Ultrasonographic morphological diagnosis of chronic liver disease: 2-dimensional shear wave elastography as an add-on test

• Published in: Ultrasonography (Seoul, Korea) Vol. 39; no. 3; pp. 272 - 280
• Main Authors: Cho, Young Seo, Jeong, Woo Kyoung, Kim, Yongsoo
• Format: Journal Article
• Language: English
• Published: Korea (South) Korean Society of Ultrasound in Medicine 01.07.2020; 대한초음파의학회
• Subjects: Abdomen; Biopsy; elastography; hepatic fibrosis; Hepatitis B; Liver cirrhosis; Liver diseases; Morphology; Original; Patients; Surveillance; ultrasonography; 방사선과학; United States > US; Hepatic fibrosis; Elastography; Liver cirrhosis; Ultrasonography
• ISSN: 2288-5919; 2288-5943
• DOI: 10.14366/usg.20009

This study investigated the impact of liver stiffness measurements (LSM) made using 2-dimensional (2D) shear wave elastography (SWE) on the diagnosis and grading of hepatic fibrosis and liver cirrhosis (LC) using grayscale ultrasonography (US). This retrospective study included 46 patients who underwent liver biopsy for chronic liver disease and 33 normal subjects with an aspartate aminotransferase-to-platelet ratio index <0.50 and a Forns score <4.21. Two abdominal radiologists reviewed randomized grayscale hepatic sonogram sets with and without LSM, separated by a 4-week interval. They graded the features of echogenicity, echotexture, surface nodularity, and edge blunting and classified patients by fibrosis grade. Interobserver agreement and correlations with the reference standards were compared before and after the reviewers were informed regarding LSM, and the impact of LSM on diagnostic performance was evaluated. The standard diagnoses were no or mild fibrosis (F0-1, n=39), moderate to advanced fibrosis (F2-3, n=23), or LC (n=17). The correlations between US and the diagnostic reference standard increased significantly with LSM incorporation (0.499 and 0.312 to 0.782 and 0.804, P<0.01 for both reviewers), as did interobserver agreement (0.318 to 0.753, P<0.01). The areas under the receiver operating characteristic curve (AUCs) for the diagnosis of significant fibrosis increased when LSM was included (0.682 and 0.591 to 0.855 and 0.907, P<0.01 for both reviewers), while the AUCs for the diagnosis of LC did not change significantly (0.891 and 0.783 to 0.904 and 0.900, P=0.849 and P=0.166). Incorporating LSM values obtained by 2D-SWE improved the diagnostic accuracy and interobserver agreement of grayscale US for hepatic fibrosis.