磁共振扩散成像复杂心肌纤维结构解析与重建研究

The spatial structure of the myocardial fibers is of great clinical value and significance to the clinical diagnosis of heart diseases and study of the heart physiological characteristics. Diffusion magnetic resonance imaging (DMRI) is currently the only imaging technology to measurement of the diffusion motion of water molecules in the myocardial tissue in vivo.But Diffusion tensor magnetic resonance imaging (DT-MRI) that is currently used can only represent a single fiber traveling direction, can not describe the complex fiber structure of the myocardial fibers cross, bifurcation..This project will used multiple tensors and higher order tensor theory to model the structure of the complex myocardium fiber. We want to get the orientations of the complex myocardial fibers by solution and optimization of the model parameters.According to the information that we get, we will improved the algorithm based on probability for tracking the complex myocardium fiber to reconstruction the cardic fiber image. We will evaluate the algorithm by quantitative analysis of the simulation data and qualitative assessment of the doctors, while we will adopt grid technology to solve the vast computational complexity of data in the process of modeling, simulation, and myocardium fiber tracking problems. And we will get more accurate and fast reconstruction of the whole myocardium fiber structure ultimately, provide more reliable scientific basis for the clinical diagnosis of heart diseases.

获得心肌纤维的空间结构信息，对心脏的生理特性研究及心脏病的诊断具有重要的科学意义和临床价值。磁共振扩散成像(DMRI)是目前唯一能够在活体上测量心肌组织内水分子扩散运动与成像的技术，但目前使用的磁共振扩散张量成像(DT-MRI)方法只能表示单一的纤维走行方向，无法描述心肌纤维交叉、分叉等复杂纤维结构。.本课题拟采用多张量及高阶张量理论对复杂心肌纤维结构进行建模，通过模型参数求解及优化，获取体素内复杂心肌纤维结构的多个纤维走向，进而利用基于概率的纤维束追踪算法实现对复杂心肌纤维结构的重建。采用仿真数据定量分析结合医生定性评定的方案来评价算法的有效性，同时引入网格技术解决建模、仿真及心肌纤维追踪等过程中的海量数据计算问题，最终实现对整体心肌纤维结构的更准确、快速的重建。为心脏疾病的临床诊断提供更可靠的科学依据。

基于磁共振扩散成像技术(DMRI)的心肌纤维成像方法是目前唯一能够获得心肌纤维结构信息的无创探测手段，然而传统上该技术只能描述单一的纤维走行方向，无法表示心肌纤维交叉、分叉等复杂结构。. 本项目针对低角分辨率扩散成像条件下现有方法易受噪声影响的问题，通过体素内复杂纤维结构的建模、求解及优化，提高了复杂纤维结构估计的准确性。由于组织结构及噪声的影响，纤维重建过程存在着较大的不稳定性的问题，项目引入粒子滤波技术来实现复杂心肌纤维结构的重建与纤维束追踪。心脏DMRI成像技术存在着数据采集时间长、空间分辨率和信噪比低以及对运动敏感等问题，研究心肌DMRI快速成像技术，可缩短数据采集时间，减少运动的影响。此外，其它生理现象如水分子的灌注信息也是磁共振扩散成像研究的内容。