3T磁共振非高斯模型扩散成像评价直肠癌放疗疗效的定量化分析

项目名称： 3T磁共振非高斯模型扩散成像评价直肠癌放疗疗效的定量化分析

项目编号： No.81471640

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 医药、卫生

项目作者： 孙应实

作者单位： 北京大学

项目金额： 75万元

中文摘要： 建立早期、准确的肿瘤疗效评价方法和明确其理论基础是指导肿瘤治疗的关键。影像学手段对于直肠癌新辅助治疗早期评效意义重大。本课题组前期研究显示磁共振扩散成像指标ADC值可以用来早期监测直肠癌放疗疗效。但基于高斯模型推导的ADC值在肿瘤成分混杂时,由于组织辨识率不高而导致部分肿瘤疗效出现错判。本研究利用3T磁共振多高b值DWI拟合出更接近真实扩散特征的非高斯FROC模型，通过微积分推导提取出反映肿瘤组织均匀度和扩散距离的参数，以此更好的区分肿瘤和正常组织。研究包括基础动物实验和临床应用评效两个部分，通过影像-病理对照阐明DWI非高斯模型各参数演变的病理组织学机制，挖掘并筛选更稳定、更敏感的反映直肠癌新辅助治疗效果的DWI模型参数，同时确定评效和监测的敏感时间窗，为直肠癌新辅助放疗个体化的实现提供新的思路、理论基础和依据。

中文关键词： 直肠肿瘤；磁共振成像；扩散成像；非高斯模型；疗效评价

英文摘要： It is of great importance to establish an early response evaluaion criteria to guide clinical treatment. Radiological modalities have significant value in the early evaluation of response to the neo-adjuvant therapies for patients with rectal cancer. The study applicants have found diffusion-weighted magnetic resonance imaging with the ADC value can be used to predict the response. However，further studies showed that ADC values that based on Gaussian model may perform wrong judgments of responses, because of the low soft tissue resolution caused by the heterogeneous tumor components. Our team has simulated the non- Gaussian FROC model which more approaches the real diffusion, using multi-high-b-value DWI on 3T scanner, and extracted parameters those can better discriminate the tumor and the normal tissue than the ADC value. In this, we plan to carry out a study including laboratory animal tests and clinical trials. We propose to uncover the more stable and sensitive parameters in the early prediction of treatment response to the neo-adjuvant therapies for patients with rectal cancer, and at the same time clarify the histopathological mechanism of the evolution of the FROC model parameters. Furthermore, we aim to make the time point for evaluation and surveillance confirmed, in order to provide effective parameters for early quantitative response evaluation and finally realize the personalized medicine.

英文关键词： rectal neoplasms;magnetic resonance imaging;diffusion-weighted imaging;non-Gaussian model;response evaluation