基于PET图像放射组学特征的肿瘤异质性研究

发布时间：2018-03-10 00:36

本文选题：放射组学　切入点：纹理特征　出处：《山东师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：放射组学是指借助现代计算机技术,从医学图像中提取图像数字特征用于表征描述肿瘤的显像、解剖结构和肿瘤内部代谢情况的一种非侵入式的技术。放射组学通过对图像数字特征的提取计算,全面的反映肿瘤的表型结构、时间前后变化等各个方面的信息。通过与临床指标的融合分析,找出少量参数构成的模型实现肿瘤图像的自动化分析。最初,临床医生根据病人的临床反应表现、肿瘤大小变化或者活检技术对肿瘤进行分期或者疗效分析。但是这些方式受主观因素、时间因素影响较大,误判率较高。随着医学影像技术发展、数字化分析方法的应用,上述限制因素的影响得到削弱。18F-FDG PET医学影像能够根据示踪剂在不同细胞中含量的不同反映肿瘤靶区的生化代谢的差异,量化不同子代肿瘤细胞之间和不同肿瘤之间周围微环境的异质性,这种异质性随着患者个体的不同、器官的不同而各有其不同。肿瘤异质性分析不仅能反映肿瘤靶区的生化代谢情况、辅助精准放疗计划的制定,还能够用于临床癌症病人的治疗疗效、预后、生存期等情况研究中。本文依据放射组学的原理计算能表征肿瘤形状特征、异质性特征的18F-FDG PET肺部肿瘤图像统计学参数,通过对统计学参数和肿瘤体积、标准吸收值等参数进行数学分析研究医学影像纹理特征与肿瘤异质性间的关联。本文在研究内容上:首先,说明了本文的选题背景以及本论文的选题意义,描述了放射组学的概念、研究内容包含的要素及现阶段研究情况,然后对本文的写作构思以及结构做了整体说明。其次,详细介绍了PET/CT图像的成像原理,两种不同功能图像间结合使用的优点。对18F-FDG示踪剂在PET图像显像中的作用、PET图像异质性参数简要介绍。再次,说明了纹理的概念定义并说明了纹理分析的分析过程及纹理分析分为四大类的依据并介绍了纹理分析在肿瘤学应用。最后,阐述肿瘤异质性分析的过程:图像的获取、图像的处理,分子图像的放射组学分析,特征值与临床参数的相关性分析。并对每一步实验的实验标准要求和思路做了说明。对肿瘤异质性分析中的结果进行分析和讨论,得到了未发现的新的潜在的纹理特征值用于肿瘤的异质性分析。
[Abstract]:Radiation group refers to the use of modern computer technology, digital image feature extraction from medical images used for imaging the description of characteristics of tumor, anatomical structure and internal metabolism of tumor in a noninvasive technique. Radiation group is extracted based on digital image feature calculation, reflect the tumor phenotype structure, information changes etc. all aspects of the time before and after. Through the fusion and clinical index analysis, find out a few parameters model which realizes the automation of tumor image analysis. First, according to the clinical manifestations of patients with clinicians, changes in tumor size or biopsy of the tumor staging or analysis effect. But these methods are affected by subjective factors, time factors, miscarriage of justice the rate is high. With the development of medical imaging technology, the application of digital analysis method, the influence of limiting factors to weaken.18F-FDG P ET medical image according to the difference of biochemical metabolism in different reflect tumor target tracer content in different cells, heterogeneity between tumor cells and quantify the different progeny between different tumor microenvironment around, this heterogeneity with individual difference, different organs and different heterogeneity of tumor not only. Can reflect the biochemical metabolism of the tumor target, making auxiliary precise radiotherapy, but also can be used to treat cancer patients clinical curative effect, prognosis, survival and other research method. According to the principle of radiation proteomics can characterize tumor shape features, the heterogeneity of 18F-FDG PET lung tumor image statistical parameters, by volume the statistical parameters and tumor, etc. the standard uptake value of the parameters of medical image texture features and correlation between tumor heterogeneity based on the study of mathematical analysis. Content: firstly, this part introduces the background and significance of the topic of this paper, describes the concept of radiation group, the contents of research include the elements and the present research situation, and then the writing conception and structure. Secondly, introduces the imaging principle of PET/CT image, two different the function of the image using a combination of advantages. The role of 18F-FDG PET imaging in the tracer in the image, PET image heterogeneity parameters are briefly introduced. Thirdly, explains the definition of texture and illustrates the physical analysis process and texture analysis is divided into four categories based on texture analysis and introduced finally applied in oncology. This process, analysis of tumor heterogeneity: image acquisition, image processing, image analysis of the molecular radiation group, correlation analysis and clinical parameters characteristics. And each step of experiment The standard requirements and ideas were explained. The results of tumor heterogeneity analysis were analyzed and discussed, and new potential texture feature values were not found, which were used for tumor heterogeneity analysis.

【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R817;TP391.41

【参考文献】