1.Institute of Modern Physics, Chinese Academy of Sciences, Lanzhou 730000, China 2.School of Nuclear Science and Technology, University of Chinese Academy of Sciences, Beijing 100049, China
Fund Project:Project supported by the National Natural Science Foundation of China (Grant No. 11675279) and the Program of "Light of the West", China.
Received Date:21 December 2018
Accepted Date:15 March 2019
Available Online:01 June 2019
Published Online:05 June 2019
Abstract:Compared with traditional gamma-ray imaging equipment, the Compton camera is a very promising imaging device in nuclear medicine and molecular imaging, and has a strong potential application in monitoring beams in heavy-ion-therapy because of its high efficiency feature. A demonstration device for heavy ion cancer treatment with complete intellectual property right has been built at Institute of Modern Physics, Chinese Academy of Sciences in Wuwei city of Gansu Province. At present the device is being up-graded, and the heavy ion cancer treatment is being generalized in national wide. In view of the broad prospects of heavy ion cancer treatment, the imaging resolution of Compton camera is analyzed theoretically, and three errors effecting the imaging resolution, which are energy resolution, position resolution of detector and the Doppler effect, are determined. Then the three errors are simulated by using the Geant 4 packages. The physical process in simulation is selected as the G4EMPenelopePhysics model, which makes the atomic shell cross section data for low energy physical process used directly. The Compton camera geometry consists of two layers of detectors. The layer close to $\gamma$ source is called detector and the other one is called absorption detector. The material of scatter detector is selected as low-Z silicon and carbon, and the absorb detector is high-Z germanium. The thickness value of scatter detector and absorb detector are both 20 mm. The spacing between the two layers is 100 mm. The simulation results by Geant 4 are used to reconstruct the image of point-like $\gamma$ source through using the back-projection algorithm. The simulation results and the re-constructed images indicate that the difference between the image full width at half maximum induced by 2 mm position resolution and that induced by 5.0% relative energy resolution of scatter detector is about 10%, and amount to that by the Doppler effect of Silicon. For the $\gamma$ ray with energies of several hundred keV, the energy resolution of Si detectors is easily better than 1.0% in practice. Therefore, the detector's position resolution dominates the image quality of the Compton camera. Considering the Doppler effect, manufacturing techniques and imaging efficiency, 2.0 mm-sized crystal unit and 1.0% energy resolution power is suggested for practically manufacturing the Compton camera. Keywords:Compton camera/ heavy ion therapy/ Geant 4 simulation/ back-projection imaging
根据第2节的分析, 康普顿相机的成像需要足够数目的有效事件. 因此, 为了定量研究影响康普顿相机成像分辨的每个源项, 应该观察其在大量有效事件中的分布曲线. 为此, 本工作用Geant 4软件包对各个源项进行了模拟. 为了能够模拟单一源项引起的误差, 即考虑某一源项时其他源项为一精确值, 建模如下: 康普顿相机的散射探测器由一整块Si晶体构成, 其尺寸为20 mm × 200 mm × 200 mm, 吸收探测器由一整块Ge晶体构成, 尺寸也为20 mm × 200 mm × 200 mm, 两者相距100 mm. 放射源为各向同性点源, 辐射的$ \gamma $射线能量为100—1000 keV, 放置在散射探测器正上方30 mm 处. 模拟中选用的物理模型为penelope 包. 对于一次有效事件, 该简化的探测器排布可精确模拟得到($ r_{1}, E_{1} $)和($ r_{2}, E_{2} $), 再由(1)式得到散射角$ \theta $. 对于探测器的能量分辨导致的散射角不确定度$ \Delta\theta_{\rm E} $, 模拟中, 以精确值$ E_{1} $为中心, $ \Delta E_{1} $为半高全宽(FWHM), 随机产生一个能量值用Erand代替(1)式中的$ E_{1} $, 计算得到$ \theta _{\rm E} $, 从而$ \Delta\theta_{\rm E} = \left| \theta - \theta _{\rm E} \right| $. 在以上条件下, 本工作模拟了相对能量分辨从0.3%—5%时$ \Delta\theta_{\rm E} $的分布曲线, 如图2所示. 可以看出, $ \Delta\theta_{\rm E} $的分布为指数衰减函数, 且随探测器能量分辨本领的提高, 指数衰减因子变大. 当相对分辨本领为0.3%时, $\Delta\theta_{\rm E}$集中分布在非常窄的范围内(最大几十$ \rm {mrad} $). 模拟结果与(2)式预测的“$ \Delta\theta_{\rm E} $与$ \Delta E_{1} $成正比例关系”一致. 考虑到实际制作工艺, 要求制作的探测器的能量分辨本领达到1%的水平. 对于不同能量的$ \gamma $射线, 随着射线能量的升高, $ \Delta\theta_{\rm E} $分布的指数衰减因子增大, 如图3 所示. 这与(2)式的预测结果一致. 图 3 康普顿散射角不确定度$ \Delta\theta_{\rm E} $分布的模拟结果, 相对能量分辨为1.0%, 初始$\gamma$射线能量分别为100 和1000 keV Figure3. Simulated distribution of $ \Delta\theta_{\rm E} $. The relative energy resolution is fitted to 1.0%, the initial $\gamma$-ray energy is 100 and 1000 keV, respectively.
图 2 康普顿相机的能量分辨本领引起的康普顿散射角不确定度$ \Delta\theta_{\rm E} $分布的模拟结果, 相对能量分辨$\Delta E/E$取值从0.3%至5%, 初始$\gamma$射线能量为600 keV Figure2. Simulated distribution of the uncertainty of Compton scattering angle caused by the resolving power of Compton camera. The value of $\Delta E/E$ is from 0.3% to 5%. The initial $\gamma$-ray energy is 600 keV.
对于探测器的位置分辨引起的不确定度$ \Delta\theta_{\rm P} $分布, 由模拟中精确得到的$ (r_{1} $, $ r_{2}) $以及$ \gamma $射线的初始位置$ r_{0} $, 根据三角关系得到散射角$ \theta $. 另外, 以精确值$ r_{1} $, $ r_{2} $为中心, $ \Delta x $为FWHM, 分别随机产生随机值$ r_{1}^{\rm {rand}} $和$ r_{2}^{\rm rand} $, 再结合$ r_{0} $三点求得同一事件的$ \theta_{\rm P} $, 从而$ \Delta \theta_{\rm P} = \left| \theta-\theta_{\rm P} \right| $. 模拟结果如图4所示. 由图可知, $ \Delta \theta_{\rm P} $也呈指数衰减分布, 且随着$ \Delta x $的减小, 指数衰减因子迅速增大. 这意味着实际制作探测器时, 探测单元尺寸应该尽可能做小. 对比图2的结果, 当$ \Delta x $ = 2.0 mm时, 引起的散射角误差接近$ \Delta E/E $ = 5%时导致的误差. 考虑到实际制作工艺、$ \gamma $射线在探测器中的多次响应及后续电子学读取系统的复杂性, 散射晶体单元尺寸$ \Delta x $ =2.0 mm是实际可行的, 这时候只要能量分辨本领小于5%, 位置分辨导致的散射角不确定性即为主要因素. 图 4 康普顿相机的位置分辨本领引起的康普顿散射角不确定度$ \Delta\theta_{\rm P} $分布的模拟结果, 位置分辨$\Delta x$取值范围为0.5—3.0 mm. 初始$\gamma$射线能量为600 keV Figure4. Simulated distribution of the uncertainty of Compton scattering angle caused by the position resolving power of Compton camera. The value of $\Delta x$ is from 0.5 mm to 3.0 mm. The initial $\gamma$-ray energy is 600 keV.
图5展示了不同散射材料导致的散射角不确定度分布的模拟结果. 模拟中, 由精确的(r1, E1)和(r2, E2)及康普顿散射公式(1)得到$ \theta_{1} $, 由(r1, r2)以及$ \gamma $射线的初始位置$ r_{0} $得到$ \theta_{2} $, 从而$ \Delta \theta_{\rm C} = $$ \left| \theta_{1}-\theta_{2} \right| $. 模拟结果显示, 随着散射晶体原子序数的减小, 指数衰减因子增大. 这是由于C晶体中电子的平均动量及束缚能小于Si晶体中的电子. 注意到该内禀误差的值也在几十mrad的范围, 即该内禀误差与$ \Delta $x = 2.0 mm或$ \Delta E/E $ = 5%时引起的误差相当. 图 5 康普顿散射角不确定度$ \Delta\theta_{\rm C} $分布的模拟结果, 初始$\gamma$射线能量为600 keV, 散射材料分别为C和Si Figure5. Simulated distribution of $ \Delta\theta_{\rm C} $. The initial $\gamma$ ray energy is 600 keV. The material of scattering detector is C and Si, respectively.