Exploring Tumor Heterogeneity By Using Radiomics Approach Of PET Images

Radiomics refers to the use of modern computer technology to extract digital features by non-invasive technique from medical images.Digital features are used to characterize a non-invasive technique that describes tumor imaging,anatomical structures and the tumor internal metabolic conditions of cancers.Radiomics through the extraction of image digital features could calculate a comprehensive reflection of the phenotype of the tumor structure before and after changes in various aspects of information.Through the fusion analysis with clinical indicators,finding a small number of parameters constitute the model to achieve the automatic analysis of tumor images.Initially,the doctor distinguishes the condition of clinical stages the tumors according to the patients' clinical response,changes of tumor sizes or biopsy techniques.But these methods subject to subjective factors,time factors greater impact,higher false positive rate.With the development of medical imaging technology,the application of digital analysis methods and the impact of the above factors are weakened.18F-FDG PET medical images can quantify the heterogeneity of the surrounding microenvironment between different generations of tumor cells and different tumors,depending on the difference in the biochemical and metabolic properties of the tumor target in different cells.Heterogeneity varies with the individual patient and the organ is different.Tumor heterogeneity analysis can not only reflect the biochemical and metabolic status of tumor target,the development of auxiliary radiotherapy plan,but also can be used in the clinical treatment of cancer patients,prognosis,survival and other researches.Based on the principle of radiology,the statistical parameters of 18F-FDG PET lung tumor can be used to characterize the tumor shape and heterogeneity.The statistical parameters such as statistical parameters,tumor volume and standard absorption value were analyzed by mathematical analysis in order to explore the correlation between image texture features and tumor heterogeneity.The contents of this paper are detailed as follows.Firstly,it explains the background of the topic and the significance of this thesis.It describes the concept of radiology,the elements of the research and the research situation at the present stage,and then gives a general description of the writing idea and structure of this paper.Secondly,we introduce the imaging principle of PET / CT image in detail and the benefit of the combination of two different functional images.The role of 18F-FDG tracer in PET imaging and PET image heterogeneity is briefly introduced.Thirdly,we introduce the definition of the concept of texture and description of the texture analysis of the analysis process and texture analysis is divided into four categories based on.The application of texture analysis in oncology is illustrated.Finally,the process of tumor heterogeneity analysis is described: image acquisition,image processing,radiological analysis of molecular images,correlation analysis of eigenvalues and clinical parameters.And the experimental requirements and ideas for each experiment are described.The results of tumor heterogeneity analysis are analyzed and discussed,and a new potential texture characteristic value is found for heterogeneity analysis of tumor.