Author: Lizbeth Ayala-Dominguez, Aaron L Carrel, Diego Hernando, Gabrielle L Hoffmann, Amber Possell, Ivan M. Rosado-Mendez, Jiayi Tang, Andrew Wentland 👨🔬
Affiliation: Department of Pediatrics, University of Wisconsin-Madison, Department of Radiology, University of Wisconsin-Madison, Department of Medical Physics, University of Wisconsin-Madison, Departments of Medical Physics and Radiology, University of Wisconsin-Madison 🌍
Purpose: Metabolic dysfunction-associated steatotic liver disease (MAFLD) is the leading cause of chronic liver disease in children. While biopsy remains the diagnostic standard, non-invasive imaging parameters such as magnetic resonance imaging-proton density fat fraction (MRI-PDFF) and ultrasound-derived fat fraction (UDFF) are emerging methods for MAFLD assessment. However, their clinical application in children is limited by challenges such as motion artifacts and a lack of standardized protocols. This study evaluates the inter- and intra-operator variability of UDFF in a pediatric cohort and its agreement with MRI-PDFF.
Methods: Fifteen subjects (ages 10-14) participated in this IRB-approved, prospective, cross-sectional study (ClinicalTrials.gov: NCT06354491). Subjects were grouped based on their body mass index (BMI) as high BMI (BMI>95th percentile, n=7) or healthy (5th≤BMI<85th percentile, n=8). All subjects underwent MRI-PDFF and UDFF on the same day. UDFF measurements were taken in intercostal (recommended for adult imaging) and subcostal views using a 9C2 transducer and a Siemens ACUSON Sequoia scanner. Inter-operator variability (2 sonographers) was evaluated in the healthy BMI group. Intra-operator variability was assessed in both groups with a test-retest approach. Linear multivariate analysis was used to examine the effects of B-mode frequency, view, sonographer, repetition, and BMI group on UDFF. Bland-Altman analysis was used to evaluate intra- and inter-operator agreement, and agreement between MRI-PDFF and UDFF.
Results: Significant differences were found between intercostal and subcostal UDFF (p<0.0001) and between BMI groups (p<0.0001). The mean bias between sonographers was 0.7% (95% limits of agreement (LOA), -2.7-4.0%), and between test-retest scans was -0.5% (95%LOA, -5.0-4.1%). Agreement with MRI-PDFF was better in the intercostal view (mean bias: 1.1%, 95%LOA: -3.3-5.4%) than in the subcostal view (mean bias: 5.4%, 95%LOA: 0.02-10.9%).
Conclusion: UDFF demonstrated inter- and intra-operator agreement in a pediatric cohort. Intercostal UDFF showed better agreement with MRI-PDFF, probably due to better beam orientation and reduced probe pressure.