Speaker
Abstract
Conventional SPECT for high-energy gamma photons is fundamentally bottlenecked by the resolution-sensitivity trade-off inherent to mechanical collimation. While detector self-collimation (SC) offers a high-efficiency alternative, its performance at 511 keV is severely degraded by multiplexing ambiguities and inter-crystal scattering. We propose Joint Self-Collimation Compton (JSCC) imaging, a novel methodology that synergistically integrates the geometric pathways of a multi-layer interspaced MATRICES architecture with the physical constraints of Compton scattering kinematics. By extracting Compton cone information to "screen" and disambiguate overlapped projection trajectories, JSCC significantly enhances the signal-to-noise ratio (SNR) while preserving the intrinsic high spatial resolution of the SC design. Following a Cramer-Rao Lower Bound (CRLB)-driven system parameter optimization, we developed a unified list-mode MLEM reconstruction framework for both single-photon and scattering events. Comprehensive Monte Carlo simulations demonstrate that JSCC achieves superior contrast recovery, faster convergence, and higher geometric fidelity than standalone SC, even under ultra-short 10-second dynamic acquisitions. The identified scale-dependent performance gain underscores the immense potential of JSCC for large-field-of-view clinical high-energy SPECT and theranostic applications.
摘要
传统的高能伽马光子SPECT成像技术从根本上受限于机械准直固有的分辨率-灵敏度权衡。虽然探测器自准直（SC）提供了一种高效的替代方案，但其在511 keV能量下的性能会因多路复用歧义和晶间散射而严重下降。我们提出了一种联合自准直康普顿（JSCC）成像方法，该方法巧妙地将多层间隔MATRICES架构的几何路径与康普顿散射运动学的物理约束相结合。通过提取康普顿锥信息来“筛选”和消除重叠的投影轨迹，JSCC显著提高了信噪比（SNR），同时保持了SC设计固有的高空间分辨率。基于克拉默-拉奥下界（CRLB）驱动的系统参数优化，我们开发了一种统一的列表模式MLEM重建框架，适用于单光子和散射事件。全面的蒙特卡罗模拟表明，即使在超短的10秒动态采集条件下，JSCC也能实现比独立SC更优异的对比度恢复、更快的收敛速度和更高的几何保真度。所发现的与尺度相关的性能提升凸显了JSCC在大视野临床高能SPECT和诊疗一体化应用中的巨大潜力。
| 关键词 | 单光子断层成像，高能成像，自准直，康普顿成像，诊疗一体化 |
|---|---|
| Keywords | SPECT, High energy imaging, Self-collimation, Compton imaging, Theranostics |
