Impact of acquisition geometry on ultrasound-derived fat fraction measurements: implications for standardized acquisition and measurement interchangeability.

BACKGROUND: Ultrasound-derived fat fraction (UDFF) is an emerging quantitative imaging biomarker for noninvasive assessment of hepatic steatosis. However, the influence of acquisition geometry on UDFF measurements remains incompletely characterized. This study aimed to determine the effects of anatomical measurement window, subcapsular sampling depth, and skin-to-liver capsule distance (SCD) on UDFF measurements and measurement interchangeability. METHODS: In this prospective study, 238 volunteers underwent UDFF measurements at five predefined hepatic windows and four standardized subcapsular sampling depths (1.0, 1.5, 2.0, and 3.0 cm). The right anterior intercostal window with a subcapsular sampling depth of 1.5 cm served as the prespecified internal methodological reference configuration for within-case comparisons. Reference-based within-case differences and Bland-Altman analyses were used to assess window- and depth-dependent reference-relative differences and agreement. Linear mixed-effects models were applied to evaluate associations of acquisition geometry and SCD with systematic differences and absolute deviations. Intra- and inter-operator repeatability were assessed using intraclass correlation coefficients (ICCs). RESULTS: UDFF values differed significantly across anatomical windows at a standardized depth (P<0.001). Left-lobe measurements demonstrated substantial positive bias relative to the reference window (+4.96% to +6.05%) and wide limits of agreement, whereas smaller but statistically significant window-dependent biases persisted within the right lobe. Subcapsular sampling depth exerted a systematic effect on UDFF, with shallow measurements at 1.0 cm showing positive bias (+2.03%), whereas measurements at 2.0 and 3.0 cm demonstrated small negative biases relative to the 1.5-cm reference. Greater SCD was independently associated with larger absolute deviations (β =+1.05% per cm, P<0.001). Under the standardized reference configuration, UDFF demonstrated good repeatability, with both intra- and inter-operator ICCs of 0.83. CONCLUSIONS: UDFF measurements are strongly influenced by acquisition geometry, particularly anatomical window selection and subcapsular sampling depth. Reference-based analysis demonstrates limited individual-level interchangeability across hepatic windows in the studied cohort, underscoring the importance of standardized acquisition protocols to reduce acquisition-dependent variability and to support reliable and comparable quantitative UDFF measurements in longitudinal and multicenter settings.