DDS) approach for IMRT beam angle optimization. Several different sets of parameters and search options were analyzed. Computational tests show that the final outcome is strongly influenced by these choices. This motivated the use of a cross-validation based procedure for choosing the algorithm's configuration, considering a set of ten retrospective treated cases of head-and-neck tumors at the Portuguese Institute of Oncology of Coimbra. © 2014 Springer International Publishing.