Maebatake Akira, Akamatsu Go, Miwa Kenta, Tsutsui Yuji, Himuro Kazuhiko, Baba Shingo, Sasaki Masayuki
Ann Nucl Med. 2016 Jan;30(1):68-74. doi: 10.1007/s12149-015-1032-0.
The aim of this study was to investigate the relationship between the NEC and TOF-PET image quality.
The National Electrical Manufactures Association and International Electrical Commission (NEMA IEC) body phantom with a 10-mm diameter sphere was filled with an 18F-FDG solution with a 4:1 radioactivity ratio. The PET data were acquired in the three-dimensional list mode for 20 min. We created frame data ranging from 1 to 5 min acquisition time, which were then reconstructed using the baseline ordered-subsets expectation maximization (OSEM), the OSEM + point spread function (PSF) algorithm, OSEM + time-of-flight (TOF) algorithm and OSEM + PSF + TOF algorithm. The PET images were analyzed according to the noise-equivalent count (NEC), the coefficients of variance of the background (CVBG), the maximum count (CVmax) and the contrast (CVCONT). The results were compared with the recommended value according to the guidelines for the oncology FDG-PET/CT protocol.
The NEC was higher than the recommended value at 3 min or longer acquisition time. The CVBG lower than 15% were obtained at 3 min acquisition time without TOF and at 2 min acquisition time with TOF. The CVBG of 10% or lower were obtained at 5 min or longer acquisition time without TOF and at 4 min or longer acquisition time with TOF. Both the CVmax and CVCONT lower than 10% were obtained at 3 min or longer acquisition time without TOF and at 1 min acquisition or longer with TOF. No particular relationships were observed between the frame number and degree of the variation in the image quality. The CVCONT significantly correlated with the NEC for the data reconstructed without TOF information, while there were no significant correlations between these useful metrics for the data reconstructed with TOF.
This study demonstrated that the NEC is not a useful metric for the evaluation of the image quality on TOF-PET images.
本研究旨在探讨噪声等效计数（NEC）与飞行时间（TOF）正电子发射断层扫描（PET）图像质量之间的关系。
将直径10毫米球体的美国国家电气制造商协会和国际电工委员会（NEMA IEC）体模填充放射性活度比为4:1的18F-氟代脱氧葡萄糖（18F-FDG）溶液。以三维列表模式采集PET数据20分钟。我们创建了采集时间从1到5分钟不等的帧数据，然后使用基线有序子集期望最大化（OSEM）、OSEM + 点扩散函数（PSF）算法、OSEM + TOF算法和OSEM + PSF + TOF算法进行重建。根据噪声等效计数（NEC）、背景方差系数（CVBG）、最大计数（CVmax）和对比度（CVCONT）对PET图像进行分析。将结果与肿瘤学FDG-PET/CT协议指南中的推荐值进行比较。
在采集时间为3分钟或更长时，NEC高于推荐值。在无TOF的情况下，采集时间为3分钟时CVBG低于15%；在有TOF的情况下，采集时间为2分钟时CVBG低于15%。在无TOF的情况下，采集时间为5分钟或更长时，以及在有TOF的情况下，采集时间为4分钟或更长时，CVBG低于10%。在无TOF的情况下，采集时间为3分钟或更长时，以及在有TOF的情况下，采集时间为1分钟或更长时，CVmax和CVCONT均低于10%。未观察到帧数与图像质量变化程度之间存在特定关系。对于无TOF信息重建的数据，CVCONT与NEC显著相关，而对于有TOF重建的数据，这些有用指标之间无显著相关性。
本研究表明，NEC不是评估TOF-PET图像质量的有用指标。
