Pyo, Cho, Choi, Jee, Yun, Hwang, Park, Kim, Kang, Kang, and Byun: Diagnostic performance of quantitative ultrasonography for hepatic steatosis in a health screening program: a prospective single-center study

Abstract

Purpose

This study compared the diagnostic performance of quantitative ultrasonography (QUS) with that of conventional ultrasonography (US) in assessing hepatic steatosis among individuals undergoing health screening using magnetic resonance imaging–derived proton density fat fraction (MRI-PDFF) as the reference standard.

Methods

This single-center prospective study enrolled 427 participants who underwent abdominal MRI and US. Measurements included the attenuation coefficient in tissue attenuation imaging (TAI) and the scatter-distribution coefficient in tissue scatter-distribution imaging (TSI). The correlation between QUS and MRI-PDFF was evaluated. The diagnostic capabilities of QUS, conventional B-mode US, and their combined models for detecting hepatic fat content of ≥5% (MRI-PDFF ≥5%) and ≥10% (MRI-PDFF ≥10%) were compared by analyzing the areas under the receiver operating characteristic curves. Additionally, clinical risk factors influencing the diagnostic performance of QUS were identified using multivariate linear regression analyses.

Results

TAI and TSI were strongly correlated with MRI-PDFF (r=0.759 and r=0.802, respectively; both P<0.001) and demonstrated good diagnostic performance in detecting and grading hepatic steatosis. The combination of QUS and B-mode US resulted in the highest areas under the ROC curve (AUCs) (0.947 and 0.975 for detecting hepatic fat content of ≥5% and ≥10%, respectively; both P<0.05), compared to TAI, TSI, or B-mode US alone (AUCs: 0.887, 0.910, 0.878 for ≥5% and 0.951, 0.922, 0.875 for ≥10%, respectively). The independent determinants of QUS included skinliver capsule distance (β=7.134), hepatic fibrosis (β=4.808), alanine aminotransferase (β=0.202), triglyceride levels (β=0.027), and diabetes mellitus (β=3.710).

Conclusion

QUS is a useful and effective screening tool for detecting and grading hepatic steatosis during health checkups.

Graphic Abstract

Introduction

Hepatic steatosis is characterized by an abnormal or pathologically elevated accumulation of fat in the liver [1]. This condition can arise from various causes, including nonalcoholic fatty liver disease (NAFLD), alcoholism, chemotherapy, and metabolic, toxic, or infectious factors. Hepatic steatosis can lead to serious complications such as benign inflammatory hepatitis, fibrosis, cirrhosis, and end-stage liver disease [2].
Liver biopsy remains the definitive reference standard for detecting and grading steatosis [3]. However, its invasiveness and high sampling variability limit its use in routine, repeated assessments of steatosis [4,5]. Over the past two decades, noninvasive imaging has advanced significantly, offering a safe, rapid, cost-effective, and accurate method for volumetric assessment of hepatic steatosis that is increasingly preferred in clinical settings [6,7]. Magnetic resonance imaging–derived proton density fat fraction (MRI-PDFF) is a quantitative imaging biomarker that provides accurate, repeatable, and reproducible assessments of liver fat [7,8]. However, due to its high cost and limited availability, it is not routinely used for clinical screening [9]. Conventional B-mode ultrasonography (US) estimates the severity of steatosis based on a subjective analysis of sonographic patterns [10] and is widely used to evaluate hepatic steatosis because of its safety and accessibility. However, its subjective nature and limited accuracy in detecting mild steatosis are significant drawbacks [11]. Quantitative US (QUS), based on radiofrequency data analysis, was developed to address the limitations of B-mode US [12]. QUS determines tissue composition based on acoustic signal analysis, including tissue attenuation imaging (TAI) and tissue scatter distribution imaging (TSI). A significant correlation between the QUS parameters and MRI-PDFF has been reported [12-14], and QUS measurements have demonstrated high intra-observer and interobserver repeatability [14-16]. However, previous studies primarily included patients diagnosed with hepatic steatosis. The prevalence of hepatic steatosis, often identified during health examinations and associated with various pathological states, is on the rise [17,18]. Accurately diagnosing hepatic steatosis in asymptomatic patients during health checkups is crucial for future health management.
The aim of this study was to investigate the diagnostic performance of QUS parameters (TAI and TSI) compared with conventional B-mode US in identifying hepatic steatosis in individuals undergoing health screening using MRI-PDFF as the reference standard. Subgroup analyses were conducted according to body mass index (BMI) and skin-liver capsule distance, and the clinical risk factors associated with the diagnostic performance of QUS were evaluated.

Materials and Methods

Compliance with Ethical Standards

This prospective, single-center study was approved by the institutional review board of Samsung Medical Center (SMC 2022-05-054), and written informed consent was obtained from all participants.

Study Population

Between July 2022 and November 2022, 474 participants who met the eligibility criteria were enrolled. The inclusion criteria included: (1) being 19 years of age or older, (2) undergoing abdominal MRI and US for health screening at Samsung Medical Center, and (3) providing written informed consent for participation in the clinical study. The exclusion criteria were as follows: (1) contraindication for MRI; (2) difficulty in obtaining an appropriate liver US image; (3) diagnosis or history of hepatic carcinoma; (4) previous liver surgery, including liver transplantation; (5) suspected liver cirrhosis on US; (6) suspected ascites on US; (7) history of chronic hepatitis (hepatitis B antigen positivity, anti-hepatitis C virus positivity, or other clinically diagnosed or suspected chronic liver disease); (8) severe obesity, defined according to the World Health Organization Asian-Pacific criteria [19] as BMI >35 kg/m2 (n=0); (9) pregnancy (n=0); and (10) the investigators’ determination of unsuitability for participation in this clinical trial (n=47). After excluding individuals who withdrew consent (n=1), those with missing data on PDFF measurements due to miscommunication with the laboratory performing the MRI (n=21), those with missing QUS images (n=1), and those with unsatisfactory QUS images (n=24), a total of 427 participants were ultimately evaluated in this study (Fig. 1).

Clinical Data Acquisition

The participants underwent a thorough medical checkup and screening program that included a lifestyle behavior survey, medical and medication histories, physical examinations, and various diagnostic tests such as laboratory, radiologic, and endoscopic assessments. Hypertension was defined as systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or the use of antihypertensive medication. Diabetes was defined as hemoglobin A1c (HbA1c) ≥6.5% without a known history of diabetes or the use of diabetes medication. Data used for this study were extracted from the Clinical Data Warehouse Darwin-C of the Samsung Medical Center.

Image Acquisition

US imaging acquisition using B-mode and S-Shearwave Imaging

B-mode imaging was conducted by board-certified abdominal radiologists using either LOGIQ E9 or E10 (GE Healthcare, Milwaukee, WI, USA) or RS85 (V2.0, Samsung Medison Co., Ltd., Seoul, Korea) equipped with a 3-5 MHz curved array transducer. The radiologists were not aware of the results from other studies. Participants were required to fast for at least four hours prior to the examinations. Sonograms were taken while the participants lay in a supine position, with the right arm fully abducted. US images of the right hepatic lobe were captured during breath-holding, using consistent time-gain compensations and focus positions. During the B-mode US examination, hepatic steatosis was identified and visually scored according to Hamaguchi’s scoring system, which integrates four US findings. The grading was as follows: S0 for no steatosis, S1 for mild steatosis, S2 for moderate steatosis, and S3 for severe steatosis [20].
Liver stiffness was measured using S-Shearwave Imaging (Samsung Medison Co., Ltd.) by a US technician with over 30 years of experience, under the supervision of an expert radiologist. The technician used the corresponding B-mode wave image to identify a region of interest (ROI) beneath the liver capsule in each section, ensuring the exclusion of large vessels, focal liver lesions, and reverberation artifacts. S-Shearwave Imaging provided images that included both stiffness and reliability measurement index maps. Measurements were obtained using the weighted sum of the weight equation residual and the magnitude of the shear wave [21]. Ten consecutive measurements were taken from different shear wave image frames, and the final measurement was expressed as the average stiffness value for each section in kilopascals (kPa), using an auto-profile function as recommended by the vendor. To differentiate between the fibrosis stages of the meta-analysis of histological data in viral hepatitis (METAVIR), the cutoff values from a previously published study were used [22]: 0-5.83 kPa for F1 (normal), >5.83 kPa for ≥F2 (mild fibrosis), >7.55 kPa for ≥F3 (moderate fibrosis), and >9.58 kPa for F4 (cirrhosis). The METAVIR staging system, initially developed for assessing liver fibrosis in chronic liver diseases such as viral hepatitis, has also proven effective in recent studies for evaluating liver fibrosis in NAFLD [23,24].

QUS parameter measurement

The two QUS parameters, the attenuation coefficient (AC) in TAI and the scatter-distribution coefficient (SC) in TSI, were measured by a single ultrasound technician with over 30 years of experience using an ultrasound system (RS 85, V2.0, Samsung Medison Co. Ltd.) equipped with a CA1-7S convex probe. A second ultrasound technician, with over 20 years of experience, repeated the QUS measurements. These parameters were recorded on the same day as the B-mode ultrasound and S-Shearwave Imaging sessions. Based on radiofrequency data, color-coded maps of AC-TAI and SC-TSI were generated on the corresponding B-mode images. The ROI, which excluded large vessels, focal lesions, and reverberation artifacts beneath the liver capsule, was positioned on the TSI and TAI maps of the liver parenchyma. When large blood vessels were unavoidable, these areas were excluded from the calculations (Fig. 2). The TAI and TSI measurements were repeated five times for each participant, and the median of these five measurements was used as the representative value. An operator measured the distance from the skin to the liver capsule (cm) on the B-mode image.

MRI-PDFF

MRI was performed (3.0 T, Ingenia CX, Philips Healthcare, Best, Netherlands) on participants in the supine position, utilizing both the 18-channel phased-array surface and integrated spine matrix coils. This MRI was performed concurrently with ultrasound. Participants in this study underwent additional sequences specifically for PDFF measurement. Complex-based chemical shift-encoded water-fat reconstruction techniques were employed, along with three-dimensional (3D) gradient-recalled-echo images, to estimate MRI-PDFF. The detailed parameters for the PDFF are provided in Supplementary Table 1.
The PDFF was estimated for each participant by drawing circular ROIs approximately 2 cm in diameter on the liver's fat fraction map. These ROIs avoided vessels, focal lesions, and imaging artifacts. Since QUS was only performed on the right liver and there were challenges in measuring PDFF in participants with small hepatic volumes or artifacts in segment I and the left lobe, ROI placement was limited to four Couinaud segments of the right lobe and one segment of the left lobe. The PDFF values for each ROI location were as follows: PDFF 1 in segment V, PDFF 2 in segment VI, PDFF 3 in segment VII, PDFF 4 in segment VIII, and PDFF 5 in the left lobe of the liver. An ultrasound specialist with over 30 years of experience, supervised by an expert radiologist who was blinded to the QUS results, performed this classification. The average of the five ROIs served as the reference standard for hepatic fat content [25]. Hepatic steatosis was defined as an MRI-PDFF of ≥5%, in accordance with current clinical practice guidelines [13,26]. Moderate hepatic steatosis was defined as an MRI-PDFF of ≥10%, based on a previously reported cutoff value [13,27].

Statistical Analyses

Participant characteristics are presented as number (percentage) for categorical variables and either mean±standard deviation (SD) or median (interquartile range), depending on the nature of the variables. Five measurements of TAI and TSI were obtained for each participant, and the median of these measurements was used as the representative value. To evaluate the interobserver reproducibility of the QUS parameters, the intraclass correlation coefficient (ICC) was calculated using a two-way random effects model. The differences in median levels of each QUS parameter across hepatic steatosis grade groups were analyzed using the Kruskal-Wallis test, followed by Dunn’s post-hoc test with a Bonferroni correction. Spearman’s correlation coefficients were utilized to examine the relationship between the QUS parameters and MRI-PDFF. Receiver operating characteristic (ROC) curve analyses were performed to evaluate the diagnostic performance of the QUS parameters and the visual steatosis grades in detecting hepatic fat content of ≥5% (MRI-PDFF ≥5%) and ≥10% (MRI-PDFF ≥10%) [13,16,26,27]. For each ROC analysis, the area under the ROC curve (AUC), optimal cutoff values, and performance parameters, including sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), were calculated. The optimal cutoff value for each QUS parameter was identified using the Youden index [28]. The likelihood of the PDFF being ≥5% or ≥10% was calculated using a logistic regression model that incorporated TAI, TSI, and B-mode US (combined model 1: TAI and TSI; combined model 2: TAI, TSI, and visual grade on B-mode US). The AUC for the combined model was derived from these probability values. The full regression equation, estimated regression coefficients, standard errors, and p-values for the logistic regression models are provided in Supplementary Tables 2 and 3. Multicollinearity was assessed using the variance inflation factor test, which indicated no significant correlation among the variables. AUC comparisons within the same dataset were conducted using Delong's method, while AUC comparisons between different datasets utilized the chi-squared test. Both univariate and multivariate linear regression analyses were performed to identify significant factors influencing QUS parameters. All statistical analyses were carried out using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) and the R package version 4.2.1 (http://www.R-project.org). Statistical significance was established at P<0.05. The assumption of normality for continuous variables was verified using the Kolmogorov-Smirnov test.

Results

Participants’ Characteristics

In total, 427 participants (360 men, 67 women; mean age [±SD]: 55.4 [±7.1] years) were included in this study. The participant characteristics are presented in Table 1. The mean BMI was 25.0 kg/m2 (±2.9), and the mean skin-liver capsule thickness was 1.9 cm (±0.3). The prevalence of diabetes was 18.7%. The proportion of heavy drinkers was 5.9%. The mean liver enzyme and lipid profile marker levels were within normal ranges. The mean MRI-PDFF was 7.7%±6.5% (range, 0.6% to 34.5%), and 206 (48.2%), 104 (24.4%), and 117 (27.4%) participants had an MRI-PDFF of <5%, ≥5% to <10%, and ≥10%, respectively. The prevalence of liver fibrosis (≥F2) was 10.0%.

Reproducibility of QUS Parameters

The ICC for TAI was 0.987 (95% confidence interval [CI], 0.978 to 0.991), and for TSI, it was 0.976 (95% CI, 0.971 to 0.980), indicating nearly perfect agreement between the two examiners.

Correlation between QUS Parameters and MRI-PDFF

TAI and TSI demonstrated a strong, significant, positive correlation with MRI-PDFF, with correlation coefficients of 0.759 and 0.802, respectively (95% CI, 0.715 to 0.796 and 0.764 to 0.833; P<0.001 for both). Fig. 3 illustrates the distribution of median TAI and TSI across various groups of hepatic fat content (<5%, ≥5% to <10%, and ≥10%) as assessed using MRI-PDFF. There were significant differences in TAI and TSI according to the grades of hepatic steatosis (P<0.001) (Table 2).
Considering the spatial heterogeneity of hepatic fat distribution, the correlation between QUS parameters and MRI-PDFF was analyzed across different PDFF regions (Supplementary Table 4). Among PDFFs 1-5, no specific region demonstrated a particularly high correlation with the median TAI or TSI values.

Diagnostic Performance of QUS Parameters in Determining Hepatic Steatosis

In detecting hepatic steatosis (MRI-PDFF ≥5%), the AUCs for TAI and TSI were 0.887 (95% CI, 0.857 to 0.918) and 0.910 (95% CI, 0.883 to 0.938), respectively, with cutoff values set at 0.708 for TAI and 98.79 for TSI (Table 3, Fig. 4). A TAI of ≥0.708 demonstrated a sensitivity of 75.6% and a specificity of 85.9%, while a TSI of ≥98.79 showed a sensitivity of 76.5% and a specificity of 93.2%. The AUC for detecting a hepatic fat content of ≥5% using visual grading with B-mode ultrasound was 0.878 (95% CI, 0.848 to 0.908), featuring a sensitivity of 97.8% and a specificity of 52.4%. When compared to B-mode ultrasound, the difference in AUC was statistically significant for TSI (P=0.043) but not for TAI (P=0.625).
In detecting hepatic fat content ≥10% (MRI-PDFF ≥10%), the AUCs for TAI and TSI were 0.951 (95% CI, 0.928 to 0.973) and 0.922 (95% CI, 0.897 to 0.946), respectively, with cutoff values set at 0.748 and 99.05. A TAI value >0.748 yielded a sensitivity of 88.9% and a specificity of 89.0%, while a TSI value >99.05 resulted in a sensitivity of 89.0% and a specificity of 77.7%. The AUC for visual grading using B-mode ultrasound to detect hepatic fat content ≥10% was 0.875 (95% CI, 0.848 to 0.902), with a sensitivity of 94.0% and a specificity of 75.8%. The corresponding PPV and NPV values are presented in Table 3. When detecting fat content ≥10%, both TAI and TSI demonstrated significantly higher AUCs than B-mode US (P<0.001 and P=0.002, respectively).
The AUC of combined model 1, which included TAI and TSI, was significantly higher (0.939 and 0.972 for detecting hepatic fat content of ≥5% and ≥10%, respectively) compared to B-mode US, TAI, or TSI alone (0.884, 0.887, and 0.910 for detecting hepatic fat content of ≥5%, respectively; 0.880, 0.951, and 0.922 for detecting hepatic fat content of ≥10%, respectively; P<0.05 for each). Additionally, the AUC of combined model 2, which incorporated TAI, TSI, and B-mode US, was significantly higher (0.947 and 0.975 for detecting hepatic fat content of ≥5% and ≥10%, respectively) compared to B-mode US, TAI, or TSI alone (P<0.05 for each). When comparing the two combined models, model 2 demonstrated a significantly higher AUC for detecting hepatic fat content of ≥5% than model 1 (0.947 vs. 0.939, P=0.041). However, the difference was not significant for detecting hepatic fat content of ≥10% (P=0.225) (Supplementary Table 5).

Subgroup Analyses

The subgroup analysis conducted according to BMI (<25 vs. ≥25 kg/m2) revealed a significant difference in the AUC of TAI (P=0.025); however, no significant differences were observed in the AUCs of other models. Specifically, the AUC for TAI was significantly higher in the BMI ≥25 kg/m2 group (0.917) than in the BMI <25 kg/m2 group (0.843). Subgroup analysis according to the median skin-liver capsule distance (<1.8 vs. ≥1.8 cm) indicated no significant differences in the AUCs across the models (Supplementary Table 6).

Factors Associated with QUS Parameters

Univariate linear regression analysis identified significant associations between TAI and several factors: BMI; skin-liver capsule distance; METAVIR score; levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyl transferase (GGT), triglycerides, high-density lipoprotein (HDL)–cholesterol, low-density lipoprotein-cholesterol, fasting glucose, and HbA1c; homeostatic model assessment for insulin resistance (HOMA-IR); hypertension; and diabetes. Furthermore, sex; BMI; skin-liver capsule distance; METAVIR score; levels of AST, ALT, GGT, triglycerides, HDL-cholesterol, fasting glucose, and HbA1c; HOMA-IR; hypertension; and diabetes were also significantly associated with TSI (Table 4).
Multivariate analysis revealed that the distance between the skin and liver capsule, METAVIR score, ALT level, triglyceride level, and diabetes were independent determinants in both TSI and TAI (Table 4).

Discussion

In the present study, the QUS parameters (TAI and TSI) demonstrated a strong correlation with MRI-PDFF and exhibited robust diagnostic capabilities for detecting and grading hepatic steatosis in asymptomatic individuals, using MRI-PDFF as the benchmark. When compared to conventional B-mode US, TSI significantly outperformed in detecting hepatic steatosis, with both TAI and TSI showing markedly better diagnostic performance for identifying hepatic fat content of ≥10%. Notably, the combination of TAI and TSI (combined model 1) resulted in a higher AUC than either parameter alone or B-mode US. Additionally, the integration of QUS with B-mode US (combined model 2) achieved the highest AUCs among all the models tested. Subgroup analysis revealed that TAI was the only parameter influenced by BMI. The QUS parameters were also affected by factors such as the skin-liver capsule distance, hepatic fibrosis, ALT levels, triglyceride levels, and diabetes. Moreover, the measurements showed excellent inter-examination repeatability.
To the authors’ knowledge, this is the first report on the diagnostic performance of QUS in asymptomatic individuals undergoing health screening. The diagnostic performance of QUS was evaluated and compared with that of conventional B-mode US and combined models. The diagnostic accuracy of conventional grading on B-mode US was acceptable and not significantly inferior to TAI. When TAI was combined with TSI, the diagnostic performance was superior compared to using TAI or TSI alone. Furthermore, the combination of QUS and B-mode US demonstrated the best diagnostic performance. Therefore, to achieve maximal diagnostic accuracy, both TAI and TSI should be performed. QUS can be easily implemented using the same conventional US machine; thus, employing both QUS and B-mode US can significantly enhance the diagnostic performance in detecting and grading hepatic steatosis. Previous studies on QUS focused on patients diagnosed with chronic liver disease, which may exhibit different characteristics from the population undergoing health screenings [14,16,29]. In addition, several previous studies have reported a correlation between QUS and MRI-PDFF in patients with NAFLD [14,16]. However, the current study included patients with alcohol-related fatty liver disease, who may be encountered at a health screening center.
Given that QUS parameters are influenced by tissue properties, the diagnostic performance of QUS in detecting hepatic steatosis may be affected by body fat [14,16,29,30]. Therefore, subgroup analyses were conducted to assess the diagnostic performance of QUS parameters based on BMI and skin-liver capsule distance. In the subgroup analysis by BMI, the AUC of TAI was significantly higher in the BMI ≥25 kg/m2 group than in the <25 kg/m2 group. Previous studies on overweight and obese patients have shown that the performance of the controlled attenuation parameter (CAP) is compromised at higher BMIs [31,32]. This phenomenon may be attributed to the fact that thicker subcutaneous tissue, including the subcutaneous fat layer, in patients with higher BMIs may lead to an overestimation of CAP. In addition, a prior meta-analysis on the accuracy of ultrasound AC for evaluating hepatic steatosis suggests that a higher BMI (≥25) may reduce the diagnostic accuracy of the AC [33]. However, there was significant heterogeneity in BMI across studies. The discrepancy between our findings and previous studies could be due to differences in study populations and BMI distributions. In Chan et al.'s study [31], 69 patients had a BMI <25 kg/m2 (among whom 11 patients [15.9%] had hepatic steatosis [≥5%]) and 92 patients had BMI ≥25 kg/m2 (among whom 87 [94.6%] had hepatic steatosis [≥5%]). In this study, 221 participants had a BMI <25 kg/m2, whereas 206 had a BMI ≥25 kg/m2, and the prevalence of hepatic steatosis (≥5%) in each BMI category was 89 (40.3%) and 132 (64.1%), respectively. Therefore, Chan’s study had a notably higher prevalence of hepatic steatosis among participants with a BMI ≥25 kg/m2 (94.6% vs. 64.1%). Some prior studies assessing hepatic steatosis by CAP reported a relationship with BMI only in the univariate analysis [34,35]. The skin-liver capsule distance may provide a better assessment of body fat status, which affects QUS parameters. The subgroup analysis based on skin-liver capsule distance revealed no significant differences in the AUC for each model. However, Shen et al. [29] reported that a skin capsular distance ≥2.5 cm might lead to an overestimation of steatosis. This discrepancy in outcomes could be due to different standards for skin capsule distance, and further research is needed to determine the impact of skin capsule distance on the diagnostic performance of QUS. The higher AUC of TAI in the BMI ≥25 kg/m2 group compared to the BMI <25 kg/m2 group might have been influenced by the ROI-based assessment of liver parenchyma. Visualizing a portion of the liver parenchyma free from large vessels may be more challenging in the BMI <25 kg/m2 group than in the BMI ≥25 kg/m2 group, as liver size is influenced by BMI [36].
Multivariate linear regression analysis showed that hepatic fibrosis was an independent predictor of TAI and TSI, with a positive relationship between them. The positive relationship between hepatic fibrosis and TAI in the present study is consistent with the results of previous studies [37,38]. Jeon et al. [39] reported that the correlation coefficient between TAI and MRI-PDFF was lower in patients with severe hepatic fibrosis. In principle, the attenuation of the US beam can be affected by hepatic fibrosis. However, the relationship between fibrosis and TSI varied in previous studies [12,14], which identified a negative association and explained that backscattered US signals from normal liver tissues predominantly consist of randomly distributed scatterers that behave according to a near-Rayleigh distribution. This inconsistency in results could stem from the limited number of patients with hepatic fibrosis included in these studies. Given that this study was conducted at a health screening center and excluded patients with suspected liver cirrhosis, the prevalence of hepatic fibrosis (F2 and F3) was only 10.0%, which is lower than in previous studies (11.7%-70.1%) [12,37]. Therefore, the clinical application of QUS in patients with hepatic fibrosis requires cautious interpretation. Further research is needed to explore the relationship between QUS and hepatic fibrosis more thoroughly.
The present study has several limitations. First, because this was a health screening-based study, the results may not be fully generalizable to the general population. Second, adjustment for possible confounders for hepatic steatosis, such as alcohol and medication use, was not performed. Further studies are required to clarify the association between alcohol consumption or the use of steatogenic medications and the diagnostic performance of QUS. Third, this study excluded individuals with severe obesity (BMI≥35), as the visual grade on B-mode US is only moderately effective in individuals with severe obesity [40]. The exclusion criteria also ruled out participants who exhibited signs of liver cirrhosis, referring to a previous study on QUS [12,37]. However, no participant was actually excluded due to severe obesity or liver cirrhosis. These exclusion criteria may potentially have led to selection bias. Further studies are required to clarify the association between severe obesity or liver cirrhosis and the diagnostic performance of QUS.
In conclusion, QUS is an accurate, noninvasive, and widely available diagnostic method that is suitable for detecting and grading hepatic steatosis in asymptomatic patients undergoing health screening. In this study, the combination of QUS and B-mode US achieved the highest AUC. Therefore, the clinical application of QUS in conjunction with conventional B-mode US should be considered for better assessment of hepatic steatosis in health screening programs.

Notes

Author Contributions

Conceptualization: Pyo JH, Cho SJ, Kang M, Byun YH. Data acquisition: Pyo JH, Cho SJ, Choi SC, Jee JH, Yun JY , Hwang JA, Park G, Kang M, Byun YH. Data analysis or interpretation: Pyo JH, Cho SJ, Jee JH, Hwang JA, Park G, Kim K, Kang W, Kang M, Byun YH. Drafting of the manuscript: Pyo JH, Cho SJ. Critical revision of the manuscript: Cho SJ, Choi SC, Jee JH, Yun JY Hwang JA, Park G, Kim K, Kang W, Kang M, Byun YH. Approval of the final version of the manuscript: all authors.

Conflict of Interest

The authors of Center for Health Promotion, Samsung Medical Center declare relationships with the following company: Samsung Medison Co., Ltd. (Seoul, Korea).

Acknowledgements

We thank Mal Shook Shim and Jung Min Lee for assistance with quantitative ultrasound examinations and magnetic resonance imaging–derived proton density fat fraction estimation.

Supplementary Material

Supplementary Table 1. MRI-PDFF pulse sequence parameters at 3.0 T (https://doi.org/10.14366/usg.24040).
Supplementary Table 1.
MRI-PDFF pulse sequence parameters at 3.0 T
usg-24040-Supplementary-Table-1,2.pdf
Supplementary Table 2.
Regression equation used for the combined models
usg-24040-Supplementary-Table-1,2.pdf
Supplementary Table 3.
Estimated regression coefficients, standard errors, and p-values in the logistic regression models
usg-24040-Supplementary-Table-3,4.pdf
Supplementary Table 4.
Correlation of each QUS parameter with MRI-PDFF according to PDFF regions
usg-24040-Supplementary-Table-3,4.pdf
Supplementary Table 5.
Comparison of diagnostic performance of QUS parameters and visual grade for the detection of hepatic steatosis
usg-24040-Supplementary-Table-5.pdf
Supplementary Table 6.
Subgroup analysis comparing the diagnostic performance of multiple models for determining hepatic fat content ≥5%
usg-24040-Supplementary-Table-6.pdf

References

1. Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander-Tetri BA, Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999;94:2467–2474.
crossref pmid
2. Vernon G, Baranova A, Younossi ZM. Systematic review: the epidemiology and natural history of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis in adults. Aliment Pharmacol Ther 2011;34:274–285.
crossref pmid
3. Chalasani N, Younossi Z, Lavine JE, Charlton M, Cusi K, Rinella M, et al. The diagnosis and management of nonalcoholic fatty liver disease: Practice guidance from the American Association for the Study of Liver Diseases. Hepatology 2018;67:328–357.
crossref pmid pdf
4. Tapper EB, Lok AS. Use of liver imaging and biopsy in clinical practice. N Engl J Med 2017;377:756–768.
crossref pmid
5. Ratziu V, Charlotte F, Heurtier A, Gombert S, Giral P, Bruckert E, et al. Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Gastroenterology 2005;128:1898–1906.
crossref pmid
6. Starekova J, Hernando D, Pickhardt PJ, Reeder SB. Quantification of liver fat content with CT and MRI: state of the art. Radiology 2021;301:250–262.
crossref
7. Caussy C, Reeder SB, Sirlin CB, Loomba R. Noninvasive, quantitative assessment of liver fat by MRI-PDFF as an endpoint in NASH trials. Hepatology 2018;68:763–772.
crossref pdf
8. Bonekamp S, Tang A, Mashhood A, Wolfson T, Changchien C, Middleton MS, et al. Spatial distribution of MRI-Determined hepatic proton density fat fraction in adults with nonalcoholic fatty liver disease. J Magn Reson Imaging 2014;39:1525–1532.
crossref pmid
9. Reeder SB, Hu HH, Sirlin CB. Proton density fat-fraction: a standardized MR-based biomarker of tissue fat concentration. J Magn Reson Imaging 2012;36:1011–1014.
crossref pmid pmc
10. European Association for the Study of the Liver (EASL); European Association for the Study of Diabetes (EASD); European Association for the Study of Obesity (EASO). EASL-EASD-EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease. J Hepatol 2016;64:1388–1402.
crossref
11. Lee DH. Quantitative assessment of fatty liver using ultrasound attenuation imaging. J Med Ultrason 2001;48:485–470.
crossref pdf
12. Jeon SK, Joo I, Kim SY, Jang JK, Park J, Park HS, et al. Quantitative ultrasound radiofrequency data analysis for the assessment of hepatic steatosis using the controlled attenuation parameter as a reference standard. Ultrasonography 2021;40:136–146.
crossref pmid pdf
13. Lin SC, Heba E, Wolfson T, Ang B, Gamst A, Han A, et al. Noninvasive diagnosis of nonalcoholic fatty liver disease and quantification of liver fat using a new quantitative ultrasound technique. Clin Gastroenterol Hepatol 2015;13:1337–1345.
crossref pmid
14. Ronaszeki AD, Budai BK, Csongrady B, Stollmayer R, Hagymasi K, Werling K, et al. Tissue attenuation imaging and tissue scatter imaging for quantitative ultrasound evaluation of hepatic steatosis. Medicine (Baltimore) 2022;101:e29708.
crossref pmid pmc
15. Ferraioli G, Raimondi A, De Silvestri A, Filice C, Barr RG. Toward acquisition protocol standardization for estimating liver fat content using ultrasound attenuation coefficient imaging. Ultrasonography 2023;42:446–456.
crossref pmid pmc pdf
16. Jeon SK, Lee JM, Joo I, Park SJ. Quantitative ultrasound radiofrequency data analysis for the assessment of hepatic steatosis in nonalcoholic fatty liver disease using magnetic resonance imaging proton density fat fraction as the reference standard. Korean J Radiol 2021;22:1077–1086.
crossref pmid pmc pdf
17. Stahl EP, Dhindsa DS, Lee SK, Sandesara PB, Chalasani NP, Sperling LS. Nonalcoholic fatty liver disease and the heart: JACC state-of-the-art review. J Am Coll Cardiol 2019;73:948–963.
pmid
18. Muzurovic E, Mikhailidis DP, Mantzoros C. Non-alcoholic fatty liver disease, insulin resistance, metabolic syndrome and their association with vascular risk. Metabolism 2021;119:154770.
crossref pmid
19. Haam JH, Kim BT, Kim EM, Kwon H, Kang JH, Park JH, et al. Diagnosis of obesity: 2022 update of clinical practice guidelines for obesity by the Korean Society for the Study of Obesity. J Obes Metab Syndr 2023;32:121–129.
crossref pmid pmc
20. Hamaguchi M, Kojima T, Itoh Y, Harano Y, Fujii K, Nakajima T, 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.
crossref pmid
21. Yoo J, Lee JM, Joo I, Yoon JH. Assessment of liver fibrosis using 2-dimensional shear wave elastography: a prospective study of intra- and inter-observer repeatability and comparison with point shear wave elastography. Ultrasonography 2020;39:52–59.
crossref pmid pdf
22. Yoo HW, Kim SG, Jang JY, Yoo JJ, Jeong SW, Kim YS, et al. Two-dimensional shear wave elastography for assessing liver fibrosis in patients with chronic liver disease: a prospective cohort study. Korean J Intern Med 2022;37:285–293.
crossref pmid pdf
23. Brattain LJ, Telfer BA, Dhyani M, Grajo JR, Samir AE. Objective liver fibrosis estimation from shear wave elastography. Annu Int Conf IEEE Eng Med Biol Soc 2018;2018:1–5.
crossref
24. Chowdhury AB, Mehta KJ. Liver biopsy for assessment of chronic liver diseases: a synopsis. Clin Exp Med 2023;23:273–285.
crossref pmid pdf
25. Tang A, Tan J, Sun M, Hamilton G, Bydder M, Wolfson T, et al. Nonalcoholic fatty liver disease: MR imaging of liver proton density fat fraction to assess hepatic steatosis. Radiology 2013;267:422–431.
crossref pmid pmc
26. Han A, Zhang YN, Boehringer AS, Montes V, Andre MP, Erdman JW Jr, et al. Assessment of hepatic steatosis in nonalcoholic fatty liver disease by using quantitative US. Radiology 2020;295:106–113.
crossref pmid
27. Caussy C, Alquiraish MH, Nguyen P, Hernandez C, Cepin S, Fortney LE, et al. Optimal threshold of controlled attenuation parameter with MRI-PDFF as the gold standard for the detection of hepatic steatosis. Hepatology 2018;67:1348–1359.
crossref pmid pdf
28. Youden WJ. Index for rating diagnostic tests. Cancer 1950;3:32–35.
crossref pmid
29. Shen F, Zheng RD, Shi JP, Mi YQ, Chen GF, Hu X, et al. Impact of skin capsular distance on the performance of controlled attenuation parameter in patients with chronic liver disease. Liver Int 2015;35:2392–2400.
crossref pmid pmc pdf
30. Park J, Lee JM, Lee G, Jeon SK, Joo I. Quantitative evaluation of hepatic steatosis using advanced imaging techniques: focusing on new quantitative ultrasound techniques. Korean J Radiol 2022;23:13–29.
crossref pmid pmc pdf
31. Chan WK, Nik Mustapha NR, Mahadeva S. Controlled attenuation parameter for the detection and quantification of hepatic steatosis in nonalcoholic fatty liver disease. J Gastroenterol Hepatol 2014;29:1470–1476.
crossref pmid pdf
32. Myers RP, Pollett A, Kirsch R, Pomier-Layrargues G, Beaton M, Levstik M, et al. Controlled Attenuation Parameter (CAP): a noninvasive method for the detection of hepatic steatosis based on transient elastography. Liver Int 2012;32:902–910.
crossref pmid
33. Jang JK, Choi SH, Lee JS, Kim SY, Lee SS, Kim KW. Accuracy of the ultrasound attenuation coefficient for the evaluation of hepatic steatosis: a systematic review and meta-analysis of prospective studies. Ultrasonography 2022;41:83–92.
crossref pmid pdf
34. Lee HW, Kim KJ, Jung KS, Chon YE, Huh JH, Park KH, et al. The relationship between visceral obesity and hepatic steatosis measured by controlled attenuation parameter. PLoS One 2017;12:e0187066.
crossref pmid pmc
35. Bae JS, Lee DH, Lee JY, Kim H, Yu SJ, Lee JH, et al. Assessment of hepatic steatosis by using attenuation imaging: a quantitative, easy-to-perform ultrasound technique. Eur Radiol 2019;29:6499–6507.
crossref pmid pdf
36. Kromrey ML, Ittermann T, vWahsen C, Plodeck V, Seppelt D, Hoffmann RT, et al. Reference values of liver volume in Caucasian population and factors influencing liver size. Eur J Radiol 2018;106:32–37.
crossref pmid
37. Jeon SK, Lee JM, Joo I, Yoon JH, Lee DH, Lee JY, et al. Prospective evaluation of hepatic steatosis using ultrasound attenuation imaging in patients with chronic liver disease with magnetic resonance imaging oroton density fat fraction as the reference standard. Ultrasound Med Biol 2019;45:1407–1416.
pmid
38. Lee DH, Lee JY, Lee KB, Han JK. Evaluation of hepatic steatosis by using acoustic structure quantification US in a rat model: comparison with pathologic examination and MR spectroscopy. Radiology 2017;285:445–453.
crossref pmid
39. Lin T, Ophir J, Potter G. Correlation of ultrasonic attenuation with pathologic fat and fibrosis in liver disease. Ultrasound Med Biol 1988;14:729–734.
crossref pmid
40. Wu J, You J, Yerian L, Shiba A, Schauer PR, Sessler DI. Prevalence of liver steatosis and fibrosis and the diagnostic accuracy of ultrasound in bariatric surgery patients. Obes Surg 2012;22:240–247.
crossref pmid pdf

Study flow chart.

MRI-PDFF, magnetic resonance imaging–derived proton density fat fraction.
usg-24040f1.jpg
Fig. 1.

Quantitative ultrasonographic (US) measurement of liver fat using attenuation coefficient (AC) in tissue attenuation imaging (TAI) and scatter-distribution coefficient (SC) in tissue scatter distribution imaging (TSI).

A-C. In a 54-year-old woman without hepatic steatosis on B-mode US, the AC in TAI (A) and SC in TSI (B) are 0.62 and 81.99, respectively. C. The magnetic resonance imaging–derived proton density fat fraction (MRI-PDFF) measured on the PDFF map as a reference standard is 0.98%. D-F. In a 60-year-old woman with mild hepatic steatosis on B-mode US, the AC in TAI (D) and SC in TSI (E) are 0.72 and 100.67, respectively. F. MRI-PDFF measured on the PDFF map as a reference standard is 8.42%. G-I. In a 56-year-old man with moderate hepatic steatosis on B-mode US, the AC in TAI (G) and SC in TSI (H) are 0.96 and 108.18, respectively. I. MRI-PDFF measured on the PDFF map as a reference standard is 21.06%
usg-24040f2.jpg
Fig. 2.

The distribution of tissue attenuation imaging (TAI) (A) and tissue scatter distribution imaging (TSI) (B) stratified by the hepatic fat content of magnetic resonance imaging–derived proton density fat fraction (MRI-PDFF).

usg-24040f3.jpg
Fig. 3.

Diagnostic performance for detecting hepatic steatosis (magnetic resonance imaging–derived proton density fat fraction [MRIPDFF] ≥5%) (A) and hepatic fat content ≥10% (MRI-PDFF ≥10%) (B).

TAI, tissue attenuation imaging; TSI, tissue scatter distribution imaging; US, ultrasonography.
usg-24040f4.jpg
Fig. 4.
usg-24040f5.jpg
Table 1.
Characteristics of the participants (n=427)
Variable Value
Age (year) 55.4±7.1
Sex
 Male 360 (84.3)
 Female 67 (15.7)
BMI (kg/m2) 25.0±2.9
Skin-liver capsule distance (cm) 1.9±0.3
Hypertension 178 (41.7)
Diabetes 80 (18.7)
Alcohol consumption
 None to moderate 402 (94.1)
 Heavy (≥30 g/day for men, ≥20 g/day for women) 25 (5.9)
AST (U/L) 25.1±9.8
ALT (U/L) 28.3±15.4
GGT (U/L) 38.9±28.4
Total cholesterol (mg/dL) 186.6±39.5
Triglyceride (mg/dL) 132.6±80.2
HDL-cholesterol (mg/dL) 56.5±14.5
LDL-cholesterol (mg/dL) 116.3±37.1
Fasting glucose (mg/dL) 102.7±17.1
HbA1c (%) 5.8±0.7
HOMA-IR 2.1±1.2
Visual grade of hepatic steatosis
 S0 (no steatosis) 113 (26.5)
 S1 (mild) 129 (30.2)
 S2 (moderate) 168 (39.3)
 S3 (severe) 17 (4.0)
Hepatic steatosis at MRI-PDFF
 MRI-PDFF <5% 206 (48.2)
 MRI-PDFF ≥5% to <10% 104 (24.4)
 MRI-PDFF ≥10% 117 (27.4)
Fibrosis grade according to shearwave elastography (METAVIR score)
 F1 (normal) 384 (89.9)
 F2 (mild) 39 (9.1)
 F3 (moderate) 4 (0.9)

Values are presented as mean±SD or number (%).

BMI, body mass index; AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, gamma-glutamyl transferase; HDL, high-density lipoprotein; LDL, low-density lipoprotein; HbA1c, hemoglobin A1c; HOMA-IR, homeostatic model assessment for insulin resistance (=fasting plasma glucose [mg/dL]×fasting plasma insulin [μIU/mL]/405); MRI-PDFF, magnetic resonance imaging–derived proton density fat fraction; METAVIR, meta-analysis of histological data in viral hepatitis; SD, standard deviation.

Table 2.
Comparison of median QUS parameters among hepatic steatosis grade groups
Hepatic steatosis grade (based on MRI-PDFF) Median (interquartile range)
P-value Post-hoc P-value (Bonferroni-corrected)
<5% (n=225) ≥5% to <10% (n=106) ≥10% (n=110) <5% vs. ≥5% to <10% ≥5% to <10% vs. ≥10%
TAI 0.64 (0.61-0.68) 0.71 (0.68-0.75) 0.83 (0.77-0.89) <0.001 <0.001 <0.001
TSI 90.2 (82.9-96.1) 99.2 (95.9-101.7) 103.7 (101.5-105.8) <0.001 <0.001 <0.001

QUS, quantitative ultrasonography; MRI-PDFF, magnetic resonance imaging–derived proton density fat fraction; TAI, tissue attenuation imaging; TSI, tissue scatter distribution imaging.

Table 3.
Diagnostic performance of QUS parameters and visual grade for the detection of hepatic steatosis
AUC (95% CI) Cutoff value Sensitivity (95% CI) Specificity (95% CI) PPV (95% CI) NPV (95% CI)
MRI-PDFF≥5%
 TAI 0.887 (0.857-0.918) ≥0.708 0.756 (0.699-0.812) 0.859 (0.812-0.907) 0.852 (0.802-0.902) 0.766 (0.712-0.812)
 TSI 0.910 (0.883-0.938) ≥98.79 0.765 (0.709-0.812) 0.932 (0.898-0.966) 0.923 (0.885-0.962) 0.787 (0.736-0.838)
 B-mode US 0.878 (0.848-0.908) ≥1 0.977 (0.958-0.997) 0.552 (0.456-0.592) 0.688 (0.637-0.739) 0.956 (0.918-0.994)
 Combined model (TAI+TSI) 0.939 (0.918-0.959) ≥0.513 0.851 (0.804-0.898) 0.874 (0.828-0.919) 0.879 (0.835-0.922) 0.845 (0.796-0.894)
 Combined model (TAI+TSI+B-mode US) 0.947 (0.929-0.966) ≥0.426 0.896 (0.856-0.9386) 0.835 (0.784-0.886) 0.853 (0.808-0.899) 0.882 (0.837-0.927)
MRI-PDFF≥10%
 TAI 0.951 (0.928-0.973) ≥0.748 0.889 (0.832-0.946) 0.890 (0.856-0.925) 0.754 (0.682-0.826) 0.955 (0.931-0.979)
 TSI 0.922 (0.897-0.946) ≥99.05 0.949 (0.909-0.989) 0.777 (0.731-0.824) 0.617 (0.556-0.688) 0.976 (0.957-0.995)
 B-mode US 0.875 (0.848-0.902) ≥2 0.940 (0.897-0.983) 0.758 (0.710-0.806) 0.595 (0.524-0.665) 0.971 (0.950-0.992)
 Combined model (TAI+TSI) 0.972 (0.959-0.985) ≥0.269 0.949 (0.909-0.989) 0.897 (0.863-0.931) 0.776 (0.708-0.845) 0.979 (0.962-0.996)
 Combined model (TAI+TSI+B-mode US) 0.975 (0.963-0.987) ≥0.311 0.932 (0.886-0.977) 0.929 (0.900-0.958) 0.832 (0.768-0.896) 0.973 (0.954-0.991)

QUS, quantitative ultrasonography; AUC, area under the receiver operating characteristic curve; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; MRI-PDFF, magnetic resonance imaging–derived proton density fat fraction; TAI, tissue attenuation imaging; TSI, tissue scatter distribution imaging; US, ultrasonography.

Table 4.
Factors associated with QUS parameters
Factor Univariable model
Multivariable model
Beta Standard error P-value Beta Standard error P-value
TAI
 Age 0.010 0.071 0.888
 Sex (Ref: male) -0.564 1.366 0.680
 BMI 1.326 0.159 <0.001
 Skin-liver capsule distance 10.976 1.377 <0.001 7.134 1.193 <0.001
 METAVIR score (Ref: 1, normal) 7.926 1.606 <0.001 4.808 1.314 <0.001
 AST 0.280 0.049 <0.001
 ALT 0.283 0.029 <0.001 0.202 0.026 <0.001
 GGT 0.083 0.017 <0.001
 Total cholesterol 0.020 0.013 0.109
 Triglyceride 0.041 0.006 <0.001 0.027 0.005 <0.001
 HDL-cholesterol -0.211 0.033 <0.001
 LDL-cholesterol 0.031 0.013 0.019
 Fasting glucose 0.174 0.028 <0.001
 HbA1c 3.750 0.735 <0.001
 HOMA-IR 3.659 0.389 <0.001
 Hypertension 2.489 1.001 0.013
 Diabetes 5.147 1.249 <0.001 3.710 1.001 <0.001
TSI
 Age -0.058 0.057 0.309
 Sex (Ref: male) -8.015 1.041 <0.001
 BMI 1.487 0.119 <0.001
 Skin-liver capsule distance 11.949 1.051 <0.001 7.100 1.244 <0.001
 METAVIR score (Ref: 1, normal) 4.984 1.321 <0.001 5.159 1.360 <0.001
 AST 0.145 0.041 <0.001
 ALT 0.202 0.024 <0.001 0.206 0.027 <0.001
 GGT 0.093 0.013 <0.001
 Total cholesterol -0.001 0.010 0.903
 Triglyceride 0.035 0.005 <0.001 0.028 0.005 <0.001
 HDL-cholesterol -0.230 0.026 <0.001
 LDL-cholesterol 0.010 0.011 0.342
 Fasting glucose 0.131 0.023 <0.001
 HbA1c 2.778 0.601 <0.001
 HOMA-IR 2.328 0.336 <0.001
 Hypertension 3.008 0.806 <0.001
 Diabetes 4.111 1.061 <0.001 3.881 1.038 <0.001

QUS, quantitative ultrasonography; TAI, tissue attenuation imaging; BMI, body mass index; METAVIR, meta-analysis of histological data in viral hepatitis; AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, gamma-glutamyl transferase; HDL, high-density lipoprotein; LDL, low-density lipoprotein; HbA1c, hemoglobin A1c; HOMA-IR, Homeostatic Model Assessment for Insulin Resistance; TSI, tissue scatter distribution imaging.

TOOLS
METRICS
0
Crossref
0
Scopus
797
View
74
Download
Editorial Office
A-304 Mapo Trapalace, 53 Mapo-daero, Mapo-gu, Seoul 04158, Korea
TEL : +82-2-763-5627   FAX : +82-2-763-6909   E-mail : office@ultrasound.or.kr
About |  Browse Articles |  Current Issue |  For Authors and Reviewers
Copyright © Korean Society of Ultrasound in Medicine.                 Developed in M2PI
Zoom in Close layer