Wu Hao, Chen Qin, Liu Yingxian, Chen Jidong, Deng Wanyue
Ultrasound Department, Sichuan Academy of Medical Sciences and Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, Sichuan 610072, P.R. China.
Ultrasound Department, Emergency General Hospital, Beijing 100028, P.R. China.
Oncol Lett. 2020 Nov;20(5):248. doi: 10.3892/ol.2020.12111. Epub 2020 Sep 16.
The present study aimed to evaluate the reproducibility and accuracy of the optimized algorithm of shear-wave elastography (SWE) in diagnosing solid thyroid nodules. Two hundred and sixty-three solid thyroid nodules in 248 patients who underwent conventional ultrasound and SWE, respectively, by two operators were scheduled for fine-needle aspiration or surgery. Elasticity indices of the mean, minimum and maximum of nodules (EI) and thyroid parenchyma (EInorm) were measured respectively in the same frame of elastographic images for three times by both operators. The intraobserver and interobserver reproducibility of the optimized algorithm were assessed by intraclass correlation coefficients (ICC). Diagnostic performance of the optimized algorithm was compared with that of conventional SWE measurements by receiver-operating characteristic (ROC) curves. Among a total of 243 nodules included, 121 were benign nodules and 122 were papillary thyroid carcinoma (PTC). Intraobserver reliability for EId and EIr was nearly perfect (ICC>0.80). Interobserver agreement for MEANd, MAXd, MEANr and MAXr was nearly perfect (ICC>0.80). MAXd had the largest areas under the ROC curve which was 0.82. Compared with conventional SWE, the optimized algorithm of SWE shows better reproducibility and performance in diagnosing solid thyroid nodules.
本研究旨在评估剪切波弹性成像（SWE）优化算法在诊断甲状腺实性结节中的可重复性和准确性。248例患者的263个甲状腺实性结节分别由两名操作者进行了常规超声和SWE检查，之后这些患者被安排进行细针穿刺或手术。两名操作者在弹性成像图像的同一帧中分别对结节的平均、最小和最大弹性指数（EI）以及甲状腺实质（EInorm）进行了三次测量。通过组内相关系数（ICC）评估优化算法的观察者内和观察者间的可重复性。通过受试者操作特征（ROC）曲线将优化算法的诊断性能与传统SWE测量的诊断性能进行比较。在总共纳入的243个结节中，121个为良性结节，122个为甲状腺乳头状癌（PTC）。EId和EIr的观察者内可靠性近乎完美（ICC>0.80）。MEANd、MAXd、MEANr和MAXr的观察者间一致性近乎完美（ICC>0.80）。MAXd的ROC曲线下面积最大，为0.82。与传统SWE相比，SWE优化算法在诊断甲状腺实性结节方面显示出更好的可重复性和性能。
