Research On Key Techniques Of Nasopharyngeal Cancer Brachytherapy Treatment Planning And Simulation System

The development of present-day science and technology was more and more characterized by the crossover of multi-disciplinarity.Virtual reality and simulation as a new area of rearch encompasses a wide spectrum of techniques including medicine,computer graphic technology,computer vision technology,mathematic analyzing and so on.Going with the development of computer technology and studying of medical image,cancer treatment technology go through a great change.Just as americal physical scientist J.Purdy described:radiation therapy has entrance a new era,the era of 3DRT(Three-Dimensional Radiation Therapy).Nasopharyngeal Cancer is one of the most common form of human cancer and the leading cause of human cancer deaths.Because of its characteristic,radiation therapy has become its major treatment option and Brachytherapy:which is a type of internal radiation thrapy given by placing radioactive sources directly into the cancerous tissue,is often used.Because there are many important tissues such as vas,nerve,lie around the nasopharynx,the dosimetric treatment planning must be designed carefully to achieves a desired dose distribution to the target,while sparing the critical structures.The image based treatment planning and treatment simulation for brachytherapy which can improve treatment quality,was studied by many researchers in recent years.The results of these research are of great value to treatment teaching,doctor training,clinical diagnosis and therapy.Firstly,we introduced the background of this area,then the key technology of image segmentation of soft tissues,three-dimensional modeling,calculation and optimization of HDR radiation dose,needle insertion simulation based on mass-springs model was researched.The main contents and contributions of this dissertation are as follows:(1) An algorithm based on the active shape model is proposed for the segmentation of medical image series.The edge of soft tissue in medical images usually are not very clearly,so it's difficult to segment them only by the computer.But a tissue's shape transform sequentially in medical image series,we asked doctor to select some images from image series and segment the tissue in these images,then we used the results of the segmentation as training maps,created a active shape model by PCA Algorithm.Next we used this model to segemnt other images.The experimental results show the accuracy of this method.(2) A parametric shape modeling using deformable superellipses for segmentation of medical image was proposed.Superellipses can represent the smooth and symmetrical contour perfectly, Most of anatomical structures having such characters can be approached by a superellipse.Prior shape information been got from a statistical modal analysis of a training set.This information was used to restrict the transformation of the surperellipse parameters.We used Hybrid Genetic Optimization Algorithm to find the optimal superellipse parameters,and snake algorithm embed with these shape information to segment the image.With shape guidance,this algorithm is less sensitive to initial contour placement and more robust even in the presence of large boundary gaps.The experiments show the efficiency of this method.(3) A new hybrid genetic algorithm based on a continuous-parameter GA with linear constraints (GENOCOP) and POWELL algorithm was proposed.The genetic algorithm has been widely used in many fields as an easy robust global search and optimization method.To optimizing the dose distribution,a new hybrid genetic algorithm based on a continuous-parameter GA with linear constraints(GENOCOP) and POWELL algorithm was proposed.The POWELL algorithm was modified to support the linear constraints and was embed in hybrid genetic algorithm as a local search operation,and the operators of GENOCOP were amended to improve its ability of global search.The simulation results show that compared to traditional GENOCOP methods,this hybrid genetic algorithm has obvious potential on many respects,such as convergence speed,solution accuracy,ability of global optimization.(4) An algorithm to optimize the dwell time of the radioactive source was proposed. Brachytherapy is the treatment of cancer by means of radioactive sources that are placed at short distances from the target cells.The computer-based HDR planning system of brachytherapy is becoming more and more widely used.To get the solution that best satisfies the clinical dose constraints,we proposed an algorithm to optimize the dwell time of the radioactive source.Firstly we considered the dwell time as a continuous function of the dwell position,computed the dose values of referenced dose points.The minimal difference between these dose values and demanded dose values was used as the objective of an optimization algorithm to get the optimized function parameters.The function curve can be separated to several sects.At last the optimized curve was dispersed using compound trapezoid formula to get the dwell positions and dwell times.This method not only avoids the emergence of negative dwell time,but also reduces the dwell time gradient.And during the process of dispersing the curve,we can get different dwell times and dwell positions.This makes the brachytherapy treatment plan more flexible.(5) Simulation of needle insertion based on mass-spring model was developed to facilitate surgeon training and planning for brachytherapy.Inserting a needle into soft tissues causes the tissues to displace and deform:ignoring these effects during seed implantation leads to imprecise dose distribution.We construct a 2-D dynamic mass-spring model using 400 springs and simulated the needle insertion.The deformation of the tissue and the force can be calculated.Simulation achieved 100 flames per second on a 2Ghz Pentium PC.