Research On Objective Function Based On EUD And Extraction Of Convolution Kernel By MC Method

Together with surgery and chemotherapy, radiotherapy plays an important role in oncology, both in the definitive and palliative aspects of treatment. It treats with about 60%-70% patients suffering from cancer and cures 40% of them. It is a historic advancement that the classical three-dimensional (3D) conformal radiotherapy (3DCRT) evolved into the intensity modulated radiotherapy (IMRT). IMRT is an inverse treatment planning, that is, the planner defines certain constrains for the desired dose distribution, and the optimization process the computer trying to find the treatment setup that matches the constraints as closely as possible. However, the advantages of IMRT have not been fully utilized yet, due to the complicated clinical conditions.This paper mainly concerns two critical issues of IMRT: the objective function and the convolution kernels based on Monte Carlo method. Objective function is an important designation to optimize and evaluate the treatment plan. It doesn't only connect the practical dose-output and expected goals, but indicates the extent of the treatment plan to be accepted. Physical function is the most widely used nowadays, but it fails to imply the influence of the dose inhomogeneity to the treatment planning while the function based on equivalent uniform dose(EUD) works better in this aspect.This paper proposes the application of the objective function based on EUD in IMRT. EUD is defined as "that dose which, when distributed uniformly across the target volume, causes the survival of the same number of colognes as the delivered non-uniform dose distribution". The background of EUD, the mathematical analysis of EUD is given in chapter3. Both EUD-based and dose-based objective functions are applied to optimize the same IMRT plans. It is found that EUD-based criteria provides better target coverage and is capable of improving the spring of critical structure beyond the specified requirements. Equivalent uniform dose based objective function needs only a small number of parameters and allows exploration of a much larger universe of solutions. It has nice derivability and convexity. It also can be a surrogate of biologic index such as tumor control probability and normal tissue complication probability. It dose have a bright future in IMRT.Dose calculation is the kernel and the soul of the treatment planning system. Accurate dose calculation for irregular field has always been an essential topic in IMRT. Convolution is nowadays a popular dose calculation method by which dose is calculated by convolving the released photon energy with an energy deposition kernel.Monte Carlo is the most perfect method in theory to implement the energy deposition kernel. Chapter 4 and Chapter 5 illustrate the Monte Carlo method, the EGS Monte Carlo Code and finally apply the EGS codes to achieve an energy deposition kernel. The EGS Monte Carlo energy deposition kernel and the pencil beam kernel extracted by deconvolution method are used to compute the dose distribution for same plans. It is found that there is remarkable dose inhomogeneity in target region when employing the EGS Monte Carlo kernel.The end part of the paper gives an overview and puts forward the issues to do in the future.