Guo Peiyuan, Spindler Simon, Rawlik Michal, Lu Jincheng, Men Longchao, Hong Mingzhi, Stampanoni Marco, Yin Hongxia, Xu Yan, Wang Zhenchang, Zhang Li, Wang Zhentian
Department of Engineering Physics, Tsinghua University, Beijing, China.
Key Laboratory of Particle & Radiation Imaging (Tsinghua University) of Ministry of Education, Beijing, China.
Med Phys. 2025 Apr;52(4):2155-2166. doi: 10.1002/mp.17630. Epub 2025 Jan 27.
X-ray grating-based dark-field imaging can sense the small angle scattering caused by object's micro-structures. This technique is sensitive to the porous microstructure of lung alveoli and has the potential to detect lung diseases at an early stage. Up to now, a human-scale dark-field CT (DF-CT) prototype has been built for lung imaging.
This study aimed to develop a thorough optimization method for human-scale dark-field lung CT and guide the system design.
We introduced a task-based metric formulated as the contrast-to-noise ratio (CNR) between normal and lesioned alveoli for system parameter optimization and designed a digital human-thorax phantom to fit the task of lung disease detection. Furthermore, a computational framework was developed to model the signal propagation in DF-CT and established the link between system parameters and the CNR metric.
We showed that for a DF-CT system, its CNR first increases and then decreases with the system auto-correlation length (ACL). The optimal ACL is mostly independent of system's visibility, and is only related to the phantom's properties, that is, its size and absorption. For our phantom, the optimal ACL is about 0.35 µm at the design energy of 60 keV. As for system geometry, increasing source-detector and isocenter-detector distance can extend the system's maximal ACL, making it easier for the system to meet the optimal ACL and relaxing the grating pitches. We proposed a set of parameters for a projective fringe system that can satisfy the simulated optimal ACL.
This study introduced a task-based metric and a process for DF-CT optimization. We demonstrated that for a given phantom, the detection performance of the system is optimized at a specific ACL. The optimization method and design principles are independent from the underlying dark-field imaging method and can be applied to DF-CT system design using different grating-based implementations such as Talbot-Lau interferometer (TLI) or projective fringe method.
基于X射线光栅的暗场成像能够检测由物体微观结构引起的小角度散射。该技术对肺泡的多孔微观结构敏感，具有早期检测肺部疾病的潜力。到目前为止，已构建了用于肺部成像的人体尺寸暗场CT（DF-CT）原型。
本研究旨在开发一种针对人体尺寸暗场肺部CT的全面优化方法，并指导系统设计。
我们引入了一种基于任务的度量标准，将其表述为正常肺泡和病变肺泡之间的对比度噪声比（CNR），用于系统参数优化，并设计了一个数字人体胸部模型以适应肺部疾病检测任务。此外，还开发了一个计算框架来模拟DF-CT中的信号传播，并建立系统参数与CNR度量之间的联系。
我们表明，对于DF-CT系统，其CNR随系统自相关长度（ACL）先增加后减小。最佳ACL大多与系统的可见性无关，仅与模型的属性有关，即其大小和吸收率。对于我们的模型，在60 keV的设计能量下，最佳ACL约为0.35 µm。至于系统几何结构，增加源探测器和等中心探测器的距离可以扩展系统的最大ACL，使系统更容易满足最佳ACL并放宽光栅间距。我们提出了一组适用于投影条纹系统的参数，该参数可以满足模拟的最佳ACL。
本研究介绍了一种基于任务的度量标准和DF-CT优化过程。我们证明，对于给定的模型，系统的检测性能在特定的ACL下得到优化。该优化方法和设计原则与底层暗场成像方法无关，可应用于使用不同基于光栅的实现方式（如Talbot-Lau干涉仪（TLI）或投影条纹法）的DF-CT系统设计。
