Computer optimization of class solutions designed on a beam segmentation basis.

Abstract

A method for analytically solving the optimization of beam weighting in radiotherapy treatments using beam segmentation is presented. A technique has been elaborated permitting the optimization of a flexible objective function which is defined by considering 'importance factors' for each anatomical region and 'allowed deviations' from the prescribed dose. As any change in these importance factors may lead to very different solutions, a statistical tool has been developed which varies the objective function automatically and iteratively to get the best possible results compatible with the chosen class solution. In addition, the spatial symmetry found in many anatomical sites is taken advantage of. Furthermore, freeware code has been written to run this optimization approach. The guidelines to design the beam segmentation used in our institution using organ avoidance criteria, and hence the suitable class solution for different anatomical sites, are given. A treatment-planning study for three anatomical sites is presented and, for two of them, the results obtained with both the suggested and the classical inverse approach are presented. The work presented might be used for beam weighting optimization in any radiotherapy treatment and furthermore, the suggested procedure may successfully confront the intensity-modulated radiation therapy (IMRT) problem and the obtained dose distributions fit the clinical constraints for all anatomical sites studied. When comparing with the classical inverse approach, both results are comparable in terms of dose distribution, but the suggested technique reduces the integral dose as the total number of monitor units is lower. The developed code performs the optimization with a very low time cost and in addition, this process can be carried out with a conventional treatment-planning system with no need of dedicated IMRT software.