Ultrasound-derived fat fraction for detection of hepatic steatosis and quantification of liver fat content

Abstract

Abstract Purpose To compare ultrasound (US) and US-derived fat fraction (UDFF) with magnetic resonance proton density fat fraction (MRI-PDFF) for the detection of hepatic steatosis and quantification of liver fat content. Materials and methods Between October and December 2022, 149 patients scheduled for an abdominal MRI agreed to participate in this study and underwent MRI-PDFF, US and UDFF. Inclusion criteria were: (a) no chronic liver disease or jaundice; (b) no MRI motion artifacts; (c) adequate liver examination at US. Exclusion criteria were: (a) alcohol abuse, chronic hepatitis, cirrhosis, or jaundice; (b) MRI artifacts or insufficient US examination. The median of 10 MRI-PDFF and UDFF measurements in the right hepatic lobe was analyzed. UDFF and MRI-PDFF were compared by Bland–Altman difference plot and Pearson’s test. Sensitivity, specificity, positive and negative predictive values, accuracy, and area under the receiver-operator curve (AUC-ROC) of US and UDFF were calculated using an MRI-PDFF cut-off value of 5%. p values ≤ 0.05 were statistically significant. Results 122 patients were included (61 men, mean age 60 years, standard deviation 15 years). The median MRI-PDFF value was 4.1% (interquartile range 2.9–6); 37.7% patients had a median MRI-PDFF value ≥ 5%. UDFF and MRI-PDFF had high agreement ( p = 0.11) and positive correlation ( ⍴ = 0.81, p < 0.001). UDFF had a higher diagnostic value than US for the detection of steatosis, with AUC-ROCs of 0.75 (95% CI 0.65, 0.84) and 0.53 (95% CI 0.42, 0.64), respectively. Conclusions UDFF reliably quantifies liver fat content and improves the diagnostic value of US for the detection of hepatic steatosis.