常规光源荧光CT的Geant4模拟及图像重建算法研究

X-ray Fluorescence Computed Tomography (XFCT) is an experimental technique for investigating the spatial distribution of elements in a sample. It has been investigated as a novel molecular imaging to diagnose cancer. XFCT can provide high resolution images, and its resolution is about two orders of magnitude higher than PET which is much used for diagnosing cancer without injecting radioactively labelled substance. It can improve diagnostic capacity of early-stage cancer greatly. But some factors such as depending on synchrotron radiation, self-absorption and too long imaging time influence the application of XFCT seriously. So XFCT based on conventional light source is urgent needed and related reconstructed method such as attenuation correction method should be discussed. In this project, monte carlo method, iterative reconstruction method and compressed sensing theory are used to explore the method of extracting desired signal, the attenuation correction reconstruction method of XFCT based on conventional light source and sparsely reconstruction method. It has significant sense for building XFCT based on conventional light source, extracting desired signal and reconstructing image, and hopes to reconstruct images accurately and quickly.

荧光CT是一种无损获取样品中元素三维分布的计算机断层扫描成像技术，该技术在癌症诊断中具有重大的潜在应用价值。它可以在不注射放射性标记物的情况下提供比肿瘤诊断中常用的正电子发射断层扫描成像术分辨率还高2个数量级的分子影像，这可以极大地提高肿瘤的早期诊断能力。但是荧光CT对于同步辐射装置的依赖、自吸收效应明显和成像时间过长等因素严重制约了荧光CT的应用。这就迫切需要建立基于常规光源的荧光CT，探索自吸收校正和稀疏投影重建的方法。本项目拟采用蒙特卡罗方法、迭代重建算法和压缩感知理论，研究常规光源荧光CT的有效信号提取方法；研究常规光源荧光CT的自吸收校正方法及CUDA技术在图像重建过程中的加速效果；研究荧光CT的稀疏角度图像重建方法。以期能够对常规光源荧光CT系统的搭建、信号的筛选和图像重建等过程提供科学依据，并实现常规荧光CT的快速高质量成像。

常规光源荧光CT在早期癌症诊断中具有重大的潜在应用价值。但是一方面荧光信号的产额较低，成像需要较强的入射光束或较长的扫描时间；另一方面基体对于荧光信号的吸收较强，导致自吸收效应严重，这些因素严重制约了其临床应用。为探索上述问题的解决方案，本项目采用蒙特卡罗方法、迭代重建算法和压缩感知理论，探索了常规光源荧光CT的有效信号提取方法，确立了基于多项式拟合的有效数据提取方法；研究了荧光信号在成像过程中的产生和吸收过程，建立了常规光源荧光CT的自吸收校正方法，提高了常规光源荧光CT的成像质量，并利用加速技术确保图像重建的速度；在自吸收算法的基础上研究了常规荧光CT的稀疏角度图像重建方法，基于Split-Bregman算法建立了稀疏角度扫描情况下的图像重建方法，确保了元素含量三维分布的快速有效获取。本工作为常规光源荧光CT系统的搭建、信号的筛选和图像重建等过程提供科学依据，同时对实现常规荧光CT的快速高质量成像具有重要意义。