Department of Internal Medicine 1, University of Erlangen, Germany.
Ultraschall Med. 2018 Dec;39(6):667-674. doi: 10.1055/a-0612-7887. Epub 2018 Jun 7.
This pilot study aimed at assessing interobserver agreement with two contrast-enhanced ultrasound (CEUS) algorithms for the diagnosis of hepatocellular carcinoma (HCC) in high-risk patients.
Focal liver lesions in 55 high-risk patients were assessed independently by three blinded observers with two standardized CEUS algorithms: ESCULAP (Erlanger Synopsis of Contrast-Enhanced Ultrasound for Liver Lesion Assessment in Patients at risk) and ACR-CEUS-LI-RADSv.2016 (American College of Radiology CEUS-Liver Imaging Reporting and Data System). Lesions were categorized according to size and ultrasound contrast enhancement in the arterial, portal-venous and late phase. Interobserver agreement for assessment of enhancement pattern and categorization was compared between both CEUS algorithms. Additionally, diagnostic accuracy for the definitive diagnosis of HCC was compared. Histology and/or CE-MRI and follow-up served as reference standards.
55 patients were included in the study (male/female, 44/ 11; mean age: 65.9 years). 90.9 % had cirrhosis. Histological findings were available in 39/55 lesions (70.9 %). Reference standard of the 55 lesions revealed 48 HCCs, 2 intrahepatic cholangiocellular carcinomas (ICCs), and 5 non-HCC-non-ICC lesions. Interobserver agreement was moderate to substantial for arterial phase hyperenhancement (ĸ = 0.53 - 0.67), and fair to moderate for contrast washout in the portal-venous or late phase (ĸ = 0.33 - 0.53). Concerning the CEUS-based algorithms, the interreader agreement was substantial for the ESCULAP category (ĸ = 0.64 - 0.68) and fair for the CEUS-LI-RADS category (ĸ = 0.3 - 0.39). Disagreement between observers was mostly due to different perception of washout.
Interobserver agreement is better for ESCULAP than for CEUS-LI-RADS. This is mostly due to the fact that perception of contrast washout varies between different observers. However, interobserver agreement is good for arterial phase hyperenhancement, which is the key diagnostic feature for the diagnosis of HCC with CEUS in the cirrhotic liver.
本初步研究旨在评估两种对比增强超声（CEUS）算法在高危患者肝细胞癌（HCC）诊断中的观察者间一致性。
对 55 例高危患者的局灶性肝脏病变，由 3 名盲法观察者分别使用两种标准化的 CEUS 算法进行评估：ESCAPUL（风险患者肝脏病变的对比增强超声综述）和 ACR-CEUS-LI-RADSv.2016（美国放射学院 CEUS-肝脏成像报告和数据系统）。根据动脉期、门静脉期和晚期的大小和超声造影增强情况对病变进行分类。比较两种 CEUS 算法对增强模式和分类的观察者间一致性。此外，还比较了对 HCC 明确诊断的诊断准确性。组织学和/或 CE-MRI 和随访作为参考标准。
本研究共纳入 55 例患者（男/女，44/11；平均年龄：65.9 岁）。90.9%有肝硬化。39/55 例病变有组织学发现（70.9%）。55 个病变的参考标准显示 48 个 HCC、2 个肝内胆管细胞癌（ICC）和 5 个非 HCC-非 ICC 病变。动脉期高增强的观察者间一致性为中度至高度（ĸ=0.53-0.67），门静脉期或晚期对比洗脱的观察者间一致性为适度至中度（ĸ=0.33-0.53）。对于基于 CEUS 的算法，ESCAPUL 类别观察者间一致性较高（ĸ=0.64-0.68），CEUS-LI-RADS 类别观察者间一致性一般（ĸ=0.3-0.39）。观察者之间的分歧主要是由于对洗脱的不同感知。
ESCAPUL 的观察者间一致性优于 CEUS-LI-RADS。这主要是因为不同观察者对对比洗脱的感知不同。然而，动脉期高增强的观察者间一致性较好，这是肝硬化患者使用 CEUS 诊断 HCC 的关键诊断特征。
