Liver Fat Quantification: Agreement and Reproducibility of Ultrasound-Derived Fat Fraction Measurements Using 5C1 and DAX Probes in Subjects Stratified by BMI

Abstract

Purpose

This study aimed to assess the agreement between the 5C1 and deep abdominal transducer (DAX) ultrasound probes when measuring ultrasound-derived fat fraction (UDFF) across different body mass index (BMI) categories and to evaluate the intra- and inter-observer reproducibility of liver fat measurements.

Materials and Methods

In this prospective study 63 subjects (32 with BMI < 25 and 31 with BMI ≥ 25) underwent UDFF measurements. Two observers performed five measurements per subject using both probes (5C1 and DAX). Intra- and inter-observer reliability were assessed using intraclass correlation coefficients (ICC). Agreement between probes and observers was evaluated using Spearman correlation and Bland-Altman analysis, stratified by BMI subgroup.

Results

The intra- and inter-observer reliability of UDFF measurements were excellent, with no relevant systematic bias between observers or probes. The 5C1 and DAX probes showed strong correlations (r = 0.92 and 0.93). In normal-weight subjects, differences between probes were minimal, with limits of agreement (LoA) of approximately ±2%. In subjects with BMI ≥ 25, the mean differences between probes were low (up to 2.6%), but the limits of agreement were wider (–8% to +13%), comparable to the inter-observer variability.

Conclusion

UDFF measurements with both the 5C1 and DAX probes showed consistently high reliability across BMI categories. Although variability increased with a higher BMI and liver fat, no systematic bias was observed, and both probes performed comparably. These findings support their interchangeable use in clinical settings. Validation against a reference standard is warranted to assess accuracy in challenging patient populations.

Zusammenfassung

Ziel

Ziel der Studie war es, die Übereinstimmung zwischen den Ultraschallsonden 5C1 und DAX (deep abdominal transducer) bei der Messung der ultraschallbasierten Fettfraktion (UDFF) über verschiedene BMI-Kategorien hinweg zu untersuchen und die intra- und interindividuelle Reproduzierbarkeit der Leberfettmessung zu bewerten.

Methodik

Insgesamt wurden 63 Probanden untersucht (32 mit BMI < 25, 31 mit BMI ≥ 25). Zwei Untersucher führten jeweils fünf Messungen pro Proband mit beiden Schallköpfen (5C1 und DAX) durch. Die Reproduzierbarkeit wurde mittels Intraklassen-Korrelation (ICC) bewertet. Die Übereinstimmung zwischen Sonden und Untersuchern wurde durch Spearman-Korrelation und Bland-Altman-Analyse untersucht, jeweils getrennt nach BMI-Gruppen.

Ergebnisse

Die Intraklassen-Korrelationskoeffizienten für Intra- und Inter-observer-Korrelationen waren exzellent, ohne relevante systematische Abweichungen zwischen Untersuchern oder Sonden. Auch zwischen den Sonden zeigten sich starke Korrelationen (r = 0.92 und 0.93). Bei normalgewichtigen Probanden waren die Unterschiede zwischen den Sonden minimal (LoA ~±2 %). Bei übergewichtigen Personen blieben die mittleren Unterschiede gering (bis zu 2.6 %), jedoch mit größeren Streubereichen (–8 % bis +13 %), vergleichbar zur Inter-observer-Variabilität.

Schlussfolgerung

UDFF-Messungen mit beiden Schallköpfen zeigten eine hohe Reproduzierbarkeit über alle BMI-Kategorien hinweg. Trotz erhöhter Variabilität bei höherem BMI und Leberfett bestand kein klinisch relevanter systematischer Unterschied. Beide Sonden lieferten vergleichbare Ergebnisse und erscheinen in der klinischen Praxis austauschbar. Eine Validierung gegenüber einem Referenzstandard bleibt jedoch notwendig.

Keywords

ultrasound-derived fat fraction - liver fat quantification - metabolic dysfunction-associated steatotic liver disease

Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD), previously known as non-alcoholic fatty liver disease (NAFLD), represents a significant and growing global health concern [1]. MASLD has emerged as the most common liver disease worldwide, with its prevalence growing in conjunction with the increasing rates of obesity and type 2 diabetes [2] [3]. The rising prevalence of MASLD and its potential to progress to complications like steatohepatitis, cirrhosis, and hepatocellular carcinoma highlight the need for screening. Developing simple and accessible tools is essential for early detection and follow-up monitoring of MASLD [4]. Liver biopsy is the current diagnostic gold standard for precise histological quantification of fat content in the liver [5]. Magnetic resonance imaging proton density fat fraction (MRI-PDFF) is the gold standard for noninvasive fat quantification. However, its use in clinical practice is often limited by high costs, lengthy scan times, and restricted availability [6] [7]. Ultrasound-based techniques provide a promising alternative for noninvasive assessment of hepatic steatosis, combining affordability, portability, and user-friendly application. Despite their clinical utility, both the Hepatorenal Index (HRI) and the Hamaguchi scoring system are affected by subjectivity, image quality, and inter-observer variability, which can limit their reproducibility and diagnostic accuracy [8] [9]. Ultrasound-derived fat fraction (UDFF) combines attenuation with the backscatter coefficient (BSC) for image-based fat quantification [10]. UDFF is integrated into a conventional high-end ultrasound system. For the assessment of obese patients, a specialized XL probe, the deep abdominal transducer (DAX) with a frequency range of 1.0–3.5 MHz, was developed. While most UDFF validation studies have used the DAX probe, the UDFF software can also be operated with the standard 5C1 convex probe, which is routinely used for abdominal ultrasound. Whether both probes provide comparable UDFF measurements has not yet been investigated, although this question is highly relevant for clinical workflow and implementation. Studies have found a strong correlation between UDFF and MRI-PDFF, with UDFF demonstrating high reproducibility [11] [12] [13] [14] [15] [16]. Given that the majority of MASLD patients have a BMI ≥ 25 kg/m², it is important that liver fat quantification methods perform reliably in this population [17]. Recent studies show that UDFF performs well in obese MASLD patients, with strong agreement with histology and high diagnostic accuracy across steatosis grades [18] [19]. However, the influence of BMI, subcutaneous fat, and measurement depth on UDFF performance remains under investigation, and a standardized protocol is essential [20] [21]. The choice of transducer may also affect measurement quality, particularly in patients with increased body fat. To our knowledge, a direct comparison of different ultrasound probes for UDFF acquisition has not yet been conducted.

This study aimed to evaluate the intra- and inter-observer reproducibility of ultrasound-derived fat fraction (UDFF) measurements and to assess the agreement between the 5C1 and DAX probes across different BMI categories.

Materials and Methods

Study Design and Observers

The single-center, cross-sectional study was conducted at a hospital. A total of 63 subjects were prospectively included: 32 normal-weight individuals (BMI < 25 kg/m²) with no history of liver disease, in particular no MASLD, recruited voluntarily from the faculty, and 31 overweight or obese patients (BMI ≥ 25 kg/m²), recruited from a specialized clinic for metabolic and nutritional disorders, such as MASLD. In the elevated BMI group, 7 patients were overweight (BMI 25–30) and 24 obese (BMI > 30), including 5 individuals with morbid obesity. The BMI distribution of the elevated BMI group is illustrated in [Fig. 3] for transparency. All measurements were performed between April and December 2024. While no formal power analysis was conducted, the sample size was determined based on feasibility. All individuals provided oral informed consent for participation in the study.  The study was approved by the local ethics committee. BMI, sex, age, liver disease etiology, skin-to-liver capsule distance (SCD), and steatosis on B-mode ultrasound were recorded.

Ultrasound Protocol and Equipment

Ultrasound examinations were conducted using the Siemens ACUSON Sequoia system (Siemens Healthineers Ultrasound, Issaquah, Washington, USA). After at least 4 hours of fasting, each subject was examined sequentially by two observers in a random order. Observer 1 (O1) had 5 years of ultrasound experience, while observer 2 (O2) had 25 years of experience. Both examiners had at least three months of UDFF measurement training prior to the study.

The study workflow is presented in [Fig. 1]. Before UDFF measurements, all subjects underwent conventional B-mode liver ultrasound using the 5C1 probe (1.0–5.7 MHz), a standard abdominal convex transducer. Throughout the examination, subjects were positioned in a supine position with their right arm raised behind their head to ensure full exposure of the intercostal space. The SCD was measured, and the presence of steatosis was determined using Hamaguchi's scoring system [9]. Afterwards, UDFF was performed using the 5C1 and DAX probe in a random order. For UDFF, 5 measurements were acquired and the median was calculated. The measurements were performed in the right liver lobe. The transducer was placed in a right intercostal approach in posterior axillary line 90 degrees to the liver capsule. Special care was taken to minimize motion artifacts during the procedure.

Fig. 1 Study workflow. Classification of subjects by BMI, B-mode ultrasound assessment, UDFF measurements with DAX and 5C1 probes, and 1×15 Auto pSWE acquisition with DAX. Auto pSWE: automated point shear wave elastography; DAX: deep abdominal transducer; UDFF: ultrasound-derived fat fraction; SCD: skin-to-liver capsule distance.

For UDFF, the dimensions of the region of interest (ROI) are fixed for DAX and for 5C1 based on the manufacturer’s specifications. Specific quality criteria were applied: The ROI was carefully selected to avoid artifacts such as vessels or rib shadows. To facilitate accurate positioning of the ROI (1.5 cm below the liver capsule), a marking bar integrated into the software was aligned with the echogenic liver capsule ([Fig. 2]). UDFF was measured at a frequency of 3 MHz as predefined by the manufacturer. Measurements were conducted during resting in an end-expiratory position for 1–2 seconds without requiring forced breath-holding. During the fifth UDFF measurement with the DAX probe, an additional automated 1×15 point shear wave elastography (Auto pSWE) measurement was performed simultaneously. If UDFF measurements were invalid, additional measurements were taken until five valid readings were obtained. The medians and IQRs of five valid measurements were calculated. The threshold for defining steatosis in the reviewed papers was typically between 4.5% and 7.6%. Thus, any value above 6% was considered to be steatosis in this study [12] [14] [16] [17] [22] [23].

Fig. 2 UDFF ultrasound images in the right liver lobe obtained with DAX (upper panel) and 5C1 (lower panel). DAX: deep abdominal transducer; UDFF: ultrasound-derived fat fraction.

Statistical Analysis

Descriptive statistics were generated to summarize the demographic characteristics of study subjects, including age, sex, BMI, SCD, steatosis in B-mode and Auto pSWE. Quantitative variables were reported as the mean and standard deviation, while qualitative variables were presented as numbers of cases. The intraclass correlation coefficient (ICC) was calculated to assess intra- and inter-observer reliability. Intra-observer reliability was based on five repeated measurements per observer, using both ultrasound probes separately, while inter-observer reliability compared measurements for each probe between the observers. A two-way random effects model (ICC (2.1)) for absolute agreement was used to account for variability between observers and probes. An ICC value above 0.90 indicates very high reliability, with excellent consistency between measurements. An ICC between 0.75 and 0.90 is considered high, showing good reliability. Values between 0.50 and 0.75 reflect moderate reliability, while ICC values below 0.50 suggest low reliability, meaning the measurements may be inconsistent. ICCs were reported along with their 95% confidence intervals (CI). Spearman’s rank correlation was used to assess the association between UDFF values obtained with the DAX and 5C1 probes; correlation coefficients and p-values were reported. Bland-Altman plots were used to evaluate mean differences and 95% limits of agreement (LoA) between observers and between probes, stratified by BMI subgroups (<25 vs. ≥25 kg/m²). Additionally, a descriptive plot illustrated within-subject variation across the four measurement conditions (DAX/5C1 × O1/O2), with values plotted against each subject’s mean UDFF. All statistical analyses were performed using R.

Results

Subject Characteristics

The study cohort included 63 subjects categorized into a normal-weight group (BMI < 25 kg/m², n=32) and an overweight/obese group (BMI ≥ 25 kg/m², n=31). Baseline characteristics are presented in [Table 1]. The normal-weight group included 16 women, while the BMI ≥ 25 kg/m² group consisted of 20 women. Participants in the normal-weight group had a mean BMI of 22.5 ± 2.0 kg/m² and an average age of 35.5 ± 10.3 years, while those in the BMI ≥ 25 kg/m² group had a mean BMI of 34.7 ± 5.3 kg/m² and were older, with a mean age of 47 ± 14.1 years. Within the BMI ≥ 25 kg/m² group, 7 subjects had BMI values between 25 and 30 kg/m², while the majority (24 subjects) had BMI values above 30 kg/m². As shown in [Fig. 3], SCD increased with BMI, reflecting greater subcutaneous tissue thickness in individuals with a higher BMI.

Table 1 Subject characteristics.

	BMI < 25 (n=32)	BMI ≥ 25 (n=31)
Summary of subject characteristics by BMI subgroup (BMI < 25 and BMI ≥ 25). Values are presented as mean ± standard deviation or number of cases. Variables include sex, age, BMI, SCD, presence of steatosis on B‑mode ultrasound, and elastography results (Auto-pSWE, m/s). Auto pSWE: automated point shear wave elastography; SCD: skin-to-liver capsule distance.
Sex, women/men	16/16	20/11
Age, years	35.5 ± 10.3	47 ± 14.1
BMI, kg/m2	22.5 ± 2.0	34.7 ± 5.3
SCD, cm	15.6 ± 2.8	26.5 ± 5.8
Steatosis B-mode Yes/No	0/32	26/5
Auto pSWE, m/s	1.16 ± 0.15	1.04 ± 0.13

Fig. 3 Scatter plot of BMI versus SCD for individual participants. Vertical dashed lines mark BMI thresholds of 25 and 30 kg/m². SCD: skin-to-liver capsule distance.

The SCD was 15.6 ± 2.8 mm in the normal-weight group and 26.5 ± 5.8 mm in the BMI ≥ 25 kg/m² group. B-mode ultrasound detected steatosis exclusively in the BMI ≥ 25 kg/m² group (26 of 31 subjects). Elastography values (Auto pSWE) were below 1.3 m/s in all subjects, with mean values of 1.16 ± 0.15 m/s in the normal-weight group and 1.04 ± 0.13 m/s in the BMI ≥ 25 kg/m² group, indicating the absence of liver fibrosis.

Ultrasound-Based Measurements

[Table 2] presents the mean ± SD of UDFF for the BMI subgroups, measured with the 5C1 and DAX probes by O1 and O2. The ICC for intra-observer reliability ([Table 3]) was excellent in the overall cohort (0.988 to 0.994).

Table 2 Mean ± standard deviation of UDFF medians.

	BMI < 25 (n=32)	BMI ≥ 25 (n=31)
Median UDFF (%) values obtained by O1 and O2 using DAX and 5C1 probes in both BMI subgroups (BMI < 25 and BMI ≥ 25). Data are presented as mean ± standard deviation of per‑subject medians. O1: observer 1; O2: observer 2; UDFF: ultrasound-derived fat fraction
UDFF DAX (Median) O1, %	3.75 ± 0.84	16.77 ± 8.69
UDFF DAX (Median) O2, %	3.84 ± 0.85	17.65 ± 9.06
UDFF 5C1 (Median) O1, %	3.78 ± 1.07	14.16 ± 6.66
UDFF 5C1 (Median) O2, %	3.72 ± 0.89	16.26 ± 8.55

Table 3 ICC for intra-observer reliability.

	All (n=63)
Intra‑observer reliability of UDFF measurements for each observer (O1, O2) and probe (DAX, 5C1). ICCs with 95% CI are reported for the entire study cohort (all). ICC: intraclass correlation coefficient; O1: observer 1; O2: observer 2; UDFF: ultrasound-derived fat fraction
UDFF DAX O1	0.994 (0.992, 0.996)
UDFF DAX O2	0.988 (0.983, 0.992)
UDFF 5C1 O1	0.990 (0.986, 0.993)
UDFF 5C1 O2	0.989 (0.984, 0.993)

Inter-observer Reliability

Inter-observer reliability ([Table 4]) was evaluated between O1 and O2 for both probes. The DAX probe showed an ICC of 0.963 (95% CI: 0.939–0.977), while the 5C1 probe had an ICC of 0.905 (95% CI: 0.843–0.943), indicating excellent agreement. To further explore the inter-observer agreement and potential measurement differences, Bland-Altman analyses (Supplementary Figure 1) were conducted separately for BMI subgroups (normal weight and BMI ≥ 25). Among normal-weight individuals, the mean differences between observers were negligible (DAX: –0.10%, 5C1: 0.03%) with narrow LoA (approximately ±1.2% to ±2.3%), reflecting very consistent inter-observer measurements. In participants with BMI ≥ 25, the average inter-observer bias remained low (DAX: –0.87%, 5C1: –2.10%). However, the limits of agreement were substantially wider, particularly for the DAX probe (LoA = –7.66% to 5.91%) and the 5C1 probe (LoA = –10.65% to 6.45%). This indicates greater variability in inter-observer measurements in overweight and obese individuals, with higher liver fat, although no clinically relevant systematic bias between observers was detected.

Table 4 ICC for inter-observer reliability.

	All (n=63)
Inter‑observer reliability of UDFF measurements between observers for the DAX and 5C1 probes. ICCs with 95% CI are reported for the entire study cohort (all). ICC: intraclass correlation coefficient; UDFF: ultrasound-derived fat fraction
UDFF DAX	0.963 (0.939–0.977)
UDFF 5C1	0.905 (0.843–0.943)

Correlation and agreement between DAX and 5C1

Spearman’s rank correlation ([Fig. 4]) revealed a strong positive correlation between the DAX and 5C1 probes for both observers. For O1, the correlation coefficient r was 0.93 (p < 0.001), and for O2, r was 0.92 (p < 0.001), reflecting a strong monotonic relationship between probes.

Fig. 4 Scatter plots illustrating the relationship between UDFF measurements obtained with the DAX and 5C1 probes. (A) O1 (DAXO1 vs. 5C1O1); (B) O2 (DAXO2 vs. 5C1O2). Each dot represents a paired measurement from one subject. The dashed identity line (x = y) indicates perfect agreement between probes. DAX: deep abdominal transducer; O1: observer 1; O2: observer 2; UDFF: ultrasound-derived fat fraction.

Bland-Altman analysis (Supplementary Figure 2) revealed no relevant systematic bias between DAX and 5C1 measurements in either BMI group, with mean differences close to zero for both observers in normal-weight subjects (–0.03% and 0.10%, LoA within ±1.9%). In subjects with BMI ≥ 25, slightly higher mean differences were observed (O1: 2.61%, O2: 1.39%), without a clinically relevant directional bias. However, the limits of agreement were notably wider in the BMI ≥ 25 kg/m² subgroup (O1: –7.42% to 12.64%; O2: –7.83% to 10.60%), indicating increased variability between probes in this group. In an exploratory analysis restricted to subjects with BMI > 30 kg/m² (n = 24), the mean difference between DAX and 5C1 measurements was 3.25% for O1 (LoA: –7.80% to 14.30%) and 2.25% for O2 (LoA: –6.71% to 11.21%).

Variability in UDFF Measurements Across Devices and Observers

Supplementary Figure 3A displays individual UDFF values from all four measurement combinations of probe and observer (DAXO1, DAXO2, 5C1O1, 5C1O2), plotted against each subject’s mean UDFF. While measurements in subjects with low mean UDFF values are relatively consistent and closely clustered, variability increases noticeably with a higher mean UDFF. This indicates reduced within-subject agreement at higher liver fat levels, as reflected by increased measurement variability. Supplementary Figure 3B focuses on a subgroup of subjects with a BMI ≥ 25 and no steatosis detected on B-mode ultrasound. The dashed horizontal line marks the 6% UDFF threshold used in this study to define steatosis. While some measurements fall below this cut-off, some values also exceed it across different probe-observer combinations. This variability illustrates the diagnostic uncertainty in borderline cases, where minor differences between measurements can influence the classification of steatosis, despite negative findings on conventional ultrasound.

Discussion

This study aimed to evaluate the performance of two ultrasound probes, the 5C1 and the DAX, with respect to measuring UDFF across BMI categories. Given the growing prevalence of MASLD, reliable and accessible methods of liver fat quantification are essential [1] [2]. The DAX probe (1.0–3.5 MHz) is designed for deeper imaging in overweight and obese patients. Compared to the conventional 5C1 convex probe (63.3 × 18.2 mm), it has a wider contact area (57.7 × 30.2 mm) and can reach depths of up to 55 cm, supporting better lesion detection in deeper tissues [24] [25]. Since the DAX probe has been used in most UDFF validation studies but the UDFF software can also be operated with the standard 5C1 probe, the question of interchangeability is clinically relevant.

Intra- and inter-observer reliability of UDFF measurements was consistently high, in line with previous studies [11] [14]. Intra-observer ICCs ranged from 0.988 to 0.994; inter-observer ICCs were 0.963 (DAX) and 0.905 (5C1), indicating excellent agreement. The two probes showed strong correlation (r = 0.93 for O1 and 0.92 for O2), and measurements were highly consistent in normal-weight subjects. In individuals with a BMI ≥ 25 kg/m² and those with higher liver fat content, measurement variability increased. However, the degree of variability between probes was comparable to that between observers using the same probe, with wider LoA but no systematic bias in either case. In subjects with a BMI ≥ 25 kg/m², the mean differences remained small (ranging from 1.4–2.6%), while the LoA ranged from –7.8% to 12.6%, indicating greater spread but no clinically relevant offset. An exploratory Bland–Altman analysis restricted to subjects with a BMI > 30 kg/m² showed slightly larger mean differences but a comparable pattern of variability and no additional clinically relevant bias. This range is consistent with previous findings, suggesting that the observed variation is not probe-specific but reflects general challenges in liver fat assessment in overweight and obese populations [11]. Prior studies have shown that increased hepatic fat and anatomical factors can reduce reproducibility. UDFF is derived from both backscatter and attenuation coefficients, which are known to exhibit depth-dependent variability [20] [21]. Factors such as motion artifacts and greater subcutaneous fat – common in obesity – may also contribute to inconsistencies [26]. Despite standardized positioning and ROI placement, these issues remain challenging.

In five patients with a BMI ≥ 25 kg/m², no steatosis was detected on B-mode ultrasound. However, among the four UDFF measurements taken per patient, some individual values exceeded the 6% threshold. Measurements varied depending on the examiner and the transducer used, with some values above and others below the threshold. This suggests that artifacts – more likely in overweight and obese individuals – may influence UDFF results and lead to potential overestimation. At the same time, it cannot be ruled out that mild steatosis was present but not described on B-mode. These findings highlight the need for cautious interpretation of both methods in overweight and obese patients.

We cannot determine whether one probe outperforms the other, as no reference standard such as histology or MRI-based PDFF was available. This represents a key limitation, along with the relatively small sample size. Future studies specifically targeting patients with a BMI ≥ 40 kg/m² or class III obesity (morbid obesity) would be valuable to determine whether extreme obesity leads to greater divergence between probes. However, such cohorts are not commonly encountered in everyday care for patients with steatotic liver disease, whereas our study population reflects the clinical spectrum typically seen in both specialized metabolic clinics and general outpatient practice. Nonetheless, the study has important strengths, including inter-observer and inter-probe comparisons conducted by trained clinicians involved in patient care, reflecting real-world conditions rather than controlled research settings.

In conclusion, increased variability in overweight and obese subjects or those with higher liver fat appears to be inherent to the measurement process and should be considered in clinical interpretation. Despite this, UDFF remains a valuable tool. Based on current evidence, both probes can be used interchangeably, depending on availability. Further studies comparing these probes with histological or imaging-based gold standards are warranted.

Conclusion

This study confirms excellent intra- and inter-observer reliability for UDFF measurements using both the 5C1 and DAX probes across BMI categories. Measurement variability increased with a higher BMI and liver fat, highlighting the need for cautious interpretation in these patients. As no clinically relevant systematic differences were found between the probes, they appear interchangeable for UDFF acquisition in clinical practice, at least within the patient population examined in this study. However, while both probes showed comparable performance, validation against a reference standard is needed to determine which one more accurately reflects the true liver fat content.

Contributorsʼ Statement

Ricarda Lamprecht-Bailer: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Project administration, Software, Visualization, Writing – original draft. Sarah Fischer: Investigation, Writing – original draft. Sophie Haberkamp: Investigation, Writing – original draft. Jonas Schmid: Software, Validation, Visualization, Writing – original draft. Markus F Neurath: Resources, Validation, Writing – original draft. Sebastian Zundler: Methodology, Supervision, Writing – original draft. Daniel Klett: Conceptualization, Formal analysis, Methodology, Supervision, Validation, Visualization, Writing – original draft.

Conflict of Interest

The authors declare that they have no conflict of interest.

• References

• 1 Rinella ME, Lazarus JV, Ratziu V. et al. A multisociety Delphi consensus statement on new fatty liver disease nomenclature. Hepatology 2023; 78: 1966-1986
• 2 Wong VW, Ekstedt M, Wong GL. et al. Changing epidemiology, global trends and implications for outcomes of NAFLD. J Hepatol 2023; 79: 842-852
• 3 Younossi ZM, Golabi P, Paik JM. et al. The global epidemiology of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH): a systematic review. Hepatology 2023; 77: 1335-1347
• 4 Jou J, Choi SS, Diehl AM. Mechanisms of disease progression in nonalcoholic fatty liver disease. Semin Liver Dis 2008; 28: 370-379
• 5 Archer AJ, Belfield KJ, Orr JG. et al. EASL clinical practice guidelines: non-invasive liver tests for evaluation of liver disease severity and prognosis. Frontline Gastroenterol 2022; 13: 436-439
• 6 Caussy C, Reeder SB, Sirlin CB. et al. Noninvasive, Quantitative Assessment of Liver Fat by MRI-PDFF as an Endpoint in NASH Trials. Hepatology 2018; 68: 763-772
• 7 Stine JG, Munaganuru N, Barnard A. et al. Change in MRI-PDFF and Histologic Response in Patients With Nonalcoholic Steatohepatitis: A Systematic Review and Meta-Analysis. Clin Gastroenterol Hepatol 2021; 19: 2274-2283 e2275
• 8 Hajibonabi F, Riedesel EL, Taylor SD. et al. Ultrasound-estimated hepatorenal index: diagnostic performance and interobserver agreement for pediatric liver fat quantification. Pediatr Radiol 2024; 54: 1653-1660
• 9 Hamaguchi M, Kojima T, Itoh Y. et al. The severity of ultrasonographic findings in nonalcoholic fatty liver disease reflects the metabolic syndrome and visceral fat accumulation. Am J Gastroenterol 2007; 102: 2708-2715
• 10 Han A, Zhang YN, Boehringer AS. et al. Inter-platform reproducibility of ultrasonic attenuation and backscatter coefficients in assessing NAFLD. Eur Radiol 2019; 29: 4699-4708
• 11 Kubale R, Schneider G, Lessenich CPN. et al. Ultrasound-Derived Fat Fraction for Hepatic Steatosis Assessment: Prospective Study of Agreement With MRI PDFF and Sources of Variability in a Heterogeneous Population. AJR Am J Roentgenol 2024; 222: e2330775
• 12 Qi R, Lu L, He T. et al. Comparing ultrasound-derived fat fraction and MRI-PDFF for quantifying hepatic steatosis: a real-world prospective study. Eur Radiol 2024;
• 13 Tavaglione F, Flagiello V, Terracciani F. et al. Non-invasive assessment of hepatic steatosis by ultrasound-derived fat fraction in individuals at high-risk for metabolic dysfunction-associated steatotic liver disease. Diabetes Metab Res Rev 2024; 40: e3787
• 14 Wang P, Song D, Han J. et al. Comparing Three Ultrasound-Based Techniques for Diagnosing and Grading Hepatic Steatosis in Metabolic Dysfunction-Associated Steatotic Liver Disease. Acad Radiol 2024;
• 15 Sun M, Zhong M, Luo F. et al. Noninvasive assessment of hepatic steatosis grades by ultrasound derived fat fraction in metabolic dysfunction associated steatotic liver disease. Sci Rep 2024; 14: 32044
• 16 Verdan S, Torri GB, Marcos VN. et al. Ultrasound-derived fat fraction for diagnosing hepatic steatosis: a systematic review and meta-analysis. Eur Radiol 2025;
• 17 Meng L, Yang H, Hu Y. et al. Evaluation of ultrasound derived fat fraction for metabolic associated fatty liver disease in obese patients with polycystic ovary syndrome. J Ultrasound 2025;
• 18 Dominik N, Nixdorf L, Schwarz M. et al. UDFF and Auto pSWE accurately assess liver steatosis and fibrosis risk in obese patients with MASLD. Ultraschall in Med 2025;
• 19 Kale AB, Farasat M, Pekindil G. et al. Evaluation of Hepatic Steatosis and Fibrosis in Steatotic Liver Disease Ultrasound-Derived Fat Fraction (UDFF) and Auto pSWE by Using Deep Abdominal Transducer (DAX) and Liver Biopsy Correlation. J Ultrasound Med 2025;
• 20 Ferraioli G, De Silvestri A, Torres G. et al. Ultrasound backscatter coefficient for fat quantification is affected by the measurement depth. Abdom Radiol (NY) 2024; 49: 2622-2628
• 21 Ferraioli G, Raimondi A, Maiocchi L. et al. Liver Fat Quantification With Ultrasound: Depth Dependence of Attenuation Coefficient. J Ultrasound Med 2023; 42: 2247-2255
• 22 Nakamura Y, Hirooka M, Koizumi Y. et al. Diagnostic accuracy of ultrasound-derived fat fraction for the detection and quantification of hepatic steatosis in patients with liver biopsy. J Med Ultrason (2001) 2024;
• 23 Huang Y, Li J, Liu C. et al. Noninvasive Quantification of Hepatic Steatosis Using Ultrasound-Derived Fat Fraction (CHESS2303): A Prospective Multicenter Study. MedComm (2020) 2025; 6: e70123
• 24 Hirooka M, Koizumi Y, Nakamura Y. et al. Correction to: Deep attenuation transducer to measure liver stiffness in obese patients with liver disease. J Med Ultrason (2001) 2024; 51: 383
• 25 Siemens Healthineers. DAX: Deep abdominal transducer. Siemens Healthineers website. Accessed January 04, 2025 at: https://www.siemens-healthineers.com/en-us/ultrasound/dax
• 26 Song D, Wang P, Han J. et al. Reproducibility of ultrasound-derived fat fraction in measuring hepatic steatosis. Insights Imaging 2024; 15: 254

Publication History

Received: 27 October 2025

Accepted after revision: 05 January 2026

Accepted Manuscript online:
09 January 2026

Article published online:
04 February 2026