Somlai-Schweiger I, Ziegler S I
Nuklearmedizinische Klinik und Poliklinik, Klinikum rechts der Isar, Technische Universität München, Ismaninger Strasse 22, München 81675, Germany.
Med Phys. 2015 Apr;42(4):1825-35. doi: 10.1118/1.4914857.
A new concept for a depth-of-interaction (DOI) capable time-of-flight (TOF) PET detector is defined, based only on the detection of Cherenkov photons. The proposed "CHERENCUBE" consists of a cubic Cherenkov radiator with position-sensitive photodetectors covering each crystal face. By means of the spatial distribution of the detected photons and their time of arrival, the point of interaction of the gamma-ray in the crystal can be determined. This study analyzes through theoretical calculations and Monte Carlo simulations the potential advantages of the concept toward reaching a Cherenkov-only detector for TOF-PET with DOI capability. Furthermore, an algorithm for the DOI estimation is presented and the requirements for a practical implementation of the proposed concept are defined.
The Monte Carlo simulations consisted of a cubic crystal with one photodetector coupled to each one of the faces of the cube. The sensitive area of the detector matched exactly the crystal size, which was varied in 1 mm steps between 1 × 1 × 1 mm(3) and 10 × 10 × 10 mm(3). For each size, five independent simulations of ten thousand 511 keV gamma-rays were triggered at a fixed distance of 10 mm. The crystal chosen was PbWO4. Its scintillation properties were simulated, but only Cherenkov photons were analyzed. Photodetectors were simulated having perfect photodetection efficiency and infinite time resolution. For every generated particle, the analysis considered its creation process, parent and daughter particles, energy, origin coordinates, trajectory, and time and position of detection. The DOI determination is based on the distribution of the emission time of all photons per event. These values are calculated as a function of the coordinates of detection and origin for every photon. The common origin is estimated by finding the distribution with the most similar emission time-points.
Detection efficiency increases with crystal size from 8.2% (1 × 1 × 1 mm(3)) to 58.6% (10 × 10 × 10 mm(3)) and decreases applying a photon detection threshold of 5/10/20 photons to 6.3%/4.3%/0.7% and 49.3%/30.4%/2.8%, respectively. The detection rate in the six photodetectors is uniform due to the nearly isotropic cone emission. Most cones originated after a photoelectric effect interaction, with two dominating peaks for the kinetic energy of the electron at 422.99 and 441.47 keV. The detection distance between same-event photons defines the spatial resolution of the detector required for individual photon recognition, with 20% of the detected photons having their closest neighbor within a distance of 5% of the length of the cube. Same-event photons are detected within a time window whose width is determined by the crystal size, with values of 30 and 150 ps for a 1 × 1 × 1 mm(3) and a 10 × 10 × 10 mm(3) cube, respectively. The DOI reconstruction has an accuracy of approximately 23% of the length of the cube, with an average value of 2.2 mm for a 10 × 10 × 10 mm(3) CHERENCUBE.
The proposed concept requires a detector with high photodetection efficiency. The structure of the sensitive surface of the detector should be a two dimensional array of microcells, able to provide individual detection coordinates and time stamps. The microcell size determines the ability to recognize individual photons, influencing detection efficiency. The 3D DOI recognition relies on the accuracy of the time stamps and detection coordinates, without the need for a recognition of the projected patterns of photons. The refractive index of the material defines a detector intrinsic energy-based rejection of scattered PET events at the cost of reduced sensitivity.
仅基于切伦科夫光子的探测，定义了一种用于具有相互作用深度（DOI）功能的飞行时间（TOF）正电子发射断层扫描（PET）探测器的新概念。所提出的“切伦科夫立方探测器（CHERENCUBE）”由一个立方切伦科夫辐射体组成，其每个晶面都覆盖有位置灵敏光电探测器。通过检测到的光子的空间分布及其到达时间，可以确定晶体中伽马射线的相互作用点。本研究通过理论计算和蒙特卡罗模拟，分析了该概念在实现具有DOI功能的仅基于切伦科夫的TOF - PET探测器方面的潜在优势。此外，还提出了一种DOI估计算法，并定义了所提出概念实际应用的要求。
蒙特卡罗模拟由一个立方晶体组成，每个立方体面都耦合有一个光电探测器。探测器的灵敏面积与晶体尺寸精确匹配，晶体尺寸在1×1×1 mm³至10×10×10 mm³之间以1 mm步长变化。对于每种尺寸，在10 mm的固定距离处触发了五次独立的一万次511 keV伽马射线模拟。选择的晶体是钨酸铅（PbWO4）。模拟了其闪烁特性，但仅分析切伦科夫光子。模拟的光电探测器具有完美的光电探测效率和无限的时间分辨率。对于每个生成的粒子，分析考虑其产生过程、母粒子和子粒子、能量、源坐标、轨迹以及检测时间和位置。DOI的确定基于每个事件中所有光子发射时间的分布。这些值是针对每个光子根据检测坐标和源坐标计算的。通过找到发射时间点最相似的分布来估计共同源。
探测效率随晶体尺寸从8.2%（1×1×1 mm³）增加到58.6%（10×10×10 mm³），当应用5/10/20个光子的光子探测阈值时，分别降至6.3%/4.3%/0.7%和49.3%/30.4%/2.8%。由于几乎各向同性的锥形发射，六个光电探测器中的探测率是均匀的。大多数锥形在光电效应相互作用之后产生，电子动能有两个主导峰，分别为422.99和441.47 keV。同事件光子之间的探测距离定义了单个光子识别所需的探测器空间分辨率，20%的检测光子其最近邻在立方体边长的5%距离内。同事件光子在一个时间窗口内被检测到，其宽度由晶体尺寸决定，对于1×1×1 mm³和10×10×10 mm³的立方体，分别为30和150 ps。DOI重建的精度约为立方体边长的23%，对于10×10×10 mm³的CHERENCUBE，平均值为2.2 mm。
所提出的概念需要一个具有高光电探测效率的探测器。探测器灵敏表面的结构应该是一个微单元的二维阵列，能够提供单独的检测坐标和时间戳。微单元尺寸决定了识别单个光子的能力，影响探测效率。三维DOI识别依赖于时间戳和检测坐标的精度，无需识别光子的投影模式。材料的折射率以降低灵敏度为代价定义了探测器基于能量的对散射PET事件的固有排除。
