The proposed project has the overall goal of developing techniques to aid in the assessment and optimization of computer-generated plans for external photon beam radiation therapy of cancer-using mathematical models which attempt to quantify the biological response of tissues to radiation. Biological Models will be developed: (a) for estimating tumor control probabilities (TCP) under conditions of non-uniform irradiation of heterogeneous tumors: (b) for estimating normal tissue complication probabilities (NTCP) under conditions of non-uniform irradiation of a variety of tissues and organs; (c) a technique will be developed for combining TCP's and NTCP's into a single number (objective function) which can be used o pick the preferred of two plans; (d) techniques will be developed to speed up the assessment and optimization processes. Tools will be developed for assessing radiation therapy plans which include: (a) a plan summary report, incorporating both numerical and graphical representations of one or more plans on an organ by organ basis, to facilitate the assessment of a plan and the intercomparison of two or more plans; (b) a three-dimensional graphical representation of a plan which will quickly indicate the regions of suspected over-and under-dosage. This representation will be a beam's-eye view of the 3D spatial distribution of a quantity, termed regret, which will be based on the biological models we develop. This view is termed a """"""""cloud of regret"""""""". Techniques will be developed for optimization and improvement of radiation therapy plans. This aspect of the proposed research is directed towards providing a planner with tools for designing an optimum plan of treatment, or improving an existing plan, making use of biological models of TCP and NTCP. The optimization will be limited to the following parameters: beam weight; choice of wedge(s); choice of beam energy; and adjustment of block position near a critical structure. An objective function will be developed, built from biological models of TCP and NTCP, which can be used to rank rival plans; available mathematical optimization techniques ill be investigated, and one selected with which to search for that plan which maximizes our objective unction; and techniques will be developed to suggest to the planner profitable directions in which to improve an existing plan. The assessment, improvement and optimization techniques developed will be applied to clinical cases to judge their efficacy. It is expected that the results of this research will allow better use of CT and MRI information about the patient, and will lead to increased tumor control and reduced treatment complications.
Showing the most recent 10 out of 28 publications
