Related Experiment Video
Updated: Jul 27, 2026
Segmentation and Measurement of Fat Volumes in Murine Obesity Models Using X-ray Computed Tomography
Published on: April 4, 2012
Real-time ultrasound-derived fat fraction in pediatric population: feasibility validation with MR-PDFF
Max Zalcman1, Richard A Barth2, Erika Rubesova2
1Department of Radiology, Lucile Packard Children's Hospital, Stanford University, Palo Alto, CA, USA. max.zalcman@gmail.com.
Background:
Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease in children. To avoid limitations of liver biopsy and MRI, quantitative ultrasound has become a research focus. Ultrasound-derived fat fraction (UDFF) is based on a combination of backscatter coefficient and attenuation parameter.
Objective:
The objectives of the study were to determine (1) agreement between UDFF/MRI proton density fat fraction (MR-PDFF) and (2) whether BMI and age are predictive for UDFF.
Materials And Methods:
This cross-sectional prospective study included a convenience sample of 46 children referred for clinically indicated abdominal MRI. MR-PDFF and five acquisitions of UDFF were collected. Intraclass correlation coefficient (ICC) and Bland-Altman analysis were used to assess agreement between MR-PDFF and UDFF. Receiver operating characteristic curves were calculated for UDFF prediction of liver steatosis (MR-PDFF ≥ 6%). Multivariable regression was performed to assess BMI and age as predictors for UDFF.
Results:
Twenty-two participants were male, 24 were female, and the mean age was 14 ± 3 (range: 7-18) years. Thirty-six out of 46 participants had normal liver fat fraction <6%, and 10/46 had liver steatosis. UDFF was positively associated with MR-PDFF (ICC 0.92 (95% CI, 0.89-0.96). The mean bias between UDFF and MR-PDFF was 0.64% (95% LOA, -5.3-6.6%). AUROC of UDFF for steatosis was of 0.95 (95% CI, 0.89-0.99). UDFF cutoff of 6% had a sensitivity of 90% (95% CI, 55-99%) and a specificity of 94% (95% CI, 81-0.99%). BMI was an independent predictor of UDFF (correlation: 0.55 (95% CI, 0.35-0.95)).
Conclusions:
UDFF shows strong agreement with MR-PDFF in children. A UDFF cutoff of 6% provides good sensitivity and specificity for detection of MR-PDFF of ≥ 6%.
