Ultrasound Derived Fat Fraction (UDFF): An US Tool for Non-Invasive Diagnosis and Quantification of Hepatic Steatosis

The rising prevalence of metabolic dysfunction-associated steatohepatitis (MASH) and the arrival of medical treatments have highlighted the need for accurate, accessible, and cost-effective methods for hepatic steatosis quantification. This study evaluates the performance of ultrasound-derived fat fraction (UDFF) for diagnosing and quantifying hepatic steatosis using MRI-derived proton density fat fraction (MRI-PDFF) and liver biopsy as reference standards. UDFF was also compared with the controlled attenuation parameter (CAP).

A total of 114 patients with chronic liver disease were prospectively included. All patients underwent UDFF and MRI-PDFF; 78 patients also had liver biopsy. CAP measurements were available for 104 patients. The correlation of UDFF with MR-PDFF and the area under the receiver operating characteristic curve (AUC) of UDFF for steatosis were calculated and compared with CAP. Factors influencing UDFF measurements were evaluated through multivariate analysis.

UDFF demonstrated a moderate to strong correlation with MRI-PDFF (r = 0.60 [0.46; 0.71], p < 0.001) and a moderate correlation with liver biopsy (η = 0.31). The AUCs for UDFF in diagnosing steatosis (≥ grade 1) were 0.79 (MRI-PDFF reference) and 0.82 (biopsy reference), outperforming CAP for moderate (grade 2) steatosis (p = 0.043). UDFF consistently overestimated steatosis by 5% compared to MRI-PDFF. Parietal thickness > 30 mm was the only factor influencing measurement accuracy. UDFF had no measurement failures, unlike CAP, highlighting its robustness.

Although the correlation between UDFF and MR-PDFF is moderate to strong, this prospective study does not provide results as good as those of the few previous studies evaluating the UDFF technique. These results highlight the importance of standardizing techniques and measurements in order to position US steatosis quantification tools in the management of patients with MASLD and MASH.