Study On Fusion Of Glioma Image Segmentation And Radiation Dose Field

The treatment of malignant tumor mainly includes operation,chemotherapy and radionuclide therapy.According to statistics,about 70% of cancer patients need radiotherapy in the process of treating cancer,and about 40% of the pain can be cured by radiotherapy.Radiotherapy plays an increasingly prominent role in the treatment of cancer and becomes one of the major methods to treat malignant tumors.How to combine image processing technology with radionuclide therapy to provide effective and intuitive data analysis for medical treatment becomes the focus of research.Aiming at the problem of radiation treatment for glioma,this paper will study the medical image segmentation and brachytherapy simulation of glioma.In the segmentation of glioma,the basic U-Net network in deep learning does not achieve the ideal effect of image segmentation.In this paper,the U-Net network is improved by adopting the residual network structure in the encoder part and adding the dense connection in the jump connection part to improve the accuracy of tumor image segmentation.Through processing the Brats2019 data sets,the two-dimensional slice data sets are obtained.Based on the data sets,different network models are trained and compared,and the effectiveness of the improved U-Net is proved by experiments.The capsule source model of radiotherapy is established and simulated at the drug end of the integrated capsule structure of radiotherapy and chemotherapy.In this paper,the complex geometry model is equivalent to a ring,and the Monte Carlo method with random response is used to simulate the particle transport of γ rays produced by radioactive drugs.The radiotherapy capsule model is built based on MCNP5 software platform.Meanwhile,the dose field and dose rate distribution of the radiotherapy capsule with the integrated chemotherapy capsule structure are obtained through simulation,and the dose field is visualized in two and three dimensions.Two-dimensional images are processed by Matlab functions for threedimensional reconstruction.The edges of the images are taken out,and then superimposed according to the layers,and the three-dimensional envelope surface is taken,so as to carry out three-dimensional reconstruction.In Matlab software,the integration with dose field distribution is completed by corresponding spatial coordinates.Through the segmentation algorithm based on U-Net,the two-dimensional MRI images of glioma are segmtured,and the integrated capsule radiotherapy end of glioma is simulated,and then the three-dimensional reconstruction model of glioma is fused with the three-dimensional radiotherapy dose field.The fused images have both fine anatomical structure and rich physiological and biochemical function information,which provide effective information for determining and searching the precise location of tumor lesions,positioning of integrated capsule implantation with radiotherapy and making radiotherapy plan.