面向小动物成像的扩散荧光层析先进方法研究

发布时间：2018-05-23 07:57

本文选题：扩散荧光层析成像 + 扩散方程　；参考：《天津大学》2013年博士论文

【摘要】：扩散荧光层析成像(DFT)技术作为一种能够定量提供特异性分子探针空间分布信息的高灵敏度光学分子成像手段，被逐渐地应用于疾病小动物模型体内生化过程的在体跟踪和检测。生物医学研究的深入发展对DFT技术的应用范围、功能和指标都提出了更高的要求。本文致力于研究面向小动物成像的DFT先进方法，，其主要内容包括基于稳态扩散方程的差分DFT成像方法研究和实验验证，基于时域扩散方程的特征数据和全时间分辨DFT先进成像方法研究和实验验证，以及基于先进光子输运模型的DFT成像方法研究和模拟验证。本文发展了一种用于pH值敏感性检测的稳态差分DFT成像方法。该方法基于一种特殊设计的仿CT扫描模式的光子计数检测系统和Born归一化的差分图像重建算法，来实现对pH异常引起的荧光参数变化图像的快速测量和重建。在仿体实验中，通过定量性实验验证了该成像方法具有高精确度的定量能力，并在三种不同的目标体与背景荧光对比度下进行了pH敏感性检测实验验证。基于时域扩散方程，本文分别发展了基于多级离散小波变换的特征数据时域DFT成像方法和基于重叠时间门技术的全时间分辨DFT成像方法。前者利用多级离散小波变换来减少逆向模型中的重建参数，从而改善计算模型的病态性，提高荧光参数重建图像质量。后者使用了全时间分辨数据，采用重叠时间门技术最大限度地挖掘测量数据中所包含的信息，从而提高荧光产率重建图像的分辨率及量化度。此外，采用了基于时间相关单光子计数技术的时域测量系统对发展的时域DFT成像方法进行了实验验证。实验结果表明两种DFT成像方法均能够有效地提高荧光参数重建图像的量化度和空间分辨率。为了克服扩散近似光子输运模型本身的限制条件，本文分别发展了基于稳态辐射传输方程的DFT成像方法和基于早期到达光荧光输运模型的DFT成像方法。前者联合了离散立体角元法和有限差分法数值求解二维稳态辐射荧光传输方程，并采用了自然边界条件及准直光源模型。利用荧光蒙特卡罗模拟方法产生的正向数据对该成像方法进行了数值模拟验证，并且与基于扩散方程的稳态DFT算法的重建结果进行了比较。后者将早期到达光近似为直线传输的弹道光发展了荧光早期到达光输运模型。并利用全时间分辨时域DFT成像方法的正向数据进行了数值模拟验证，模拟结果证明了该方法的有效性以及对于荧光产率重建图像在空间分辨率上的提高。
[Abstract]:Diffusion fluorescence Tomography (DFT), as a highly sensitive optical molecular imaging method which can provide quantitative information on the spatial distribution of specific molecular probes, has been gradually applied to the in vivo tracking and detection of biochemical processes in small animal models of disease. The further development of biomedical research has put forward higher requirements for the application scope, function and index of DFT technology. This paper is devoted to the study of advanced DFT methods for small animal imaging, including the research and experimental verification of differential DFT imaging method based on steady-state diffusion equation. The characteristic data based on time-domain diffusion equation and the full-time resolved DFT advanced imaging method are studied and verified by experiments, and the DFT imaging method based on advanced photon transport model is studied and simulated. In this paper, a steady-state differential DFT imaging method for pH sensitivity detection is developed. This method is based on a specially designed photon counting detection system imitating CT scanning mode and Born normalized differential image reconstruction algorithm to realize the fast measurement and reconstruction of the fluorescence parameter change image caused by pH anomaly. The quantitative experiments show that the imaging method has high accuracy, and the pH sensitivity is tested under three different object and background fluorescence contrast. Based on the diffusion equation of time domain, this paper develops the time domain DFT imaging method based on multistage discrete wavelet transform and the full time resolved DFT imaging method based on overlapping time gate technique. In the former, multilevel discrete wavelet transform is used to reduce the reconstruction parameters in the reverse model, thus improving the ill-condition of the computational model and improving the quality of the reconstructed image with the fluorescence parameters. The latter uses full time resolved data and uses overlapping time gate technique to mine the information contained in the measurement data to improve the resolution and quantization of the reconstructed image. In addition, a time-domain measurement system based on time-dependent single-photon counting technique is used to verify the developed time-domain DFT imaging method. Experimental results show that both DFT imaging methods can effectively improve the quantization and spatial resolution of reconstructed images with fluorescence parameters. In order to overcome the limitation of the diffusion approximation photon transport model, the DFT imaging method based on the steady state radiation transfer equation and the DFT imaging method based on the early arrival photoluminescence transport model are developed in this paper. The former combines the discrete solid element method and the finite difference method to numerically solve the two-dimensional steady-state radiative fluorescence transfer equation and adopts the natural boundary conditions and the collimation light source model. The forward data generated by the fluorescence Monte Carlo simulation method are used to validate the proposed method, and the results are compared with the reconstruction results of the steady-state DFT algorithm based on diffusion equation. The latter developed an early arrival light transport model by approximating the early arrival light as a linear ballistic light. The forward data of full time resolved time domain DFT imaging method is simulated and validated. The simulation results show that the method is effective and the spatial resolution of the reconstructed image is improved.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：R310

【参考文献】