Wang Jia-Yu, Feng Shao-Yang, Xu Jian-Wei, Li Jun, Chu Liang, Cui Xin-Wu, Dietrich Christoph F
Sino-German Tongji-Caritas Research Center of Ultrasound in Medicine, Department of Medical Ultrasound, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China.
Department of Ultrasound, Sixth People's Hospital of Zhengzhou, Zhengzhou, China.
J Ultrasound Med. 2020 Aug;39(8):1537-1546. doi: 10.1002/jum.15242. Epub 2020 Feb 20.
To evaluate the usefulness of the contrast-enhanced ultrasound (CEUS) Liver Imaging Reporting and Data System (LI-RADS) in diagnosing focal liver lesions (FLLs) by inexperienced radiologists.
Images and clinical data from 258 patients at risk for hepatocellular carcinoma who underwent CEUS were collected retrospectively. Two trained inexperienced radiologists and 2 experienced radiologists reviewed all CEUS clips. Each inexperienced radiologist assigned a CEUS LI-RADS category for each observation and labeled it benign or malignant independently. Each experienced radiologist labeled each lesion malignant or benign independently using a conventional diagnostic method. Interobserver agreement of CEUS LI-RADS was analyzed by the κ test. The overall diagnostic accuracy of the LI-RADS category and conventional diagnosis was described by the sensitivity, specificity, positive predictive value, and negative predictive value. All test results were considered significant at P < .05.
A κ value of 0.774 indicated that the CEUS LI-RADS algorithm resulted in substantial consistency between the inexperienced radiologists. For the diagnosis of hepatocellular carcinoma, the sensitivity, specificity, positive predictive value, and negative predictive value were improved significantly in inexperienced radiologists using the CEUS LI-RADS compared to conventional methods. The overall diagnostic accuracy of the experienced radiologists was almost equal to that of CEUS LI-RADS categories assigned by the inexperienced radiologists.
The CEUS LI-RADS algorithm can not only obtain substantial consistency among inexperienced radiologists but also have excellent diagnostic efficacy in the differentiation of benign from malignant FLLs compared to conventional methods. As a comprehensive algorithm, the CEUS LI-RADS can act as a guide for trainees in learning how to diagnose FLLs.
评估对比增强超声（CEUS）肝脏影像报告和数据系统（LI-RADS）在经验不足的放射科医生诊断肝脏局灶性病变（FLLs）中的实用性。
回顾性收集258例有肝细胞癌风险且接受CEUS检查患者的图像和临床数据。两名经过培训但经验不足的放射科医生和两名经验丰富的放射科医生对所有CEUS图像片段进行了回顾。每位经验不足的放射科医生为每个观察结果指定一个CEUS LI-RADS类别，并独立标记为良性或恶性。每位经验丰富的放射科医生使用传统诊断方法独立标记每个病变为恶性或良性。通过κ检验分析CEUS LI-RADS的观察者间一致性。LI-RADS类别和传统诊断的总体诊断准确性用灵敏度、特异度、阳性预测值和阴性预测值来描述。所有检验结果在P < 0.05时被认为具有统计学意义。
κ值为0.774表明CEUS LI-RADS算法在经验不足的放射科医生之间产生了高度一致性。对于肝细胞癌的诊断，与传统方法相比，使用CEUS LI-RADS的经验不足的放射科医生的灵敏度、特异度、阳性预测值和阴性预测值有显著提高。经验丰富的放射科医生的总体诊断准确性几乎与经验不足的放射科医生指定的CEUS LI-RADS类别相同。
CEUS LI-RADS算法不仅能在经验不足的放射科医生之间获得高度一致性，而且与传统方法相比，在鉴别FLLs的良恶性方面具有出色的诊断效能。作为一种综合算法，CEUS LI-RADS可作为培训学员学习如何诊断FLLs的指南。
