Discovery of ultrasound-derived fat fraction as a non-invasive tool for MASLD diagnosis
摘要
This study aimed to evaluate the diagnostic performance of ultrasound-derived fat fraction (UDFF) for metabolic dysfunction-associated steatotic liver disease (MASLD), by directly comparing it with the non-invasive gold standard MRI-PDFF and the established ultrasound method CAP.
MethodsThe diagnostic criteria for MASLD were used, we included 103 individuals with 53 MASLD patients and 49 healthy controls. All participants underwent liver MRI for MRI-PDFF quantification; ultrasound elastography for UDFF measurement; and vibration-controlled transient elastography for CAP assessment. Receiver operating characteristic (ROC) curves were generated to evaluate the performance of UDFF and CAP for predicting MRI-PDFF ≥ 5%, the imaging threshold defining hepatic steatosis in MASLD diagnosis. Finally, we analyzed the correlation between UDFF and serological indicators and the consistency between UDFF and MRI-PDFF.
ResultsAcross all 103 participants, the median UDFF value was 6.0%, median CAP was 246 dB/m, and median MRI-PDFF was 5.44%. Overall, UDFF was positively correlated with MRI-PDFF (R = 0.876; P < 0.001), and CAP was positively correlated with MRI-PDFF (R = 0.792, P < 0.001). For diagnosing 5% and above MRI-PDFF, UDFF was noted to have a mean AUC of 0.981. Additionally, CAP was noted to have an AUC of 0.932 in diagnosing MASLD. In addition, the Bland–Altman difference plots showed overall UDFF and MRI-PDFF, with a mean deviation of -0.2%. A linear regression model between the difference and the mean suggested proportional error.
ConclusionsUltrasound testing for UDFF provides a simple and noninvasive clinical tool to quantify the extent of hepatic steatosis. The diagnostic efficacy of UDFF is superior to CAP. Notably, while UDFF shows strong agreement with MRI-PDFF overall, Bland–Altman analysis revealed a proportional bias, with UDFF tending to underestimate MRI-PDFF at higher fat content levels. Despite this limitation, UDFF demonstrates potential as a noninvasive alternative to MRI-PDFF for MASLD diagnosis.