Interobserver Agreement Between Primary Sonographers and Secondary Overreaders for Screening and Surveillance Liver Ultrasounds Using Ultrasound Liver Imaging Reporting and Data System

Abstract

The authors aim to identify if primary sonographers and secondary reviewers, both radiologists and sonographers, are likely to assign the same Ultrasound Liver Imaging Reporting and Data System (US LI-RADS) scores for liver surveillance ultrasounds. Institutional review board approval was obtained. Sonographers were familiarized with US LI-RADS via radiologist-led lectures. Three sonographers prospectively scored 170 screening examinations using US LI-RADS recommendations. Scans were retrospectively rescored by a fourth sonographer and a radiologist, both of whom were blinded to the original scores. Results were analyzed with weighted and nonweighted Cohen kappa statistical analysis methods. There was near-perfect agreement between primary and secondary sonographers and primary sonographer and radiologist (kappa of 0.87 and 0.92, respectively) for US LI-RADS category (cat) scores. However, only substantial and moderate agreements were noted for visualization (vis) scores between primary and secondary sonographers and primary sonographer and radiologist (weighted kappa of 0.73 and 0.48, respectively). There was vis score disagreement between the primary sonographer and radiologist in 60 (35.3%) cases. In 35 (20%) cases, the radiologist assigned a lower/more conservative vis score. There was vis score disagreement between the primary and secondary reviewing sonographers in 30 (17.6%) cases. In 12 (7%) cases, the secondary sonographer assigned a more conservative vis score. Although a good degree of concordance was noted between the groups, radiologists will need to generate their own US LI-RADS scoring to accurately reflect their impression and appropriately steer management.

Quality Assessment of Ultrasound and Magnetic Resonance Imaging for Hepatocellular Carcinoma Surveillance: A Systematic Review and Meta-Analysis

“…Variations in body habitus also contribute to heterogeneity and could not be quantified due to a paucity of data in the included articles. Additionally, the semi-quantitative scoring of liver visualization using LI-RADS may suffer from inter-reader variability, while standardized may not be entirely reliable between studies [49, 50]. There were limited studies that provided data for visualization quality for aMRI, only one for complete MRI, and no studies for CT.…”