Original ResearchFree Access

Comparison of Quantitative Liver US and MRI in Patients with Liver Disease

Published Online:https://doi.org/10.1148/radiol.212995

Abstract

Background

Quantitative US techniques can be used to identify changes of liver disease, but data regarding their diagnostic performance and relationship to MRI measures are sparse.

Purpose

To define associations between quantitative US and MRI measures of the liver in children, adolescents, and young adults with liver disease and to define the predictive ability of quantitative US measures to detect abnormal liver stiffening and steatosis defined with MRI.

Materials and Methods

In this prospective study, consecutive patients aged 8–21 years and known to have or suspected of having liver disease and body mass index less than 35 kg/m2 underwent 1.5-T MRI and quantitative liver US during the same visit at a pediatric academic medical center between April 2018 and December 2020. Acquired US parameters included shear-wave speed (SWS) and attenuation coefficient, among others. US parameters were compared with liver MR elastography and liver MRI proton density fat fraction (PDFF). Pearson correlation, multiple logistic regression, and receiver operating characteristic curve analyses were performed to assess associations and determine the performance of US relative to that of MRI.

Results

A total of 44 study participants (mean age, 16 years ± 4 [SD]; age range, 8–21 years; 23 male participants) were evaluated. There was a positive correlation between US SWS and MR elastography stiffness (r = 0.73, P < .001). US attenuation was positively correlated with MRI PDFF (r = 0.45, P = .001). For the prediction of abnormal (>2.8 kPa) liver shear stiffness, SWS (1.56 m/sec [7.3 kPa] cutoff) had an area under the receiver operating characteristic curve (AUC) of 0.95 with 91% sensitivity (95% CI: 71, 99) (20 of 22 participants) and 95% specificity (95% CI: 76, 99) (20 of 21 participants). For the prediction of abnormal (>5%) liver PDFF, US attenuation (0.55 dB/cm/MHz cutoff) had an AUC of 0.75 with a sensitivity of 73% (95% CI: 39, 94) (eight of 11 participants) and a specificity of 73% (95% CI: 55, 86) (24 of 33 participants).

Conclusion

In children, adolescents, and young adults with known or suspected liver disease, there was moderate to high correlation between US shear-wave speed (SWS) and MR elastography–derived stiffness. US SWS predicted an abnormal liver shear stiffness with high performance.

© RSNA, 2022

Online supplemental material is available for this article.

See also the editorial by Khanna and Alazraki in this issue.

Summary

In young individuals with a body mass index less than 35 kg/m2 who were known to have or suspected of having liver disease, there was moderate to high correlation between US shear-wave speed and MR elastography–derived shear stiffness.

Key Results

  • ■ In a prospective study of 44 young individuals with body mass index less than 35 kg/m2 and known to have or suspected of having liver disease, US shear-wave speed (SWS) and MR elastography–derived stiffness showed a moderate to high positive correlation (r = 0.73, P < .001).

  • ■ For the prediction of abnormal liver stiffness with MR elastography (>2.8 kPa), US SWS (>1.56 m/sec [7.3 kPa]) had an area under the receiver operating characteristic curve (AUC) of 0.95.

  • ■ For the prediction of abnormal liver MRI proton density fat fraction (>5%), US attenuation (>0.55 dB/cm/MHz) had an AUC of 0.75.

Introduction

Pediatric chronic liver disease may be caused by a wide range of conditions. Currently, nonalcoholic fatty liver disease is the most common cause of liver disease in children (1,2). The final common pathway of most chronic liver diseases is fibrosis, cirrhosis, and ultimately liver failure (3). Steatosis, congestion, and inflammation may coexist with fibrosis or even predominate in the early phases of disease. Although liver biopsy remains the reference standard with which to diagnose and define liver disease severity, imaging methods have shown promise in the detection and characterization of pediatric chronic liver disease (4). For example, US shear-wave elastography (SWE) and MR elastography can be used to identify liver stiffening indicative of fibrosis and have been used to distinguish patients with fibrosis from those with simple steatosis (59).

The greatest body of literature exists for SWE and MR elastography as quantitative measures of liver disease, specifically liver stiffening as a manifestation of fibrosis (7,8,10). There is also a substantial body of literature supporting the use of MRI liver proton density fat fraction (PDFF) to quantify hepatic steatosis (1113). US manufacturers have begun to develop and market additional quantitative techniques, including measurement of shear-wave dispersion, acoustic attenuation, and backscatter, as other means to identify changes in diffuse liver disease. To date, data regarding the diagnostic performance of these techniques and their relationship to MRI measures are sparse. While US- and MRI-based measures of diffuse liver disease are not interchangeable, understanding the relative performance and agreement of the two modalities is relevant, particularly as MR elastography is rapidly becoming accepted as a surrogate reference standard for clinically relevant hepatic stiffening and MRI PDFF is becoming accepted as a reference standard for steatosis in both clinical and research settings (9,1114). On the basis of existing literature, we hypothesized that there would be substantial correlations between US-measured shear-wave speed (SWS) and MR elastography–derived liver stiffness—and between US-measured acoustic attenuation and MRI PDFF.

The purpose of our study was to define the associations between quantitative US measures and MRI measures of chronic liver disease in children, adolescents, and young adults known to have or suspected of having chronic liver disease. Further, we sought to define the predictive ability of quantitative US measures to detect the presence of abnormal liver stiffening and steatosis as defined by the reference standard of MRI.

Materials and Methods

This prospective study was approved by the institutional review board at the Cincinnati Children’s Hospital Medical Center. All research activities were compliant with the Health Insurance Portability and Accountability Act. Written informed consent was obtained in person from parents or guardians of participants younger than 18 years. In addition, written assent was obtained from all pediatric participants aged at least 11 years. This study was financially supported by Canon Medical Systems USA. The authors of this work, none of whom is an employee of Canon Medical Systems, had complete control of study procedures, data, data analysis, and the content of this report.

Study Participants

Consecutive outpatients aged 8–21 years with known or clinically suspected chronic liver disease and previous MR elastography or US elastography at our institution were considered for inclusion in this study. Exclusion criteria included body mass index (BMI) greater than 35 kg/m2, inability to cooperate with US or MRI examinations, and pregnancy. Recruitment was targeted to encompass the range of clinically encountered liver shear stiffness values, with an approximate distribution of 20 with MR elastography–measured liver shear stiffness of at least 3 kPa, 25 with liver shear stiffness of more than 3 to 5 kPa, and five with liver shear stiffness greater than 5 kPa. A study coordinator used an imaging report search engine (Illuminate Insight; Softek) to identify patients who had previously undergone US elastography or MR elastography at our institution between 2014 and 2020. Patients who were eligible for enrollment were contacted by mail, email, or phone to solicit involvement in the study. Study visits occurred between April 2018 and December 2020 at a pediatric academic medical center.

Study Visits

Study procedures included a nonsedated MRI examination, height and weight measurements, and a US examination, all performed during the same visit. Participants were asked to fast for at least 4 hours prior to the study visit. BMI percentiles were calculated using the Centers for Disease Control and Prevention calculator (15). The electronic medical record (Epic Hyperspace; Epic Systems) was reviewed to collect participant sex, self-reported race and ethnicity, and liver disease diagnosis. Race and ethnicity data were collected to provide demographics relevant to generalizability of the results of this study.

MRI Examination

MRI examinations were performed with a 1.5-T Ingenia scanner (Philips Healthcare). The following MRI pulse sequences were performed: two-dimensional axial gradient-recalled echo (GRE) MR elastography, two-dimensional axial spin-echo echo-planar imaging (EPI) MR elastography, and axial mDixon Quant (to quantify PDFF). Protocol details are provided in Table E1 (online).

All MRI postprocessing was performed by a clinical image analyst with 7 years of experience in processing these examinations in the clinical environment, who was blinded to US data. After image acquisition, liver shear stiffness was manually measured for each of the GRE and EPI MR elastography sequences by drawing irregular regions of interest in the right hepatic lobe on each of four axial MRI scans through the middle of the liver. Regions of interest were placed at least 1 cm deep to the liver capsule while avoiding large vessels and areas of artifact and were guided by the magnitude images and scanner-generated 95% confidence maps. Results were expressed as a mean of mean stiffness values. Liver PDFF was manually measured on each of four axial PDFF parametric map images through the middle of the liver by placing an approximately 4 × 2 cm ovoid region of interest in the right hepatic lobe while avoiding the liver capsule, major vessels, and areas of artifact (IntelliSpace; Philips Healthcare).

US Examination

US was performed immediately before or immediately after the MRI examination using an Aplio i800 US machine (Canon Medical Systems) and an i8CX1 (PVI-475BX) transducer. Examinations were performed by a group of technologists (n = 3); all technologists had more than 5 years of experience using the Canon system in the clinical environment, and all were blinded to MRI data. Collected US measures included liver SWS (in meters per second), shear-wave dispersion (in meters per second per kilohertz), acoustic attenuation coefficient (in decibels per centimeter per megahertz), normalized local variance (NLV) (unitless), and hepatorenal index (unitless). At the time of this study, all these US techniques were U.S. Food and Drug Administration 510 (k) cleared. Specific US protocol details are provided in Appendix E1 (online). Figures 1 and 2 show examples of the quantitative US and MRI measures obtained in study participants.

Representative MR elastography and quantitative US images in a                         16-year-old boy with Fontan-associated liver disease and elevated liver                         shear stiffness (5.4 kPa with gradient-recalled echo MR elastography). (A)                         Axial MR elastogram with 95% confidence map overlay shows a stiff (5.4 kPa)                         heterogeneous liver. Image colors are indicative of stiffness (in                         kilopascals) according to the scale in the left of the elastogram image. (B)                         Transverse two-dimensional shear-wave elastography US image with shear-wave                         speed of 1.73 m/sec (9.0 kPa). (C) Transverse shear-wave dispersion map with                         dispersion of 16.86 m/sec/kHz. (D) Split-screen transverse image shows liver                         attenuation measurement of 0.46 dB/cm/MHz. (E) Split-screen transverse image                         shows normalized local variance measurement of 1.24. (F) Split-screen                         longitudinal image shows hepatorenal index measurement of 1.07. ATI-Pen =                         attenuation imaging penetration mode frequency, DR = dynamic range, fps =                         frames per second, G = gain, i8CX1 = Canon Apio i8CX1 transducer, Ml =                         mechanical index, NLV-Pen = normalized local variance penetration mode                         frequency, SF = spatial filter, Sw = shear-wave frequency.

Figure 1: Representative MR elastography and quantitative US images in a 16-year-old boy with Fontan-associated liver disease and elevated liver shear stiffness (5.4 kPa with gradient-recalled echo MR elastography). (A) Axial MR elastogram with 95% confidence map overlay shows a stiff (5.4 kPa) heterogeneous liver. Image colors are indicative of stiffness (in kilopascals) according to the scale in the left of the elastogram image. (B) Transverse two-dimensional shear-wave elastography US image with shear-wave speed of 1.73 m/sec (9.0 kPa). (C) Transverse shear-wave dispersion map with dispersion of 16.86 m/sec/kHz. (D) Split-screen transverse image shows liver attenuation measurement of 0.46 dB/cm/MHz. (E) Split-screen transverse image shows normalized local variance measurement of 1.24. (F) Split-screen longitudinal image shows hepatorenal index measurement of 1.07. ATI-Pen = attenuation imaging penetration mode frequency, DR = dynamic range, fps = frames per second, G = gain, i8CX1 = Canon Apio i8CX1 transducer, Ml = mechanical index, NLV-Pen = normalized local variance penetration mode frequency, SF = spatial filter, Sw = shear-wave frequency.

Representative proton density fat fraction (PDFF) parametric map and                         quantitative US images in an 11-year-old girl with nonalcoholic fatty liver                         disease and hepatic steatosis as defined by MRI PDFF of 15.6%. (A) Axial                         PDFF parametric map of a fatty liver (PDFF = 15.6%). (B) Transverse                         two-dimensional shear-wave elastography image with shear-wave speed of 1.08                         m/sec (3.5 kPa). (C) Transverse shear-wave dispersion map with dispersion of                         5.92 m/sec/kHz. (D) Split-screen transverse image shows liver attenuation                         measurement of 0.65 dB/cm/MHz. (E) Split-screen transverse image shows                         normalized local variance measurement of 1.01. (F) Split-screen longitudinal                         image shows hepatorenal index measurement of 1.11. ATI-Pen = attenuation                         imaging penetration mode frequency, DR = dynamic range, fps = frames per                         second, G = gain, i8CX1 = Canon Apio i8CX1 transducer, Ml = mechanical                         index, NLV-Pen = normalized local variance penetration mode frequency, SF =                         spatial filter, Sw = shear-wave frequency.

Figure 2: Representative proton density fat fraction (PDFF) parametric map and quantitative US images in an 11-year-old girl with nonalcoholic fatty liver disease and hepatic steatosis as defined by MRI PDFF of 15.6%. (A) Axial PDFF parametric map of a fatty liver (PDFF = 15.6%). (B) Transverse two-dimensional shear-wave elastography image with shear-wave speed of 1.08 m/sec (3.5 kPa). (C) Transverse shear-wave dispersion map with dispersion of 5.92 m/sec/kHz. (D) Split-screen transverse image shows liver attenuation measurement of 0.65 dB/cm/MHz. (E) Split-screen transverse image shows normalized local variance measurement of 1.01. (F) Split-screen longitudinal image shows hepatorenal index measurement of 1.11. ATI-Pen = attenuation imaging penetration mode frequency, DR = dynamic range, fps = frames per second, G = gain, i8CX1 = Canon Apio i8CX1 transducer, Ml = mechanical index, NLV-Pen = normalized local variance penetration mode frequency, SF = spatial filter, Sw = shear-wave frequency.

Image Review

All US and MRI scans were reviewed for quality after acquisition by a board-certified radiologist with more than 7 years of clinical and research experience with quantitative liver US and MRI (A.T.T.). Individual measurements deemed unacceptable on the basis of sample location or presence of artifact were excluded from the analysis.

Statistical Analyses

For the purpose of analyses, MRI served as the reference standard. MRI was analyzed as a continuous outcome as well as a categorical (normal vs abnormal) outcome. For categorical analyses, a liver shear stiffness greater than 2.8 kPa for both GRE and spin-echo EPI MR elastography and a liver PDFF greater than 5% were considered abnormal (14,16).

A descriptive analysis was performed to summarize participant characteristics. Quantitative US and MRI measures were summarized using means, medians, SDs, and IQRs, as appropriate. For the primary analysis, Pearson correlation (r value) was used to assess univariable associations between US and MRI measurements. T tests were used for the comparison of means across continuous variables.

Multiple logistic regression modeling with forward variable selection was used to generate models for the prediction of categorical MRI outcomes based on quantitative US results. Variables were allowed to enter the model if P < .05, and they exited the model if P > .1. Receiver operating characteristic curve analyses were performed to establish the sensitivity and specificity of continuous quantitative US measures to determine the presence of abnormal liver stiffening or hepatic steatosis. Optimal thresholds were selected on the basis of the Youden index.

Based on our planned primary analysis of the correlation between quantitative US and MRI measures and assuming a null correlation of 0, an alternative hypothesis of 0.5, α = .05, and 90% power, the calculated sample size was 37 participants. For 50 participants and assuming a correlation coefficient of 0.8 for any US-MRI comparison, the associated 95% CI was predicted to be 0.67, 0.88. Therefore, we targeted recruitment of up to 50 participants.

All statistical analyses were performed with MedCalc statistical software (version 20.009; MedCalc Software) and GraphPad Prism (version 8.0.0 for Windows; GraphPad Software). P < .05 was indicative of a significant difference.

Results

A total of 70 patients were contacted for participation in this study, and 45 agreed to participate. One female study participant was subsequently excluded due to her inability to complete MRI because of discomfort (Fig 3). After this one exclusion, a total of 44 study participants (23 male, 21 female) were evaluated. Because of technical issues, NLV and hepatorenal index data were not collected for two participants, while GRE MR elastography data were not analyzable for one participant because of high liver iron content. After review of the acquired measurements, one attenuation and seven individual hepatorenal index (one measurement in each of seven participants) measurements were excluded because of technical inadequacy. No SWS, SWS dispersion, or NLV measurements were excluded. The three most common causes of liver disease in the recruited study sample were Fontan-associated liver disease (seven of 44 participants, 16%), known fatty liver disease (six of 44 participants, 14%), and autoimmune liver disease (six of 44 participants, 14%) (Table 1).

Study sample flow diagram. Boxes on right denote excluded participants                     based on established exclusion criteria.

Figure 3: Study sample flow diagram. Boxes on right denote excluded participants based on established exclusion criteria.

Table 1: Participant Characteristics, Summary Statistics for Quantitative Liver Imaging, and Clinical Diagnoses

Table 1:

Participant demographic data, as well as summary MRI and US measures, are detailed in Table 1. Mean age and BMI were 16 years ± 4 (SD) and 21.9 kg/m2 ± 4.6. Mean liver shear stiffness with GRE MR elastography was 3.3 kPa ± 1.4, and mean liver PDFF was 5.6% ± 4.7 (Fig 4). Liver shear stiffness was higher in male participants than in female participants (GRE MR elastography, 3.8 kPa ± 1.7 kPa for male participants and 2.7 kPa ± 0.8 for female participants; P = .001). Similarly, liver PDFF was higher for male participants versus female participants (mean, 5.8% ± 5.7 for male participants and 5.3% ± 3.4 for female participants; P = .02).

(A) Histogram shows the distribution of liver shear stiffness values                     measured with MR elastography (MRE) in the study sample. Measurements above 2.8                     kPA suggest abnormal liver shear stiffness according to a study by Trout et al                     (14). (B) Histogram shows the distribution of liver steatosis values measured                     with MRI proton density fat fraction (PDFF) in the study sample. PDFF greater                     than 5% indicates hepatic steatosis.

Figure 4: (A) Histogram shows the distribution of liver shear stiffness values measured with MR elastography (MRE) in the study sample. Measurements above 2.8 kPA suggest abnormal liver shear stiffness according to a study by Trout et al (14). (B) Histogram shows the distribution of liver steatosis values measured with MRI proton density fat fraction (PDFF) in the study sample. PDFF greater than 5% indicates hepatic steatosis.

Univariable Associations among Quantitative US Measures

Table 2 details univariable associations among quantitative US measures. There was a strong positive correlation between median liver SWS and median dispersion (r = 0.94, P < .001). Attenuation negatively correlated with NLV (r = –0.37, P = .01) and positively correlated with hepatorenal index (r = 0.46, P = .003).

Table 2: Pearson Correlation Coefficients between Quantitative US Measures

Table 2:

Uni- and Multivariable Associations between Quantitative US Measures and MRI

Table 3 details univariable associations between quantitative US measures and MRI. There was a positive correlation between median liver SWS and liver shear stiffness as measured with both GRE MR elastography (r = 0.73, P < .001) and spin-echo EPI MR elastography (r = 0.71, P < .001) (Fig 5). Median liver SWS was also negatively correlated with liver PDFF (r = –0.35, P = .03). Median dispersion was similarly positively correlated with liver shear stiffness as measured with both GRE MR elastography (r = 0.74, P < .001) and spin-echo EPI MR elastography (r = 0.70, P < .001)—and negatively correlated with liver PDFF (r = –0.30, P = .04). Attenuation was positively correlated with liver PDFF (r = 0.45, P = .001), and NLV was negatively correlated with liver PDFF (r = –0.52, P = .001). There was no significant correlation between hepatorenal index and any MRI measure (r = –0.02 to 0.27, all P > .05).

Table 3: Pearson Correlation Coefficients between Quantitative US Measurements and MRI

Table 3:
Scatterplot of liver stiffness measured with US median shear-wave                         speed (SWS) (in meters per second) versus MR elastography (MRE) mean shear                         stiffness (in kilopascals) shows a moderate to high correlation (r = 0.73, P                         < .001). Thick solid line is the linear fit line for presented data.                         Curved dotted lines reflect 95% CIs. Vertical dotted line represents optimal                         SWS threshold of 1.56 m/sec to predict liver shear stiffness greater than                         2.8 kPa with MR elastography, shown by the horizontal dotted                         line.

Figure 5: Scatterplot of liver stiffness measured with US median shear-wave speed (SWS) (in meters per second) versus MR elastography (MRE) mean shear stiffness (in kilopascals) shows a moderate to high correlation (r = 0.73, P < .001). Thick solid line is the linear fit line for presented data. Curved dotted lines reflect 95% CIs. Vertical dotted line represents optimal SWS threshold of 1.56 m/sec to predict liver shear stiffness greater than 2.8 kPa with MR elastography, shown by the horizontal dotted line.

At multivariable modeling with forward selection to predict a categorically abnormal liver shear stiffness with MR elastography (>2.8 kPa), liver SWS was the only independent predictor, with an odds ratio of 2.9 (95% CI: 1.5, 5.5) (P = .001). In the prediction of liver PDFF greater than 5%, median attenuation was the only independent predictor, with an odds ratio of 2.6 (95% CI: 1.1, 6.0) (P = .03).

Diagnostic Performance of Quantitative US Measures for Categorical Prediction of Abnormal Liver Shear Stiffness

For the prediction of liver shear stiffness greater than 2.8 kPa with GRE MR elastography, US median liver SWS had an area under the receiver operating characteristic curve (AUC) of 0.95 (95% CI: 0.84, 0.99), with a sensitivity of 91% (95% CI: 71, 99) (20 of 22 participants) and a specificity of 95% (95% CI: 76, 99) (20 of 21 participants) with a cutoff of greater than 1.56 m/sec (7.3 kPa) (Fig 6). US liver dispersion similarly had an AUC of 0.92 (95% CI: 0.79, 0.98), with a sensitivity of 91% (95% CI: 71, 99) (20 of 22 participants) and specificity of 81% (95% CI: 58, 95) (17 of 21 participants) with a cutoff of greater than 10.4 m/sec/kHz. Similar statistics for spin-echo EPI MR elastography are presented in Table 4.

(A) Receiver operating characteristic curve for US shear-wave speed                         prediction of abnormal liver shear stiffness as defined by MR elastography                         greater than 2.8 kPa in the study sample. (B) Receiver operating                         characteristic curve for US attenuation prediction of abnormal liver                         steatosis as defined by MRI proton density fat fraction greater than 5% in                         the study sample. AUC = area under the receiver operating characteristic                         curve.

Figure 6: (A) Receiver operating characteristic curve for US shear-wave speed prediction of abnormal liver shear stiffness as defined by MR elastography greater than 2.8 kPa in the study sample. (B) Receiver operating characteristic curve for US attenuation prediction of abnormal liver steatosis as defined by MRI proton density fat fraction greater than 5% in the study sample. AUC = area under the receiver operating characteristic curve.

Table 4: Diagnostic Performance of Quantitative US Measurements for Abnormally High Liver Shear Stiffness (>2.8 kPa*) as Defined by MR Elastography

Table 4:

Diagnostic Performance of Quantitative US Measures for Categorical Prediction of Hepatic Steatosis

For the prediction of a liver PDFF greater than 5.0%, US liver dispersion had an AUC of 0.69 (95% CI: 0.54, 0.83), with a sensitivity of 91% (95% CI: 59, 100) (10 of 11 participants) and a specificity of 52% (95% CI: 34, 70) (17 of 33 participants) with a cutoff of no more than 12.1 m/sec/kHz. US attenuation had an AUC of 0.75 (95% CI: 0.60, 0.87), with a sensitivity of 73% (95% CI: 39, 94) (eight of 11 participants) and a specificity of 73% (95% CI: 55, 86) (24 of 33 participants) with a cutoff of more than 0.55 dB/cm/MHz (Fig 6). US NLV and hepatorenal index did not have significant predictive performance for steatosis (P = .07 and P = .29, respectively) (Table 5).

Table 5: Diagnostic Performance of Quantitative US Measurements for Hepatic Steatosis (Liver PDFF >5%)

Table 5:

Discussion

Quantitative MRI and US play an important and growing role in the assessment of liver disease (8,9,1719). However, there is a paucity of data comparing the modalities, and data that exist are largely related to elastography, with newer quantitative US techniques largely unstudied. Our study provides important data on the diagnostic performance of quantitative US in children with chronic liver disease, using MRI as a reference standard. In our study sample of children, adolescents, and young adults with known or suspected liver disease, US-assessed median liver shear-wave speed (SWS) and median liver shear-wave dispersion were highly positively correlated, and each had a moderate to high positive correlation with liver shear stiffness measured with MR elastography. At multivariable logistic regression, SWS was the only independent predictor of abnormal MRI liver shear stiffness (>2.8 kPa), and a liver SWS cutoff of 1.56 m/sec (7.3 kPa) predicted abnormal MRI liver shear stiffness with an area under the receiver operating characteristic curve (AUC) of 0.95 (95% CI: 0.84, 0.99). Median US acoustic attenuation was moderately positively correlated with MRI proton density fat fraction (PDFF), and normalized local variance, another putative metric of steatosis, was moderately negatively correlated with MRI PDFF. At multivariable logistic regression analysis, median acoustic attenuation was the one independent predictor for abnormal liver fat (>5%), with a cutoff of more than 0.55 dB/cm/MHz to predict abnormal MRI PDFF, with an AUC of 0.75 (95% CI: 0.60, 0.87).

The moderate to high positive correlation between US liver SWS and MRI liver shear stiffness shown in our study differs somewhat from a prior prospective study by Trout et al (7) that demonstrated a lower correlation (r = 0.33, P = .01) between liver SWS measured with point SWE and GRE MRI liver shear stiffness in a sample of children, adolescents, and young adults. A study by Levitte et al (20) reported a moderate correlation between SWE and MR elastography using 6-MHz point SWE (r = 0.52; 95% CI: 0.20, 0.74) but no significant correlation between 9-MHz point SWE or two-dimensional SWE and MR elastography (point SWE, r = 0.09; two-dimensional SWE, r = 0.02). The same study reported an SWS cutoff of 1.45 m/sec to have an AUC of 0.94 for the identification of an abnormal MRI liver shear stiffness of more than 3 kPa (20). This is slightly lower than the cutoff of 1.56 m/sec seen in our study for the identification of MRI liver shear stiffness of more than 2.8 kPa but has similar performance. The difference in results between our study and these prior studies may reflect technique differences (two-dimensional SWE vs point SWE), US manufacturer differences, or the fact that our study limited inclusion to participants with a BMI of less than 35 kg/m2.

Our finding of high correlation between US SWS and shear-wave dispersion warrants discussion. US shear-wave dispersion is thought to reflect tissue viscosity, a characteristic different from, but related to, tissue stiffness (8). There is some thought that changes in viscosity may be reflective of inflammation, potentially allowing the separation of inflammation from fibrosis in diseased liver (21,22). However, in our study, liver SWS and shear-wave dispersion were correlated, as they were in a prior study of healthy children and adults by Trout et al (8). Further study is needed to define whether shear-wave dispersion sufficiently diverges from liver SWS to be a useful complementary imaging marker and whether there is disease dependence of the diagnostic performance of dispersion.

Our results demonstrating a positive association between US attenuation coefficient and MRI PDFF are concordant with the literature (2326). In adults, Paige et al (25) described a similar strength of correlation (r = 0.69) using an Acuson S3000 US system (Siemens Healthineers). Recently, D’Hondt et al (24) reported an attenuation coefficient threshold of 0.54 dB/cm/MHz to achieve a sensitivity of 80% and specificity of 82% (AUC = 0.86) for MRI PDFF greater than 5% in a pediatric sample. Finally, in a meta-analysis including adult patients scanned with different US systems using biopsy or MRI PDFF as a reference standard, the pooled sensitivity and specificity of attenuation coefficient for any grade of steatosis were 76% and 84%, respectively (26).

There are limited prior data for NLV. In a prospective study of adults, Bae et al (27) reported NLV to have an AUC of 0.91 and optimal cutoffs of 1.095 for histopathologic steatosis of grade 1 or higher and 1.055 for steatosis of grade 2 or higher. The direction of the relationship shown by Bae et al with low NLV in patients with higher degrees of steatosis matches what we observed, although the relationship was weaker in our sample.

Our study had limitations. Sample size was small, resulting in large CIs for measured associations. We did not have liver biopsy findings available as a reference standard; instead, we benchmarked the performance of US using MRI measures of liver disease (MRE-derived liver stiffness and MRI PDFF) as a reference standard. This study used only one manufacturer’s US system and one vendor’s MRI system. Our recruitment emphasized the inclusion of the range of clinically encountered liver shear stiffness values, without specific recruitment across the range of hepatic steatosis. The relatively low frequency of steatosis in our sample (25%, 11 of 44 participants) and the relatively narrow range of included liver fat fractions may have impacted the strength of associations between US and MRI measures of liver fat.

In conclusion, we have shown that in children, adolescents, and young adults with body mass index less than 35 kg/m2 and known or suspected chronic liver diseases, there is a moderate to high correlation between US liver shear-wave speed (SWS) and MR elastography–derived stiffness, and there is a moderate correlation between US attenuation coefficient and liver MRI proton density fat fraction (PDFF). US liver SWS is used to predict a liver shear stiffness greater than 2.8 kPa with an area under the receiver operating characteristic curve (AUC) of 0.95 (95% CI: 0.84, 0.99), whereas US liver attenuation is used to predict a liver PDFF greater than 5% with an AUC of 0.75 (95% CI: 0.60, 0.87). Further research is needed to validate our findings and to assess the predictive performance of these US measures across different US system manufacturers.

Disclosures of conflicts of interest: V.P.V.A. No relevant relationships. J.R.D. Grants from Siemens Healthineers, Perspectum, and Canon Medical Systems USA. J.A.T. No relevant relationships. P.S.B. No relevant relationships. S.A.X. No relevant relationships. A.T.T. Grants from Siemens Healthineers, Perspectum, and Canon Medical Systems USA.

Acknowledgments

We acknowledge the contribution of Victoria Goodwin, RT(R),(CT),(MR); Samantha Summers, BS, MA; and Julie Plummer, BS (Department of Radiology, Cincinnati Children’s Hospital Medical Center) for their time and effort to help accomplish this study.

Author Contributions

Author contributions: Guarantors of integrity of entire study, V.P.V.A, A.T.T.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, V.P.V.A, J.R.D., J.A.T., A.T.T.; clinical studies, V.P.V.A, P.S.B., A.T.T.; statistical analysis, V.P.V.A, J.R.D., A.T.T.; and manuscript editing, V.P.V.A, J.R.D., J.A.T., S.A.X., A.T.T.

Supported by Canon Medical Systems USA.

Data sharing: Data generated or analyzed during the study are available from the corresponding author by request.

References

  • 1. Lavine JE , Schwimmer JB . Nonalcoholic fatty liver disease in the pediatric population. Clin Liver Dis 2004;8(3):549–558,viii–ix. Crossref, MedlineGoogle Scholar
  • 2. Huang JS , Barlow SE , Quiros-Tejeira RE , et al. Childhood obesity for pediatric gastroenterologists. J Pediatr Gastroenterol Nutr 2013;56(1):99–109. Crossref, MedlineGoogle Scholar
  • 3. Feldstein AE , Charatcharoenwitthaya P , Treeprasertsuk S , Benson JT , Enders FB , Angulo P . The natural history of non-alcoholic fatty liver disease in children: a follow-up study for up to 20 years. Gut 2009;58(11):1538–1544. Crossref, MedlineGoogle Scholar
  • 4. Vos MB , Abrams SH , Barlow SE , et al. NASPGHAN Clinical Practice Guideline for the Diagnosis and Treatment of Nonalcoholic Fatty Liver Disease in Children: Recommendations from the Expert Committee on NAFLD (ECON) and the North American Society of Pediatric Gastroenterology,. Hepatology and Nutrition (NASPGHAN). J Pediatr Gastroenterol Nutr 2017;64(2):319–334. Crossref, MedlineGoogle Scholar
  • 5. Schwimmer JB , Behling C , Angeles JE , et al. Magnetic resonance elastography measured shear stiffness as a biomarker of fibrosis in pediatric nonalcoholic fatty liver disease. Hepatology 2017;66(5):1474–1485. Crossref, MedlineGoogle Scholar
  • 6. Mouzaki M , Trout AT , Arce-Clachar AC , et al. Assessment of Nonalcoholic Fatty Liver Disease Progression in Children Using Magnetic Resonance Imaging. J Pediatr 2018;201:86–92. Crossref, MedlineGoogle Scholar
  • 7. Trout AT , Dillman JR , Xanthakos S , et al. Prospective assessment of correlation between US acoustic radiation force impulse and MR elastography in a pediatric population: Dispersion of US shear-wave speed measurement matters. Radiology 2016;281(2):544–552. LinkGoogle Scholar
  • 8. Trout AT , Xanthakos SA , Bennett PS , Dillman JR . Liver Shear Wave Speed and Other Quantitative Ultrasound Measures of Liver Parenchyma: Prospective Evaluation in Healthy Children and Adults. AJR Am J Roentgenol 2020;214(3):557–565. Crossref, MedlineGoogle Scholar
  • 9. Pirmoazen AM , Khurana A , El Kaffas A , Kamaya A . Quantitative ultrasound approaches for diagnosis and monitoring hepatic steatosis in nonalcoholic fatty liver disease. Theranostics 2020;10(9):4277–4289. Crossref, MedlineGoogle Scholar
  • 10. Fang C , Sidhu PS . Ultrasound-based liver elastography: current results and future perspectives. Abdom Radiol (NY) 2020;45(11):3463–3472. Crossref, MedlineGoogle Scholar
  • 11. Ferraioli G , Barr RG . Quantification of liver steatosis: Is CT equivalent to PDFF? AJR Am J Roentgenol 2021;216(4):W14. Crossref, MedlineGoogle Scholar
  • 12. Chen G , Jiang J , Wang X , et al. Evaluation of hepatic steatosis before liver transplantation in ex vivo by volumetric quantitative PDFF-MRI. Magn Reson Med 2021;85(5):2805–2814. Crossref, MedlineGoogle Scholar
  • 13. Qi Q , Weinstock AK , Chupetlovska K , et al. Magnetic resonance imaging-derived proton density fat fraction (MRI-PDFF) is a viable alternative to liver biopsy for steatosis quantification in living liver donor transplantation. Clin Transplant 2021;35(7):e14339. Crossref, MedlineGoogle Scholar
  • 14. Trout AT , Anupindi SA , Gee MS , et al. Normal liver stiffness measured with MR elastography in children. Radiology 2020;297(3):663–669. LinkGoogle Scholar
  • 15. BMI Percentile Calculator for Child and Teen. U.S. Centers for Disease Control and Prevention. https://www.cdc.gov/healthyweight/bmi/calculator.html. Accessed September 9, 2021. Google Scholar
  • 16. Tang A , Tan J , Sun M , et al. Nonalcoholic Fatty Liver Disease: MR Imaging of Liver Proton Density Fat Fraction to Assess Hepatic Steatosis. Radiology 2013;267(2):422–431. LinkGoogle Scholar
  • 17. Mosca A , Panera N , Crudele A , Alisi A . Noninvasive diagnostic tools for pediatric NAFLD: where are we now? Expert Rev Gastroenterol Hepatol 2020;14(11):1035–1046. Crossref, MedlineGoogle Scholar
  • 18. Ferraioli G , Barr RG , Dillman JR . Elastography for Pediatric Chronic Liver Disease:. A Review and Expert Opinion. J Ultrasound Med 2021;40(5):909–928. Crossref, MedlineGoogle Scholar
  • 19. Alhashmi GH , Gupta A , Trout AT , Dillman JR . Two-dimensional ultrasound shear wave elastography for identifying and staging liver fibrosis in pediatric patients with known or suspected liver disease: a clinical effectiveness study. Pediatr Radiol 2020;50(9):1255–1262. Crossref, MedlineGoogle Scholar
  • 20. Levitte S , Lee LW , Isaacson J , et al. Clinical use of shear-wave elastography for detecting liver fibrosis in children and adolescents with cystic fibrosis. Pediatr Radiol 2021;51(8):1369–1377. Crossref, MedlineGoogle Scholar
  • 21. Sugimoto K , Moriyasu F , Oshiro H , et al. The role of multiparametric US of the liver for the evaluation of nonalcoholic steatohepatitis. Radiology 2020;296(3):532–540. LinkGoogle Scholar
  • 22. Lee DH , Cho EJ , Bae JS , et al. Accuracy of Two-Dimensional Shear Wave Elastography and Attenuation Imaging for Evaluation of Patients With Nonalcoholic Steatohepatitis. Clin Gastroenterol Hepatol 2021;19(4):797–805.e7. Crossref, MedlineGoogle Scholar
  • 23. Ogino Y , Wakui N , Nagai H , Igarashi Y . The ultrasound-guided attenuation parameter is useful in quantification of hepatic steatosis in non-alcoholic fatty liver disease. JGH Open 2021;5(8):947–952. Crossref, MedlineGoogle Scholar
  • 24. D’Hondt A , Rubesova E , Xie H , Shamdasani V , Barth RA . Liver Fat Quantification by Ultrasound in Children: A Prospective Study. AJR Am J Roentgenol 2021;217(4):996–1006. Crossref, MedlineGoogle Scholar
  • 25. Paige JS , Bernstein GS , Heba E , et al. A pilot comparative study of quantitative ultrasound, conventional ultrasound, and MRI for predicting histology-determined steatosis grade in adult nonalcoholic fatty liver disease. AJR Am J Roentgenol 2017;208(5):W168–W177https://doi.org/10.2214/AJR.16.16726. Crossref, MedlineGoogle Scholar
  • 26. 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(1):83–92. Crossref, MedlineGoogle Scholar
  • 27. Bae JS , Lee DH , Lee JY , et al. Quantitative Assessment of Fatty Liver using Ultrasound with Normalized Local Variance Technique. Ultraschall Med 2021;42(6):599–606. Crossref, MedlineGoogle Scholar

Article History

Received: Nov 24 2021
Revision requested: Jan 18 2022
Revision received: Mar 28 2022
Accepted: Apr 8 2022
Published online: May 24 2022
Published in print: Sept 2022