An inter-laboratory comparison study of image quality of PET scanners using the NEMA NU 2-2001 procedure for assessment of image quality
An inter-laboratory comparison study of image quality of PET scanners using the NEMA NU 2-2001 procedure for assessment of image quality
复制标题
DOI：
10.1088/0031-9155/50/10/001
复制
发表时间：
2005-05-21
影响因子：
3.5
通讯作者：
Samal, M
中科院分区：
文献类型：
--
作者：
Bergmann, H;Dobrozemsky, G;Samal, M
在奥地利进行了一项实验室间比较研究，以评估PET扫描仪的图像质量。调查包括专用PET扫描仪（D-PET, n = 8）和重合相机（GC-PET, n = 7）。图像质量的测量基于NEMA（国家电气制造商协会）NU 2-2001协议和IEC（国际电工委员会）体模。后者包含六个直径从37毫米到10毫米不等的可填充球体和一个“肺”插入物。2个最大的病灶L1-2模拟冷病灶，4个较小的病灶L3-6充满F-18，活性浓度与背景的比值分别为8:1和4:1。研究中的获取和重建采用了参与机构的标准肿瘤处理方案。球体的对比计算采用全自动程序进行。对比质量指数（cqi）通过对单个对比值求和得到反映全局性能的对比质量指数。根据NEMA协议计算的其他图像质量参数是背景变异性和用于校正衰减和散射的相对误差。得到的对比值L1为61 +/- 16和37 +/- 14 (D-PET和GC-PET分别为1%的对比SD)， L2为57 +/- 16和29 +/- 16，L3为46 +/- 10和26 +/- 6.3，L4为37 +/- 10和15 -/+ 4.3，L5为26 +/- 11.5和6.1 +/- 2.5，L6为14 +/- 7.1和2.6 +/- 2.6，D-PET系统始终优于GC-PET系统。CQIs允许扫描仪的排名，也证明了D-PET和GC-PET系统之间的明确区别。GC-PET系统的背景变异性最大；仅在D-PET系统中，衰减和散射校正的相对误差与图像质量显著相关。该研究表明，不仅在GC-PET和D-PET系统之间，而且在单个D-PET系统之间，图像质量存在相当大的差异，这可能会对图像的临床解释和定量指标（如标准化摄取值）的测量产生影响。该研究为参与者提供了有价值的反馈，并为提高不同实验室之间PET图像和定量指标的互换性提供了基线数据。
An inter-laboratory comparison study was conducted to assess the image quality of PET scanners in Austria. The survey included both dedicated PET scanners (D-PET, n = 8) and coincidence cameras (GC-PET, n = 7). Measurement of image quality was based on the NEMA (National Electrical Manufacturers Association) NU 2-2001 protocol and the IEC (International Electrotechnical Commission) body phantom. The latter contains six fillable spheres ranging in diameter from 37 mm down to 10 mm and a 'lung' insert. The two largest lesions L1-2 simulate cold lesions, the four smaller ones (L3-6) are filled with F-18 and activity concentration ratios relative to background of 8:1 and 4:1, respectively. Acquisition and reconstruction in the study employed the participating institutes' standard oncological processing protocol. Calculation of contrast of the spheres was performed with a fully automated procedure. Contrast quality indices (CQIs) reflecting global performance were obtained by summing individual contrast values. Other image quality parameters calculated according to the NEMA protocol were background variability and relative error for correction of attenuation and scatter. Contrast values obtained were 61 +/- 16 and 37 +/- 14 for L1 (per cent contrast SD for D-PET and GC-PET, respectively), 57 +/- 16 and 29 +/- 16 for L2, 46 +/- 10 and 26 +/- 6.3 for L3, 37 +/- 10 and 15 -/+ 4.3 for L4, 26 +/- 11.5 and 6.1 +/- 2.5 for L5, 14 +/- 7.1 and 2.6 +/- 2.6 for L6, with D-PET systems consistently being superior to GC-PET systems. CQIs permitted ranking of the scanners, also demonstrating a clear distinction between D-PET and GC-PET systems. Background variability was largest for GC-PET systems; the relative error of attenuation and scatter correction was significantly correlated with image quality for D-PET systems only. The study demonstrated considerable differences in image quality not only between GC-PET and D-PET systems but also between individual D-PET systems with possible consequences for clinical interpretation of images and measurement of quantitative indices such as the standardized uptake value. The study provided valuable feedback to the participants as well as baseline data for improving interchangeability of PET images and of quantitative indices between different laboratories.