当前位置: X-MOL 学术EJNMMI Phys. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Quantitative implications of the updated EARL 2019 PET–CT performance standards
EJNMMI Physics ( IF 4 ) Pub Date : 2019-12-26 , DOI: 10.1186/s40658-019-0257-8
Andres Kaalep 1 , Coreline N Burggraaff 2 , Simone Pieplenbosch 3 , Eline E Verwer 3 , Terez Sera 4, 5 , Josee Zijlstra 2 , Otto S Hoekstra 3 , Daniela E Oprea-Lager 3 , Ronald Boellaard 3, 5, 6
Affiliation  

Recently, updated EARL specifications (EARL2) have been developed and announced. This study aims at investigating the impact of the EARL2 specifications on the quantitative reads of clinical PET–CT studies and testing a method to enable the use of the EARL2 standards whilst still generating quantitative reads compliant with current EARL standards (EARL1). Thirteen non-small cell lung cancer (NSCLC) and seventeen lymphoma PET–CT studies were used to derive four image datasets—the first dataset complying with EARL1 specifications and the second reconstructed using parameters as described in EARL2. For the third (EARL2F6) and fourth (EARL2F7) dataset in EARL2, respectively, 6 mm and 7 mm Gaussian post-filtering was applied. We compared the results of quantitative metrics (MATV, SUVmax, SUVpeak, SUVmean, TLG, and tumor-to-liver and tumor-to-blood pool ratios) obtained with these 4 datasets in 55 suspected malignant lesions using three commonly used segmentation/volume of interest (VOI) methods (MAX41, A50P, SUV4). We found that with EARL2 MAX41 VOI method, MATV decreases by 22%, TLG remains unchanged and SUV values increase by 23–30% depending on the specific metric used. The EARL2F7 dataset produced quantitative metrics best aligning with EARL1, with no significant differences between most of the datasets (p>0.05). Different VOI methods performed similarly with regard to SUV metrics but differences in MATV as well as TLG were observed. No significant difference between NSCLC and lymphoma cancer types was observed. Application of EARL2 standards can result in higher SUVs, reduced MATV and slightly changed TLG values relative to EARL1. Applying a Gaussian filter to PET images reconstructed using EARL2 parameters successfully yielded EARL1 compliant data.

中文翻译:

更新后的 EARL 2019 PET-CT 性能标准的定量影响

最近,更新的 EARL 规范 (EARL2) 已经制定并公布。本研究旨在调查 EARL2 规范对临床 PET-CT 研究的定量读数的影响,并测试一种方法以启用 EARL2 标准,同时仍生成符合当前 EARL 标准 (EARL1) 的定量读数。13 项非小细胞肺癌 (NSCLC) 和 17 项淋巴瘤 PET-CT 研究被用于导出四个图像数据集——第一个数据集符合 EARL1 规范,第二个数据集使用 EARL2 中描述的参数重建。对于 EARL2 中的第三个 (EARL2F6) 和第四个 (EARL2F7) 数据集,分别应用了 6 mm 和 7 mm 高斯后置滤波。我们比较了定量指标(MATV、SUVmax、SUVpeak、SUVmean、TLG、以及肿瘤与肝脏和肿瘤与血池的比率)使用三种常用的分割/感兴趣体积 (VOI) 方法(MAX41、A50P、SUV4)通过这 4 个数据集在 55 个疑似恶性病变中获得。我们发现,使用 EARL2 MAX41 VOI 方法,MATV 减少了 22%,TLG 保持不变,SUV 值增加了 23-30%,具体取决于所使用的具体指标。EARL2F7 数据集生成的定量指标与 EARL1 最一致,大多数数据集之间没有显着差异 (p>0.05)。不同的 VOI 方法在 SUV 指标方面表现相似,但观察到 MATV 和 TLG 的差异。未观察到 NSCLC 和淋巴瘤癌症类型之间存在显着差异。EARL2 标准的应用可导致更高的 SUV、降低的 MATV 和相对于 EARL1 的略微变化的 TLG 值。
更新日期:2019-12-26
down
wechat
bug