不同OSEM参数对^(99m)Tc SPECT/CT定量结果的影响  

Impact of Different Ordered-subsets Expectation Maximization Reconstruction Parameters on the Quantitative Results of^(99m)Tc SPECT/CT

作　　者：龚成鹏 胡佳[1] 李坤[1] 胡帆[1] 胡蒙蒙 高再荣[1] Gong Chengpeng;Hu Jia;Li Kun(Department of Nuclear Medicine Union Hospital,Tongji Medical College,Huazhong University of Science and Technology,Hubei Key Laboratory of Molecular Imaging Wuhan 430022,China)

机构地区：[1]华中科技大学同济医学院附属协和医院核医学科,分子影像湖北省重点实验室,武汉430022

出　　处：《华中科技大学学报（医学版）》2021年第6期766-770,799,共6页Acta Medicinae Universitatis Scientiae et Technologiae Huazhong

基　　金：华中科技大学自主创新研究基金资助项目(No.2017KFYXJJ231)。

摘　　要：目的探讨有序子集最大期望值重建算法(ordered-subsets expectation maximization,OSEM)中不同迭代次数与子集数对^(99m)Tc SPECT/CT定量结果的影响。方法在美国电气制造商协会(NEMA)国际电工委员会(IEC)体模的6个不同大小的小球(直径分别为37、28、22、17、13、10 mm)中填充入0.54 MBq/mL的^(99m)Tc与碘克沙醇混合液(含碘量约0.3%),改变罐体放射性浓度形成不同的靶本比(32∶1、16∶1、8∶1、4∶1)。使用美国GE公司Discovery NM/CT 670 Pro型SPECT/CT显像仪对不同靶本比的体模行临床协议下的SPECT/CT采集,对采集的数据进行衰减校正、散射校正、分辨率恢复校正,使用GE Xeleris Version 4.0 Q-Metrix软件以CT阈值法自动勾画各小球感兴趣体积(volume of interest,VOI),并选择以下OSEM参数对小球定量分析。固定OSEM 10个子集,分别行1、2、4、8、12、16、24、32次迭代,定量计算不同迭代次数下各小球恢复系数(recovery coefficient,RC)(RC=测量活度/真实活度);固定8次迭代,分别行2、6、10、14、18、20、22、26、32个子集,计算不同子集下各小球RC。变换靶本比模型重复扫描,RC取平均值,采用Pearson相关性分析评价各小球RC与迭代次数、子集数的相关性,分析各小球RC随迭代次数、子集数变化趋势。结果各小球定量RC与迭代次数、子集数呈正相关,均随迭代次数增加而增加(小球由大到小r=0.723、0.723、0.725、0.729、0.800、0.864,均P<0.05),随子集数增加而增加(小球由大到小r=0.800、0.709、0.744、0.742、0.832、0.865,均P<0.05);随着迭代次数、子集数不断增加RC变化越来越缓慢逐渐趋于平稳,直径≥22 mm的3个小球8次迭代10个子集后RC增加趋于平稳,直径≤17 mm的3个小球16次迭代20个子集后RC趋于平稳。结论OSEM参数选择对^(99m)Tc SPECT定量分析影响显著。临床工作中建议当热区直径≥22 mm时选择8次迭代,10个子集;直径≤17 mm时选择16次迭代,20个子集。Objective To evaluate the effects of different ordered-subsets expectation maximization(OSEM)reconstruction parameters on the quantitative results of^(99m)Tc SPECT/CT.Methods Six different spheres(diameter:37,28,22,17,13,10 mm)in National Electrical Manufacturers Association International Electrotechnical Commission(NEMA IEC)model were filled with a mixture of 0.54 MBq/mL^(99m)Tc and iodixanol.The mixture iodine content was about 0.3%,which led to different target background ratio(T/B)(32∶1,16∶1,8∶1,4∶1)by changing the radioactivity concentration of the cylinder.Routine imaging was performed on different T/B phantom which was scanned by SPECT/CT(Discovery NM/CT 670 Pro),attenuation correction,scattering correction and resolution recovery correction were performed on the collected data.The volume of interest(VOI)of each sphere was automatically delineated by CT threshold using GE Xeleris Version 4.0 Q-Metrix software,and the following OSEM parameters selected to quantitatively analyze the VOI.Fixed OSEM with 10 subsets,we performed 1,2,4,8,12,16,24,and 32 iterations respectively,and quantitatively calculated recovery coefficient(RC)(RC=measuring activity/true activity)of each sphere under different iteration frequencies.By fixing 8 iterations,we ran 2,6,10,14,18,20,22,26,and 32 subsets severally,and calculated RC of each sphere under different subsets.We changed the T/B model to repeat the scan,took the average values of RCs,used the Pearson correlation to analyze the correlation between RCs of each sphere and the number of iterations and the number of subsets,and analyzed the trend of the RC of each sphere changing with the number of iterations and the number of subsets.Results There were significant relationships between the number of iterations and RC(r values from large sphere to small:0.723,0.723,0.725,0.729,0.800,0.864,all P<0.05).There were significant relationships between the number of subsets and RC(r values from large sphere to small:0.800,0.709,0.744,0.742,0.823,all P<0.05).With the increasing

关 键 词：体层摄影术 发射型计算机 单光子 体层摄影术 X线计算机 有序子集最大期望值法 ^(99m)Tc 

分 类 号：R814[医药卫生—影像医学与核医学]