Ultra-high sensitivity, high spatial resolution single photon emission tomography using mechanical flux manipulation.

该项目的目标是开发硬件和软件，以展示开创性的能力 一种新的单光子放射性核素(SPR)成像技术，其灵敏度有可能提高1000倍 与使用平行孔的临床SPECT成像相比，体积空间分辨率提高了100倍 准直器。我们将我们的新成像方法称为机械通量操纵(MFM)。MFM利用 高分辨率像素化探测器、高带宽数据采集电子设备和一种新的图像重建 利用探测器流量信息实现50%探测效率目标的方法 用于光子撞击MFM探测器和<lt；2 mm重建图像分辨率。MFM是SPR断层扫描 成像技术。MFM与传统SPECT的两个主要区别是无准直器 并使用通量概率分布与响应线(LOR)计数来重建图像。 MFM拒绝接受这样的概念，即必须知道每个检测到的光子的方向才能准确地 从单光子放射性核素发射物体重建图像。相反，MFM收集的是 通过移动机械衰减器，记录每个晶体的流量是如何改变的 (MA)在发射对象和探测器之间。使用通量信息，每个检测到的入射方向 图像重建不需要光子。MFM与SPECT的进一步区别在于它使用全3D 图像重建，而不是2D数据堆叠。而MFM将支持一般的单光子断层扫描 成像协议，这一阶段第一阶段提案的重点是证明人类大脑成像的可行性。 该项目由三个具体目标组成。第一个目标是扩展和验证SimSET蒙特卡罗 模拟工具，模拟包括真实世界效果的MFM扫描仪。此扩展模块的主要组件 就是能够模拟连续的MA运动。另一个子目标是制造一个原型MA组件 和全功能像素化探测器面板，用于收集实验数据，用于验证SimSET蒙特卡罗 卡洛软件工具。该项目的第二个目标是将MFM图像重建软件扩展到3D 并结合所有校正以支持定量成像。扩展到全3D图像重建将 需要更多的计算资源和算法优化，这样代码才能运行 效率很高。其中一个子目标是使用GPU处理器实现重建软件。第三个目标 是使用来自特定目标1的经过验证的蒙特卡罗工具和全3D图像重建代码 在目标2中开发，以优化高分辨率人脑成像的MFM成像系统的设计。 在这个项目成功完成后，我们将寻求额外的资金来建立一个MFM系统的原型，以 使用临床可行的方案支持&lt；2 mm图像分辨率人脑成像系统。

The goal of this project is to develop both hardware and software to demonstrate the ground breaking capabilities of a new single photon radionuclide (SPR) imaging technique with the potential for >1000 times gain in sensitivity and >100 times gain in volumetric spatial resolution compared to clinical SPECT imaging using parallel-hole collimators. We refer to our new imaging methodology as mechanical flux manipulation (MFM). MFM utilizes high resolution pixelated detectors, high bandwidth data acquisition electronics and a novel image reconstruction methodology utilizing detector flux information to achieve target performance goals of >50% detection efficiency for photons impinging an MFM detector and <2 mm reconstructed image resolution. MFM is a SPR tomographic imaging technique. The two main features that differentiate MFM from traditional SPECT are collimator-less detectors and the use of flux-probability distributions versus line of response (LOR) counts to reconstruct images. MFM rejects the notion that the direction of every detected photon must be known in order to accurately reconstruct images from a single photon radionuclide emitting object. Instead, MFM collects flux information on a crystal by crystal basis and records how the flux to each crystal is altered by moving a mechanical attenuator (MA) between the emission object and the detector. Using flux information, the incident direction of each detected photon is not required for image reconstruction. MFM is further differentiated from SPECT in that it uses fully 3D image reconstruction rather than stacks of 2D data. While MFM will support general single photon tomographic imaging protocols, the focus of this Phase I proposal is to demonstrate feasibility for human brain imaging. This project is consists of three specific aims. The first aim is to extend and validate the SimSET Monte Carlo simulation tool to simulate an MFM scanner including real-world effects. The main component of this extension is to be able to simulate continuous MA motion. An additional sub-aim is to fabricate a prototype MA assembly and fully functional pixelated detector panel to collect experimental data with which to validate the SimSET Monte Carlo software tools. The second aim of the project is to expand the MFM image reconstruction software to 3D and to incorporate all corrections to support quantitative imaging. Extending to fully 3D image reconstruction will require significantly more computing resources and optimization of the algorithms so that the code can run efficiently. One of the sub-aims is to implement the reconstruction software using GPU processors. The third aim is to use the validated Monte Carlo tools from specific aim 1 and the fully 3D image reconstruction code developed in aim 2 to optimize the design of a MFM imaging system for high resolution human brain imaging. After successful completion of this project, we will seek additional funding to build a prototype MFM system to support <2mm image resolution human brain imaging system using clinically feasible protocols.