Premium
Optimization in intensity modulated radiation therapy
Author(s) -
Gaede Stewart
Publication year - 2004
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.1655708
Subject(s) - radiation treatment planning , beam (structure) , inverse , dosimetry , intensity modulation , mathematical optimization , computer science , range (aeronautics) , intensity (physics) , inverse problem , medical physics , algorithm , radiation therapy , mathematics , optics , physics , nuclear medicine , engineering , medicine , geometry , mathematical analysis , aerospace engineering , phase modulation , phase noise
Intensity modulated radiation therapy (IMRT) uses nonuniform intensity distributions to conform high dose to a tumor and low dose to surrounding sensitive structures. Because of the large number of beams (5–11) and the wide range of intensities, treatment planning is typically an inverse problem in which the intensity distributions are optimized. Three areas addressed in this thesis are plan complexity, beam directions, and dose–volume constraints. Inverse treatment planning is flexible and can deliver complex dose distributions that are sometimes not warranted. The first goal of this thesis is to demonstrate simple alternatives to inverse planning that use just enough degrees of freedom for the problem so that the solution is not overly sensitive to a slight change in dose constraints and patient geometry. With the addition of simple beam direction optimization, a suitable IMRT plan can be created while maintaining clinical practicality. The second goal of the thesis is to introduce and analyze a new algorithm which systematically analyzes and selects beam directions in the fewest number of beams possible. In IMRT, the optimization of beam directions is complicated due to the interdependence with beam intensities. Our beam direction algorithm has the capability of achieving plans that are better than standard IMRT techniques, often with a fewer number of beams. The third goal of this thesis is to propose a new formulation of the inverse treatment planning optimization problems that include dose–volume constraints which are known to destroy convexity. This is compared to a formulation that has been addressed in the literature. We solve both formulations with a new technique based on direct search optimization with a systematic search region reduction. This is compared to a standard fast simulated annealing technique. The results of using the new formulation show a direct correspondence between the minimum objective function values and the resulting dose distributions and dose–volume histograms. The research of this thesis is performed using examples of lung, prostate, and brain stem radiotherapy. We provide evidence that dose–volume based formulations of inverse treatment planning optimization for IMRT have the ability to achieve optimal plans that are clinically relevant.