Ultrasound-Derived Fat Fraction for Hepatic Steatosis Assessment: Prospective Study of Agreement With MRI PDFF and Sources of Variability in a Heterogeneous Population
Reinhard Kubale,Guenther Schneider,Carl Paul Nicolas Lessenich,Arno Buecker,Sebastian Wassenberg,Gabriela Torres,Arati Gurung,Timothy Hall,Yassin Labyed
DOI: https://doi.org/10.2214/ajr.23.30775
2024-03-20
American Journal of Roentgenology
Abstract:<b>Background:</b> Metabolic dysfunction-associated steatotic liver disease is a growing global public health concern. Quantitative ultrasound measurements, such as ultrasound-derived fat fraction (UDFF), could provide noninvasive, cost-effective, and portable steatosis evaluation. <b>Objective:</b> To evaluate utility of UDFF for steatosis assessment using PDFF as reference in patients undergoing liver MRI for heterogeneous indications, and to assess UDFF variability. <b>Methods:</b> This prospective study included a primary analysis of 187 patients (mean age, 53.8 years; 112 men, 75 women) who underwent 3-T liver MRI for any clinical indication from December 2020 to July 2021. Patients underwent investigational proton-density fat fraction (PDFF) measurement using MR spectroscopy, and investigational ultrasound with UDFF calculation (mean of 5 intercostal measurements) within 1 hour after MRI. In a subanalysis, 21 of these patients underwent additional UDFF measurements 1, 3, and 5 hours after meal consumption. The study also included repeatability and reproducibility (R&R) analysis of 30 patients (mean age, 26.3 years; 11 men, 20 women) who underwent clinical abdominal ultrasound between November 2022 and January 2023; in these patients, three operators sequentially performed UDFF measurements. <b>Results:</b> In primary analysis, UDFF and PDFF showed intraclass correlation coefficient (ICC) of 0.79. In Bland-Altmananalysis, UDFF and PDFF showed mean difference of 1.5% (95% CI, 0.6-2.4%), with 95% limits of agreement from -11.0% to 14.0%. UDFF exhibited AUC for detecting PDFF at historic thresholds of ≥6.5%, ≥17.4%, and ≥22.1% of 0.90, 0.95, and 0.95, respectively. In subanalysis, mean UDFF was not significantly different across time points with respect to meal consumption (p=.21). In R&R analysis, ICC for intraoperator repeatability ranged from 0.98 to 0.99, and for interoperator reproducibility from 0.90 to 0.96. Visual assessment of patient-level data plots indicated increasing variability of mean UDFF measurements across operators, and of intercostal measurements within individual patients, with increasing steatosis. <b>Conclusion:</b> UDFF showed robust agreement with PDFF, diagnostic performance for steatosis grades, as well as intraoperator repeatability and interoperator reproducibility. Nonetheless, UDFF exhibited bias toward slightly larger values versus PDFF, intraoperator and interoperator variation increased with increasing steatosis. <b>Clinical Impact:</b> UDFF shows promise for steatosis assessment across diverse populations, although continued optimization remains warranted.
radiology, nuclear medicine & medical imaging