项目名称: 动态PET的定量分析算法和血液函数自动提取算法的研究
项目编号: No.81071218
项目类型: 面上项目
立项/批准年度: 2011
项目学科: 武器工业
项目作者: 陈喆
作者单位: 中国科学院自动化研究所
项目金额: 10万元
中文摘要: 通过对正电子核素标记的示踪分子参与活体的生理生化过程进行PET扫描,能够从分子水平反映活体的生理生化变化。作为目前核医学诊断和研究最先进的分子显像方法,已从临床应用推广到了小动物科学实验,小动物PET在药物寻找和开发、疾病研究、基因显像等领域发挥重要作用。借助定量分析技术,对动态小动物PET图像数据做进一步的数学处理,可以得到非常有意义的定量指标,如葡萄糖代谢率、DNA合成率、蛋白质合成速率等,这些指标可以对器官的代谢功能进行准确的评测,具有非常重要的临床意义。在本课题中,本项目针对动态小鼠PET数据的定量分析中存在的问题,展开两个方面的研究,针对连续采血所带来的小动物的不适和操作处理的负担,研究不需要采血操作的血液函数自动提取算法;针对PET图像的高噪和低对比度的特点,研究对小鼠的心脏各组织成分进行自动分割的算法。实验结果从多个方面证明了提出的算法的有效性和鲁棒性。
中文关键词: 动态PET;输入函数;分割;定量分析
英文摘要: Positron Emission Tomography(PET) is a diagnostic imaging tool used by phsicians to look at metobolic processes in the body. It is based on a radio-labeled biologically active compound(tracer)being detected, and then modeling the resulting tracer activity over time to image the tissue function quantitatively. The main advantage of such dynamic functional imaging is that it enables this quantitative analysis of metabolic changes in tissues which is complementary information to that provided by the structural imaging modalitis. In this work, to address on the problems of the quantitative analysis in dynamic mouse microPET studies, we present and validate two methods. The first method is automaticlly segment cardiac components (left ventricle, myocaridium, right ventircle) based on constraint nonnegative matrix factorization, and the second one is almost noninvasively estimate the input function from dynamic mouse 18F-FDG microPET images and 1 late blood sample, which accounting for the spillover, partial-volume, delay and dispersion effects. The experimental results demonstrate the effectiveness of the proposed methods.
英文关键词: dynamic PET; input function; segmentation; quantitive analysis