Heterogeneity Studying For Tumors By Using Texture Analysis With PET Images

18F-FDG PET is a kind of functional images. It may reflect the biological heterogeneity of the tumor dynamic and objective. The patients are safe from the pathological examination and the doctors' subjective judgments through medical images. That is why we based on the PET images to study the heterogeneity of tumor. The heterogeneity of tumor has many forms. Because the different forms, the same kind tumor is difference between different people, and even more the different parts of the same people are different. These differences are very difficult to find by subjective judgment. Therefore, the study of tumor heterogeneity can not only help doctors more detailed and precise analysis of the condition of patients, but also develop a more scientific treatment planning. It gives more reasonable dose in the course of radiotherapy. At the same time, it can protect the normal organs from more damaged to achieve an optimal therapeutic effect. The heterogeneity of primary and lymphoma are analyzed by texture analysis. The tumor heterogeneity is characterized by texture feature value.Firstly, this paper introduced the background of texture analysis of tumor heterogeneity and significance of the research. The research statue and the difficult issues were summarized in these parts. Then, the main research content and the structure of the paper were described.Secondly, this part introduces what is texture. The types of texture analysis were pointed and how to use them respectively. The common methods were listed, such as histogram, gray level co-occurrence matrix, Fourier transform and wavelet transform. The definition and meaning of texture features were illustrated in the experiment.Thirdly, the PET / CT equipment was descripted very detailed. How to obtain a PET image, and the images were pre-processing. Then, the ROI of tumor was segmented automatically by MATLAB software. Quantifying tumor heterogeneity of primary and lymphoma tumors used texture analysis of PET images.Finally, the primary tumors of lung and lymphoma tumors were analyzed in deeply. The tumor heterogeneity was represented by extracted texture features. The texture features of the primary and the lymphoma tumors were compared, and the volumes of the primary and the lymphoma tumors were analyzed before radiation therapy. The correlations and Pearson correlation were calculated between the primary and the lymphoma tumors. At the same time, the tumor heterogeneity of the primary and the lymphoma tumors were compared after radiation therapy. At last, the conclusion was given about the heterogeneity of the primary and the lymphoma tumors. In the end, in addition to the correlation, the other characteristics have proved the heterogeneity of the primary tumor above the heterogeneity of the lymphoma tumor before treatment; the size relationships of characteristics' valus between pre-treatment tumors and pro-treatment tumors are no changed.