Abstract
Background:
Metabolic dysfunction-associated steatotic liver disease (MASLD) is a growing global health challenge, with evidence indicating that hepatic inflammation and fibrosis are heterogeneous processes.
Purpose:
To measure liver mechanical property heterogeneity using MR elastography (MRE) and evaluate its potential as a biomarker for tissue inflammation and fibrosis in MASLD.
Materials and Methods:
Mechanical tissue heterogeneity in MASLD was assessed via 3D vector MRE pixel-wise histogram analysis of shear stiffness and loss modulus in preclinical and clinical studies. The preclinical study involved 25 rats examined monthly, while the clinical study analyzed data from 179 participants across two prospective studies (September 2015 - November 2022), including some underwent bariatric surgery with pre- and post-treatment MRE examinations. Mean and coefficient of variation (CV) of shear stiffness and loss modulus were calculated for each exam. Non-parametric tests and Spearman correlation were used to compare MRE-derived tissue mechanics with biopsy-confirmed fibrosis/inflammation and assess correlations with portal pressure and histopathological hepatic fibrosis.
Results:
The preclinical study revealed that in cirrhotic livers, CV of loss modulus positively correlated with portal pressure and fibrosis area ratio variation (ρ=0.52, P=.008; ρ=0.55, P=.005), respectively. In 10 healthy volunteers (median age, 36.5 [IQR, 34.0–38.8] years; 5 females) and 169 participants with MASLD (median age, 50.1 [IQR, 41.0–58.2] years; 118 females), CV of sheer stiffness (from 0.12 to 0.30 in healthy to fibrosis stage 4 participants) and loss modulus (from 0.31 to 0.51 in healthy to inflammation grade 3 participants) increased with increasing severity of fibrosis and inflammation, respectively. In 36 participants who underwent bariatric surgery, the CV of sheer stiffness significantly decreased at one year follow-up from 0.16 [IQR, 0.14–0.18] to 0.14 [IQR, 0.12–0.16] (P=.009).
Conclusion:
MRE-assessed tissue mechanical heterogeneity positively correlated with MASLD progression, demonstrating potential as a biomarker for liver disease severity and therapeutic intervention.
Clinical trial registration no. NCT02565446
Summary
Spatial heterogeneity of liver viscoelasticity, measured with 3D vector MR elastography positively correlated with metabolic dysfunction-associated steatotic liver disease progression.
Introduction
Metabolic dysfunction-associated steatotic liver disease (MASLD) is a condition of increasing global significance (1, 2). Characterized by fat accumulation in the liver, MASLD may lead to progressive disease including hepatic inflammation, fibrosis, cirrhosis, portal hypertension, and hepatocellular carcinoma (3, 4). The increasing prevalence of MASLD and complications over the past decades are attributed to increasing frequency of obesity (5). With the availability of the first Food and Drug Administration approved drug, Resmetirom, for metabolic dysfunction-associated steatohepatitis (MASH) in March 2024, reliable methods to track treatment effectiveness are relevant to patient management of assessing subtle spatial variation of disease severity, monitoring short-term therapeutic response, and predicting long-term risks of liver outcomes such as hepatocellular carcinoma development.
While liver disease resulting from MASLD has traditionally been viewed as a diffuse process that can be reasonably characterized by a single liver biopsy specimen, the effect of sampling error due to disease heterogeneity is increasingly being recognized. Studies show that when two needle biopsy specimens are obtained from different regions of the same liver, the histologic stage of fibrosis may differ in approximately one-third of the specimen pairs (6, 7). Liver stiffness heterogeneity, categorized into homogeneous and heterogeneous groups based on four-slice 2D MR elastography (MRE) value ranges, affects the concordance between MRE and biopsy-based fibrosis staging (8). This suggests that comprehensive assessment of tissue spatial response of the entire organ during disease progression/regression may be helpful in establishing models of care and individualized patient management.
Several established noninvasive tools exist for monitoring chronic liver disease progression (9, 10). Among these, MRE has emerged as having the highest diagnostic performance for detecting and staging hepatic fibrosis (11). Investigators have also found that MRE-assessed viscosity-related parameters, such as loss modulus or damping ratio, are potential indicators of hepatic inflammation (12) and may have potential to distinguish MASH from simple steatosis in combination with MRI-assessed proton-density fat fraction (PDFF) (13, 14). Unlike other noninvasive imaging tools, such as Vibration-Controlled Transient Elastography and Shear Wave Elastography, which are limited to evaluating small, localized volumes of liver tissue, MRE can assess mechanical properties across the entire organ. This whole-liver assessment offers the opportunity to explore the spatial heterogeneity of tissue viscoelasticity as a biomarker for use in managing liver disease.
The purpose of this study was to measure spatial heterogeneity of liver mechanical properties using MRE and evaluate its potential as a biomarker for heterogenous tissue inflammation and fibrosis development in MASLD.
Materials and Methods
Study Design
This study comprised of two parts: a preclinical animal model and a clinical study. The animal study was approved by the institutional animal care and use committee. The clinical study was a secondary analysis of two prospective studies approved by the institutional Review Board and one registered on ClinicalTrials.gov (Identifier: NCT02565446). The written informed consent was obtained from each participant.
MASLD Preclinical Model
Twenty-five male Wistar Han rats, aged 7 weeks, were used: 20 for progressive MASLD development and 5 controls between February 2021 and June 2021. After one-week acclimatization, rats were randomly assigned to different diets based on a well-established MASLD model (15, 16). Progressive liver disease was induced by using a choline-deficient high-fat diet (CDHFD, A16092003i, Research Diets Inc, New Brunswick, NJ) from 0 to 16 weeks (referred to as the CDHFD group); while age-matched control rats were given a normal chow diet (referred to as the control group). In vivo MRI/MRE data were collected at week 0, 4, 8, 12, and 16. Portal pressure was measured with a digital pressure analyzer device (Millar, v8.1.16, ADInstruments Pty Ltd, Australia) immediately after the MRI/MRE exam before animal sacrifice as a reference of portal hypertension (17). Liver samples were obtained after euthanizing the rats at week 12 for the control group and half of the CDHFD group, and at week 16 for the remaining half of the CDHFD group, corresponding to the compensated cirrhosis (week 12) and decompensated cirrhosis with severe portal hypertension and complications (week 16) (18, 19).
Study Participants
This study was a secondary analysis of two prospective studies approved by Mayo Clinic Institutional Review Board (a test-retest technical repeatability study of MRE on 10 healthy participants without liver biopsy, and a clinical study of MRE on 188 participants with liver biopsies – ClinicalTrials.gov Identifier: NCT02565446) from September 2015 to November 2022. Detailed descriptions of MASLD participants criteria for inclusion and exclusion are available in Appendix S1. One hundred and sixty-nine participants with suspected or diagnosed MASLD had completed 3D vector MRE at 60Hz and paired liver biopsy within one week; 36 of these had bariatric surgeries after baseline data collection and one-year follow-up biopsy and MRE (Fig 1). Eighty-eight of the 169 participants have been previously reported in a study using mean liver shear stiffness to predict MASH (13), and 13 of the 36 participants with bariatric surgery were reported to have mean liver shear stiffness changes post-surgery (14). This study differs from the previous two because it introduces a new parameter, the coefficient of variation (CV), which represents the spatial heterogeneity of liver viscoelasticity. This parameter was measured using MRE on the long-term follow-up outcome after continuous recruitment. Liver-related blood markers within 30 days of MRE were collected.
Figure 1:
Flow chart of participant with metabolic dysfunction-associated steatotic liver disease (MASLD) enrollment.
MRI/MRE Operation and processing
For both preclinical models and human participants, 3D vector MRE and 6-point Dixon MRI were used, with imaging protocols detailed in Table S1. Preclinical imaging was performed at 100Hz, while human imaging was conducted at 60Hz (20). Further details on imaging procedures are provided in Appendix S2, S3.
MRE data were processed using the Helmholtz equation (i.e., 3D Direct Inversion) to calculate the complex shear modulus (G*=G’+iG”) (21). The magnitude, real and imaginary part of shear modulus were referred to as shear stiffness, storage modulus and loss modulus (22). Regions of interests (ROIs) were manually placed to calculate mean and CV of shear stiffness and loss modulus. Detailed descriptions of 3D vector MRE and PDFF measurement are available in Appendix S4, S5.
Effect of noise on CV
As tissue stiffness increases, the CV rises due to reduced spatial derivatives, leading to greater noise sensitivity even in the absence of actual heterogeneity. The magnitude of this effect was investigated in a set of simulation experiments, which are included in Appendix S6 and Figure S1.
Histologic Analysis
For preclinical samples, liver specimens were stained with hematoxylin-eosin and picrosirius red, and analyzed by two expert pathologists (R.G., with 7 and T. M., with 8 years of experience in MASLD pathology, respectively) blinded to animal group or clinical history, and MRE results. Histologic analysis of fibrosis stage was based on MASH (previously called nonalcoholic steatohepatitis) Clinical Research Network criteria (23). Digitized slides of rat liver samples from the median, right lateral and left lateral lobes with picrosirius red staining were subdivided into 2x2 mm2 segments that were close to MRE acquisition resolution and used to quantify the collagen extent as percent proportionate area, which served as a reference standard to assess the severity of fibrosis (area quantification module, HALO software, version 3.6.4134.166, Indica Labs, Albuquerque). For clinical samples, all participants underwent MRI followed by core liver biopsies from the right lobe within 1 month. These biopsies were reviewed by clinical liver pathologists, with a second review by a research pathologist blinded to the initial interpretation. Any discrepancies were resolved by a third blinded pathologist.
Statistical Analysis
The intraclass correlation coefficient and Bland–Altman plots were used to assess agreement, bias, and repeatability of mechanical tissue property measurements with 3D vector MRE. The repeatability coefficient was calculated based on the Radiological Society of North America/Quantitative Imaging Biomarkers Alliance (https://qibawiki.rsna.org/index.php/Profiles ) recommendation (24). The correlation between MRE imaging measurements and histologic results was evaluated using Spearman correlation coefficient. The strength of association can be interpreted as excellent, good, moderate, and poor with |ρ|≥0.9, 0.75≤|ρ|<0.9, 0.5≤|ρ|<0.75, and |ρ|<0.5, respectively (25). We conducted non-parametric tests to compare MRE-derived tissue mechanics with biopsy-confirmed grades, with detailed descriptions provided in Appendix S7.
Results
MASLD Preclinical Model
Hepatic viscoelasticity and spatial heterogeneity increased with disease progression
Figure 2 demonstrates typical maps of MRE-derived tissue mechanical properties and the calculated means and spatial variations with MASLD disease progression for each animal. In the MASLD rat model, body weight gradually increased from 377 [IQR, 363, 406] g to 427 [IQR, 412, 467] g throughout the lifespan, but more slowly than that of the control group. After 4 weeks of the CDHFD diet, PDFF increased from 2.9% [IQR, 2.1%, 3.3%] to 46.8% [IQR, 43.7%, 48.9%], then gradually decreased but remained steatotic from 8 to 16 weeks. Mean shear stiffness and loss modulus steadily increased from 1.06 [IQR, 0.99, 1.10] kPa to 3.09 [IQR, 2.89, 3.31] kPa, and 0.32 [IQR, 0.29, 0.36] kPa to 0.80 [IQR, 0.69, 0.93] kPa, respectively, throughout the progression toward cirrhosis in the CDHFD group. Detailed measurements are included in Table S2. Significant differences were observed in mean shear stiffness and loss modulus in the CDHFD group when compared to baseline and control group measurements (P<.01). In pixel-wise histogram analysis, CV of shear stiffness and loss modulus remained relatively stable in controls, but consistently increased in the CDHFD group as MASLD progressed. All longitudinal follow-up time points had elevated heterogeneity compared to their corresponding baselines (P<.001). Comparing the two endpoints of 12 and 16 weeks (compensated cirrhosis and decompensated cirrhosis), mean shear stiffness and loss modulus were reduced (3.09 [IQR, 2.89, 3.31] kPa vs. 2.56 [IQR, 2.09, 2.93] kPa, P=.03; 0.80 [IQR, 0.69, 0.93] kPa vs. 0.67 [IQR, 0.41, 0.79] kPa, P=.10; respectively), while there was no evidence of a difference in heterogeneity measurements of CV for shear stiffness and loss modulus (0.51 [IQR, 0.39, 0.64] vs. 0.50 [IQR, 0.40, 0.65], P=.98; 0.92 [IQR, 0.80, 1.03] vs 1.00 [IQR, 0.88, 1.13], P=.27, respectively).
Figure 2:
Mechanical tissue properties of the livers from Wistar Han male rats assessed with 3D vector MR elastography (MRE). (A) Top row: MRE magnitude images; middle row: 3D vector MRE-derived shear stiffness maps; bottom row: 3D vector MRE-derived loss modulus maps. Left blue panel: control rats from baseline to 12 weeks; right pink panel: choline-deficient high-fat diet (CDHFD) rats from baseline to 16 weeks. (B-G) Scatter/line plots of (B) body weight changes for each animal throughout their lifespan, (C) mean shear stiffness, (D) coefficient of variation (CV) of shear stiffness, (E) proton-density fat fraction (PDFF), (F) mean loss modulus, and (G) CV of loss modulus. In all plots (B-G), blue dots/lines represent MRI/MRE factors of control animals and pink dots/lines represent MRI/MRE factors of CDHFD animals. Black ***, **, * represent crosswise comparison between CDHFD and control group P-values less than .001, .01, .05, respectively; fuchsia ***, **, * represent longitudinal comparison of CDHFD group P values less than .001, .01, .05, respectively.
Comparison of viscoelastic heterogeneity and histopathological or clinical standards
Figure 3 illustrates reference standards of hepatic fibrosis (picrosirius red percent for collagen extent) and portal hypertension (portal pressure measurement), and their association with tissue mechanical heterogeneity. First, fibrosis extent varied in different lobes of the liver. There were differences in picrosirius red percent of the left, middle, and right liver lobes in the CDHFD group but not in the control group (Fig 3A). Histograms of picrosirius red percent and shear stiffness in the corresponding rats indicated that both the mean and variance of distribution curve for histologically assessed picrosirius red percent and MRE-assessed liver stiffness gradually increased with disease progression (Fig 3B). The CV of loss modulus correlated with picrosirius red percent variation and portal pressure (ρ=0.55, [95% CI: 0.17, 0.79], P=.005; ρ=0.52, [95% CI: 0.13, 0.77], P=.008, respectively) (Fig 3C). As demonstrated in Table 1, compared with mean values, spatial variation of shear stiffness and loss modulus generally had better positive correlations with portal pressure, histologically assessed total fibrosis, and fibrosis variation (for shear stiffness: ρ=0.45, P=.02 vs. ρ=0.32, P=.12; ρ=0.53, P=.008 vs. ρ=0.29, P=.16; ρ=0.42, P=.04 vs. ρ=0.30, P=.15, respectively; and for loss modulus: ρ=0.52, P=.008 vs. ρ=0.16, P=.43; ρ=0.61, P=.001 vs. ρ=0.24, P=.24; ρ=0.55, P=.005 vs. ρ=0.19, P=.37, respectively).
Figure 3:
Reference standards of hepatic fibrosis and portal hypertension and their associations with mechanical tissue heterogeneity. (A) Picrosirius red staining of the left, middle, and right lobe of the liver in rats at 12- or 16-weeks of a normal (control) or choline-deficient high-fat diet (CDHFD group). The number in the lower right corner is the calculated fibrosis area ratio (Picrosirius red stained collagen extent as proportionate area in percentage) by HALO software. A bar plot of the mean Picrosirius red stained collagen extent percentage for different lobes/groups is located on the right side. (B) Histograms of Picrosirius red stained collagen extent percentage and shear stiffness of the liver in rats with 12-week normal or CDHFD diet, or 16-week CDHFD diet (mean Picrosirius red percentage: 0.10, 0.22, 0.52; standard deviation of Picrosirius red percentage: 0.03, 0.06, 0,09; mean of shear stiffness: 1.12 kPa, 1.94 kPa, 3.16 kPa; coefficient of variation (CV) in shear stiffness: 0.17, 0.40, 0.69, respectively). (C) Scatter plots of CV in loss modulus with Picrosirius red percentage standard deviation and portal pressure. Both demonstrated moderate and significant correlations with fibrosis extent and portal hypertension severity (0.5 ≤ρ<0.75, P <.01).
Table 1:
Correlation of pathology and 3D vector MRE factors in a rat metabolic dysfunction-associated steatotic liver disease model fibrosis stage 4
| Spearman Correlation | Mean of shear stiffness | Mean of loss modulus | CV of shear stiffness | CV of loss modulus | ||||
|---|---|---|---|---|---|---|---|---|
|
| ||||||||
| ρ (95% CI) | P Value | ρ (95% CI) | P Value | ρ (95% CI) | P Value | ρ (95% CI) | P Value | |
| MRI-PDFF | 0.48 (0.32,0.61) | <.001 | 0.42 (0.25,0.56) | <.001 | 0.50 (0.34,0.63) | <.001 | 0.43 (0.26,0.57) | <.001 |
| Portal pressure | 0.32 (−0.09,0.64) | .12 | 0.16 (−0.25,0.53) | .43 | 0.45 (0.05,0.73) | .02 | 0.52 (0.13,0.77) | .008 |
| picrosirius red percent | 0.29 (−0.13, 0.62) | .16 | 0.24 (−0.17,0.59) | .24 | 0.53 (0.14,0.77) | .008 | 0.61 (0.25,0.82) | .001 |
| picrosirius red percent standard deviation | 0.30 (−0.11,0.63) | .15 | 0.19 (−0.22,0.54) | .37 | 0.42 (0.02, 0.71) | .04 | 0.55 (0.17,0.79) | .005 |
Note. – CI = confidence interval, CV = coefficient of variation, PDFF = proton density fat fraction.
Study Participants
Test-retest technical repeatability of MRE-assessed viscoelastic heterogeneity
10 healthy participants (median age, 36.5 [IQR, 34.0–38.8] years; 5 females) completed the test-retest in one day. Agreement was moderate in their CV of shear stiffness and loss modulus test-retest values (ICC: 0.50 [-0.07, 0.84], P=.04 and 0.61 [-0.02, 0.89], P=.03, respectively) (Table S3). Bland–Altman plot analyses demonstrated bias of 0.011, -0.004 with 95% limits of agreement [-0.031, 0.052], [-0.103, 0.095] for CV of shear stiffness and loss modulus, respectively (Fig S2). The repeatability coefficient of CV of shear stiffness and loss modulus were 40.8%, 32.0%, higher than 7.4%, 16.2% from the repeatability test of mean values (Table S3).
Hepatic viscoelastic heterogeneity increased with fibrosis stage in MASLD
The prospective study excluded 15 participants with missing liver biopsy and 4 participants with technical failure in Figure 1. The cohort included 10 lean healthy volunteers (median age, 36.5 [IQR, 34.0–38.8] years; 5 females) and 169 participants (median age, 50.1 [IQR, 41.0–58.2] years; 118 females) with biopsy-proven MASLD, was divided into four groups based on fibrosis stage: 1) healthy participants; 2) participants with no fibrosis; 3) participants with mild-moderate fibrosis; 4) participants with advanced fibrosis (Table 2). The size of ROI for MRE increased from 123 cm3 in healthy livers to 246 cm3 in cirrhotic livers. Participants with MASLD fibrosis were older than healthy participants. The imaging biomarker values for PDFF, mean and CV of shear stiffness were significantly higher in MASLD fibrosis participants than healthy volunteer (13.9 [IQR, 7.4–21.3] vs. 1.8 [IQR, 1.3–2.5], P<.001; 2.35 [IQR, 2.03–2.83] kPa vs. 1.92 [IQR, 1.77–2.04] kPa, P<.001; 0.15 [IQR, 0.13–0.18] vs. 0.12 [IQR, 0.12–0.13], P<.001, respectively). As fibrosis progresses to more advanced stages, the prevalence of diabetes among participants also rises. Figure 4 shows a sample elastogram from each fibrosis group and corresponding scatter/box plots of MRE measurements. Mean liver stiffness was increasingly elevated with fibrosis extent, from 1.92 kPa in lean healthy participants up to 4.58 kPa in participants with advanced fibrosis. The maximum stiffness we observed in the MASLD cohort was 12.75 kPa. Meanwhile, CV of shear stiffness also consistently increased from 0.12 to 0.30 but with less ability to differentiate F1–2 from F0.
Table 2:
Characteristics of participants included in the secondary analyses from the two prospective studies
| Characteristic | Lean healthy (N=10) | Suspected / diagnosed MASLD with F=0 (N=72) | Suspected / diagnosed MASLD with 0<F≤2 (N=66) | Suspected / diagnosed MASLD with F≥3 (N=31) | P Value* |
|---|---|---|---|---|---|
| Demographics | |||||
| Female Sex | 5 (50) | 57 (79) | 43 (65) | 18 (58) | NA |
| Male Sex | 5 (50) | 15 (21) | 23 (35) | 13 (42) | NA |
| Age (year) | 36.5 (34.0, 38.8) | 49.5 (41.0, 57.0) | 48.5 (40.2, 56.8) | 56.0 (46.0, 59.5) | .006 |
| BMI (kg/m2) | 23.2 (22.2, 24.7) | 42.5 (36.5, 46.6) | 38.6 (33.8, 44.1) | 33.7 (30.0, 40.2) | <.001 |
| Quantitative MR imaging biomarkers of the liver | |||||
| ROI for PDFF (cm2) | 41.2 (37.6, 42.3) | 47.2(37.6, 52.2) | 52.9 (43.1, 65.9) | 50.7 (43.7, 69.9) | .001 |
| PDFF (%) | 1.8 (1.3, 2.5) | 9.7 (5.5, 16.6) | 20.4 (13.2, 26.7) | 10.3 (7.5, 17.1) | <.001 |
| ROI for MRE (cm3) | 123 (90, 140) | 195 (126, 273) | 218 (150, 340) | 246 (144, 495) | .002 |
| Mean shear stiffness (kPa) | 1.92 (1.77, 2.04) | 2.08 (1.89, 2.25) | 2.48 (2.15, 2.81) | 3.96 (3.26, 5.13) | <.001 |
| CV of shear stiffness | 0.12 (0.12, 0,13) | 0.13 (0.13, 0.18) | 0.14 (0.12, 0.18) | 0.18 (0.16, 0.22) | <.001 |
| Mean storage modulus (kPa) | 1.80 (1.66, 1.93) | 1.94 (1.73, 2.10) | 2.32 (2.04, 2.64) | 3.77 (3.14, 4.84) | <.001 |
| CV of storage modulus | 0.14 (0.13, 0.14) | 0.14 (0.13, 0.18) | 0.15 (0.13, 0.18) | 0.19 (0.16, 0.21) | .002 |
| Mean loss modulus (kPa) | 0.47 (0.44, 0.51) | 0.59 (0.53, 0.64) | 0.64 (0.55, 0.72) | 0.89 (0.65, 1.15) | <.001 |
| CV of loss modulus | 0.34 (0.33, 0.40) | 0.31 (0.27, 0.36) | 0.37 (0.31, 0.47) | 0.48 (0.43, 0.63) | <.001 |
| Liver-related blood markers | |||||
| AST (IU/L) | NA | 23 (18, 29) | 37 (26, 52) | 35 (26,61) | NA |
| ALT (IU/L) | NA | 26 (18, 37) | 46 (32, 78) | 47 (33, 70) | NA |
| Platelets (109/L) | NA | 268 (236, 311) | 240 (198, 274) | 202 (170, 260) | NA |
| Diabetic diagnosis | NA | 12 (17) | 23 (35) | 19 (61) | NA |
| Histologic Analysis of the Liver | |||||
| MASH diagnosis | NA | 11 (15) | 58 (88) | 24 (77) | NA |
| Steatosis | |||||
| Grade 0 | NA | 27 (38) | 3 (5) | 3 (10) | NA |
| Grade 1 | NA | 34 (47) | 30 (45) | 19 (61) | NA |
| Grade 2 | NA | 11 (15) | 19 (29) | 7 (23) | NA |
| Grade 3 | NA | 0 | 14 (21) | 2 (6) | NA |
| Hepatic Inflammation | |||||
| Grade 0 | NA | 40 (56) | 6 (9) | 5 (16) | NA |
| Grade 1 | NA | 31 (43) | 55 (83) | 23 (74) | NA |
| Grade 2 | NA | 1 (1) | 4 (6) | 3 (10) | NA |
| Grade 3 | NA | 0 | 1 (2) | 0 | NA |
| Hepatocellular Ballooning | |||||
| Grade 0 | NA | 62 (86) | 11 (17) | 6 (19) | NA |
| Grade 1 | NA | 10 (14) | 47 (71) | 15 (48) | NA |
| Grade 2 | NA | 0 | 8 (12) | 10 (32) | NA |
Note: ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, CV = coefficient of variation, MASH = metabolic dysfunction-associated steatohepatitis, NA = not applicable, PDFF = 6-point Dixon MRI assessed proton density fat fraction, ROI = region of interest.
All MRE-assessed parameters were calculated from 3D vector MRE-based volumetric regions of interest.
P values were calculated with the Kruskal-Wallis test for continuous variables comparison between subgroup of healthy volunteers to all suspected/diagnosed MASLD patients with biopsy.
Categorical variables are presented as counts (percentage), while continuous variables are presented as median (interquartile range).
Figure 4:
MR elastography (MRE) assessed mechanical tissue properties and spatial variation of metabolic dysfunction-associated steatotic liver disease at different fibrosis stages. (A) Top row: MRE magnitude images; Bottom row: 3D vector MRE-derived shear stiffness maps. Boxplots with dots of (B) mean and (C) coefficient of variation (CV) in liver shear stiffness in different fibrosis stages. Black ***, * represent crosswise comparison P-values less than .001, .05, respectively.
Hepatic viscoelastic heterogeneity increased with inflammation grade in MASLD
MRE-based liver loss modulus was closely associated with the progression of liver inflammation. The prospective cohort based on biopsy inflammation grade was divided into five groups: 1) 10 healthy volunteers (5.6%); 2) 51 participants with no inflammation (28.5%); 3) 109 with grade 1 (60.9%); 4) 8 with grade 2 (4.5%); 5) 1 with grade 3 (0.6%). Most were diagnosed with I0 or I1 group. Figure 5 shows a sample elastogram from each group and corresponding scatter/box plots of loss modulus measurements. Mean liver loss modulus was increasingly elevated with inflammation extent, from 0.54 kPa in lean healthy participants to 0.90 kPa in participants with inflammation grade 2. The maximum loss modulus observed was 2.11 kPa. Meanwhile, CV of loss modulus also consistently increased from 0.31 to 0.46 but with less ability to differentiate no inflammation suspected / diagnosed MASLD participants from healthy volunteers.
Figure 5:
MR elastography (MRE) assessed mechanical tissue properties and spatial variation of metabolic dysfunction-associated steatotic liver disease at different inflammation grades. (A) Top row: MRE magnitude images; Bottom row: 3D vector MRE-derived loss modulus maps. Boxplots with dots of (B) mean and (C) coefficient of variation (CV) of liver loss modulus in different inflammation grades. Black ***, * represent crosswise comparison P-values less than .001, .05, respectively.
Hepatic viscoelastic heterogeneity decreased after treatment
Table 3 shows participants demographic characteristics and results from 36 participants with a one-year follow-up paired liver biopsy and 3D vector MRE. All had a reduction of body mass index from 43.6 [IQR, 41.5, 52.7] kg/m2 to 32.2 [IQR, 28.5, 37.8] kg/m2. The post-treatment heterogeneity of mechanical properties for individuals, as shown in the reconstructed shear stiffness and loss modulus maps of Figure 6A, are consistently reduced. We studied the effect of ROI size on CV calculation to adjust the influence of wave penetration depth changes by refining the confidence value threshold (Fig S3). When the calculated areas exceeded 30 cm3, both mean and CV factors tended to stabilize (Intraclass correlation coefficient for mean shear stiffness, mean loss modulus, CV of shear stiffness, CV of loss modulus: 0.99, 0.98, 0.92, 0.97, respectively). As shown in Figure 6B and 6D, the scatter plots indicated that although no evidence of differences in mean shear stiffness and loss modulus change at 60 Hz were observed (2.15 [IQR, 1.99, 2.35] kPa vs. 2.09 [IQR, 1.84, 2.29] kPa, P=.06; 0.59 [IQR, 0.54, 0.66] kPa vs. 0.57 [IQR, 0.50, 0.61] kPa, P=.27, respectively), the liver heterogeneity, assessed by CV of shear stiffness exhibited a notable decrease (0.16 [IQR, 0.14, 0.18] vs. 0.14 [IQR, 0.12, 0.16], P=.009) (Fig 6C).
Table 3:
Clinical and 3D vector MRE characteristics of 36 participants with metabolic dysfunction-associated steatotic liver disease before and one-year post-surgery
| Characteristic | Before surgery | 1 year after surgery | P Value* |
|---|---|---|---|
| Demographics | |||
| Female sex | 32 (89) | 32 (89) | NA |
| Male sex | 4 (11) | 4 (11) | NA |
| Age (year) | 50.2 (42.4, 57.6) | 51.2 (43.5, 58.6) | NA |
| BMI (kg/m2) | 43.6 (41.5, 52.7) | 32.2 (28.5, 37.8) | <.001 |
| Quantitative MR imaging biomarkers of the liver | |||
| PDFF (%) | 9.9 (6.3, 18.6) | 2.7 (2.1, 3.9) | <.001 |
| ROI for MRE (cm3) | 209 (156, 297) | 163 (120, 271) | .57 |
| Mean shear stiffness (kPa) | 2.15 (1.99, 2.35) | 2.09 (1.84, 2.29) | .06 |
| CV of shear stiffness | 0.16 (0.14, 0.18) | 0.14 (0.12, 0.16) | .009 |
| Mean loss modulus (kPa) | 0.59 (0.54, 0.66) | 0.57 (0.50, 0.61) | .27 |
| CV of loss modulus | 0.34 (0.30, 0.42) | 0.31 (0.27, 0.36) | .06 |
| Histologic Analysis of the Liver | |||
| MASH diagnosis | 11 (31) | 0 | NA |
| Hepatic fibrosis | |||
| Stage 0 | 25 (69) | 34 (94) | NA |
| Stage 1–2 | 10 (28) | 1 (3) | NA |
| Stage 3–4 | 1 (3) | 1 (3) | |
| Hepatic Inflammation | |||
| Grade 0 | 19 (53) | 29 (81) | NA |
| Grade 1 | 16 (44) | 7 (19) | NA |
| Grade 2 | 1 (3) | 0 | NA |
Note. – BMI= body mass index, CV = coefficient of variation, MASH = metabolic dysfunction-associated steatohepatitis, MRE = MR elastography, NA = not applicable, PDFF = proton density fat fraction, ROI= regions of interest.
P values were calculated with the paired Wilcoxon test for continuous variables. Categorical variables are presented as counts (percentage), while continuous variables are presented as medians (quantiles).
Figure 6:
Mechanical properties of the participants’ livers before bariatric surgery and one-year follow-up assessed with MR elastography (MRE). (A) Top row: MRE magnitude images; Middle row: 3D vector MRE-derived shear stiffness maps; Bottom row: 3D vector MRE-derived loss modulus maps. First column: reconstructed maps before bariatric surgery. Second column: reconstructed maps at one-year follow-up after surgery. The scatter/line plots of (B) the mean and (C) coefficient of variation (CV) for liver stiffness, and (D) the mean and (E) CV for loss modulus changes in before and one-year follow-up post-surgery in 36 metabolic dysfunction-associated steatotic liver disease patients. The horizontal-colored lines indicate mean values (pink: pre-operation; blue: post-operation).
Discussion
Hepatic tissue heterogeneity is a recognized phenomenon in chronic liver diseases (2, 6). In this study, by measuring spatial variations in liver mechanical properties via 3D vector MR elastography (MRE) and pixel-wise analysis of shear stiffness and loss modulus, we found that heterogeneity, represented by coefficient of variation (CV), increased with fibrosis stage (from 0.13 to 0.50 in 0 to 16 weeks, 0.12 to 0.30 in healthy to fibrosis stage 4 participants) and inflammation grade (from 0.45 to 1.00 in 0 to 16 weeks, 0.31 to 0.51 in healthy to inflammation grade 3 participants) in preclinical and clinical studies of metabolic dysfunction-associated steatotic liver disease (MASLD).
Heterogeneity assessed in shear stiffness correlated more strongly with portal pressure and fibrosis area ratio than mean values (ρ=0.45, P=.02 vs. ρ=0.12, P=.16; ρ=0.53, P=.008 vs. ρ=0.29, P=.16) in the preclinical model. Previous head-to-head comparison with MRE-assessed macroscopic mechanical tissue heterogeneity and atomic force microscopy demonstrated a heterogeneous distribution of liver stiffness in a carbon tetrachloride mouse model at microscopic levels, accompanied by heterogeneity of hepatic stellate cells subpopulations identified by single-cell RNA sequencing (26). However, atomic force microscopy requires a fresh tissue sample, which is invasive, has limited repeatability if using biopsy samples in clinical practice, and is unable to characterize the in vivo dynamic bi-phasic mechanical environment of the liver (27). The ROI of liver parenchyma in healthy volunteers to MASLD participants with advanced fibrosis, assessed using 3D vector MRE, were from 123 cm3 to 246 cm3. This volume is much higher than that sampled through other elastography methods such as shear wave elastography (20 cm3) and transient elastography (4 cm3), and nearly 100 times larger than that of a typical liver biopsy (28, 29). Assessing larger volumes of liver enables a more reliable assessment of spatial heterogeneity of parenchymal changes based on preliminary results of CV of shear stiffness and loss modulus stabilization with volumes greater than 30 cm3 (30).
Our results demonstrated that tissue mechanical heterogeneity has a positive correlation with disease progression in both preclinical and clinical settings. From the monthly MRE exams in a longitudinal study design with a sex- and age-controlled animal model, which reduces the necessity for longitudinal histologic evaluation, the mean and CV of viscoelastic properties exhibited marked increases from baseline to mild-moderate fibrosis and eventually cirrhosis. In the progression of decompensated cirrhosis, the observed reduction in mean shear stiffness and loss modulus is unexpected. One potential explanation for this decrease could be the increased tissue heterogeneity caused by ongoing liver regeneration, necroinflammation, and extracellular matrix deposition. These processes may result in localized areas of decreased stiffness, despite overall disease progression. This interplay between regenerative nodules, necrosis, and fibrotic tissue could mask the expected increase in mean shear stiffness and loss modulus. Interestingly, compared with the mean mechanical properties, the CV of loss modulus demonstrated a better correlation with portal pressure and microscopic fibrosis variation at end-stage liver fibrosis. Our findings complement prior research that loss modulus and its heterogeneity reflect increased interstitial fluid volume, as well as its spatial distribution variation that might be associated with increased microthrombosis and vascular permeability in portal hypertension (13, 31).
The results from our clinical cohort demonstrate that tissue mechanical heterogeneity, accessed by CV of shear stiffness and loss modulus, was increasingly elevated with fibrosis and inflammation extent as confirmed by liver biopsy. These findings are consistent with the results from our animal MASLD study. Bariatric surgery is the most effective and sustainable treatment for MASLD, leading to improvements in body mass, as well as histologically assessed steatosis, ballooning, and fibrosis in approximately 75% of participants (32). Our study shows a notable decrease in tissue mechanical heterogeneity following this intervention, especially in most obese participants without detectable fibrosis; therefore, no evidence of differences in mean shear stiffness and loss modulus at 60 Hz were observed. Previous findings have revealed a decrease of loss modulus at 30 Hz in obese participants (body mass index > 30 kg/m2) who had normal biopsy results with follow-up after surgery (27). Thus, spatial variation of tissue mechanical properties has the potential to complement mean measurements at 60Hz, providing similar differentiation as loss modulus at 30Hz, for comprehensively revealing tissue response during disease progression or regression.
This study had several limitations. First, the preclinical study was conducted using a simplified diet induced progressive MASLD model. In practice, MASLD often presents alongside other liver conditions, such as viral hepatitis, hemochromatosis, and alcoholic liver disease. Consequently, more diverse models encompassing various etiologies and disease states may enhance the validation of the relationships between mechanical heterogeneity and histologic findings. Second, ROI manual drawing in 3D MRE may introduce variability. Implementing automated ROI delineation can enhance consistency of CV measurements and facilitate multi-center applications in future prospective studies (33). Finally, our prospective study predominantly involved cases of mild to moderate MASLD with few cases of advanced fibrosis. However, this provided a more controlled operative environment for obtaining liver biopsies from a large number of participants.
In conclusion, the assessment of spatial heterogeneity in liver viscoelastic mechanical properties using 3D vector MR elastography (MRE), showed a positive correlation with disease progression in a preclinical model and in a clinical cohort. Spatial variation of tissue mechanical properties has the potential to complement mean measurements of these biomarkers for comprehensively assessing tissue response during metabolic dysfunction-associated steatotic liver disease (MASLD) progression or regression. However, further investigation is needed to confirm these findings of spatial heterogeneity, particularly in assessing short-term therapeutic responses and predicting long-term risks of developing hepatocellular carcinoma.
Supplementary Material
Key Results
In a preclinical longitudinal study of 20 rats with metabolic dysfunction-associated steatotic liver disease (MASLD), spatial heterogeneity of liver viscoelasticity significantly correlated with portal pressure (ρ=0.52, P=.008) and fibrosis area ratio variation (ρ=0.55, P=.005), respectively.
In a prospective study of 179 participants, spatial heterogeneity of liver viscoelasticity consistently increased with biopsy-proven fibrosis and inflammation. Among 36 participants who underwent bariatric surgery, coefficient of variation of shear stiffness significantly decreased by 13% at one-year follow-up (P=.009).
Acknowledgment:
We sincerely thank Thomas Hulshizer, Phillip Rossman and Corey Woxland for their invaluable technical support to this study, Desiree Lanzino for editorial assistance, and Takeda Pharmaceuticals Company Limited for funding support.
Funding support:
NIH R01 EB17197 (Yin); R01 DK132718 (Yin); R01 DK136731 (Yin); R37 EB001981 (Ehman); UH3 AA026887 (Shah); K23 DK115594 (Allen).
Data sharing statement:
Data generated or analyzed during the study are available from the corresponding author by request. Hao Wu, Richard L. Ehman, Meng Yin managed the data and guaranteed the integrity of entire study.
Abbreviations
- CDHFD
choline-deficient high-fat diet
- CV
coefficient of variation
- MASLD
metabolic dysfunction-associated steatotic liver disease
- MRE
MR elastography
- MASH
metabolic dysfunction-associated steatohepatitis
- PDFF
proton density fat fraction
- ROI
regions of interest
References
- 1.Paik JM, Golabi P, Younossi Y, Srishord M, Mishra A, Younossi ZM. The Growing Burden of Disability Related to Nonalcoholic Fatty Liver Disease: Data From the Global Burden of Disease 2007–2017. Hepatol Commun 2020;4(12):1769–1780. doi: 10.1002/hep4.1599 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Han SK, Baik SK, Kim MY. Non-alcoholic fatty liver disease: Definition and subtypes. Clin Mol Hepatol 2023;29(suppl):S5–S16. doi: 10.3350/cmh.2022.0424 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Takahashi Y, Dungubat E, Kusano H, Fukusato T. Pathology and Pathogenesis of Metabolic Dysfunction-Associated Steatotic Liver Disease-Associated Hepatic Tumors. Biomedicines 2023;11(10). doi: 10.3390/biomedicines11102761 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kleiner DE, Brunt EM, Wilson LA, Behling C, Guy C, Contos M, Cummings O, Yeh M, Gill R, Chalasani N, Neuschwander-Tetri BA, Diehl AM, Dasarathy S, Terrault N, Kowdley K, Loomba R, Belt P, Tonascia J, Lavine JE, Sanyal AJ, Nonalcoholic Steatohepatitis Clinical Research N. Association of Histologic Disease Activity With Progression of Nonalcoholic Fatty Liver Disease. JAMA Netw Open 2019;2(10):e1912565. doi: 10.1001/jamanetworkopen.2019.12565 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Collaborators GBDC. The global, regional, and national burden of cirrhosis by cause in 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet Gastroenterol Hepatol 2020;5(3):245–266. doi: 10.1016/S2468-1253(19)30349-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Eslam M, Sanyal AJ, George J, International Consensus P. MAFLD: A Consensus-Driven Proposed Nomenclature for Metabolic Associated Fatty Liver Disease. Gastroenterology 2020;158(7):1999–2014 e1991. doi: 10.1053/j.gastro.2019.11.312 [DOI] [PubMed] [Google Scholar]
- 7.Ratziu V, Charlotte F, Heurtier A, Gombert S, Giral P, Bruckert E, Grimaldi A, Capron F, Poynard T, Group LS. Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Gastroenterology 2005;128(7):1898–1906. doi: 10.1053/j.gastro.2005.03.084 [DOI] [PubMed] [Google Scholar]
- 8.Kawamura N, Imajo K, Kalutkiewicz KJ, Nagai K, Iwaki M, Kobayashi T, Nogami A, Honda Y, Kessoku T, Ogawa Y, Higurashi T, Hosono K, Takahashi H, Yoneda M, Saito S, Aishima S, Toyoda H, Hayashi H, Sumida Y, Ehman RL, Nakajima A. Influence of liver stiffness heterogeneity on staging fibrosis in patients with nonalcoholic fatty liver disease. Hepatology 2022;76(1):186–195. doi: 10.1002/hep.32302 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Younossi ZM, Loomba R, Anstee QM, Rinella ME, Bugianesi E, Marchesini G, Neuschwander-Tetri BA, Serfaty L, Negro F, Caldwell SH, Ratziu V, Corey KE, Friedman SL, Abdelmalek MF, Harrison SA, Sanyal AJ, Lavine JE, Mathurin P, Charlton MR, Goodman ZD, Chalasani NP, Kowdley KV, George J, Lindor K. Diagnostic modalities for nonalcoholic fatty liver disease, nonalcoholic steatohepatitis, and associated fibrosis. Hepatology 2018;68(1):349–360. doi: 10.1002/hep.29721 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Smith A, Baumgartner K, Bositis C. Cirrhosis: Diagnosis and Management. Am Fam Physician 2019;100(12):759–770. [PubMed] [Google Scholar]
- 11.Lim JK, Flamm SL, Singh S, Falck-Ytter YT, Clinical Guidelines Committee of the American Gastroenterological A. American Gastroenterological Association Institute Guideline on the Role of Elastography in the Evaluation of Liver Fibrosis. Gastroenterology 2017;152(6):1536–1543. doi: 10.1053/j.gastro.2017.03.017 [DOI] [PubMed] [Google Scholar]
- 12.Chen J, Martin-Mateos R, Li J, Yin Z, Chen J, Lu X, Glaser KJ, Mounajjed T, Yashiro H, Siegelman J, Winkelmann CT, Wang J, Ehman RL, Shah VH, Yin M. Multiparametric magnetic resonance imaging/magnetic resonance elastography assesses progression and regression of steatosis, inflammation, and fibrosis in alcohol-associated liver disease. Alcohol Clin Exp Res 2021. doi: 10.1111/acer.14699 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Allen AM, Shah VH, Therneau TM, Venkatesh SK, Mounajjed T, Larson JJ, Mara KC, Schulte PJ, Kellogg TA, Kendrick ML, McKenzie TJ, Greiner SM, Li J, Glaser KJ, Wells ML, Chen J, Ehman RL, Yin M. The Role of Three-Dimensional Magnetic Resonance Elastography in the Diagnosis of Nonalcoholic Steatohepatitis in Obese Patients Undergoing Bariatric Surgery. Hepatology 2020;71(2):510–521. doi: 10.1002/hep.30483 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Allen AM, Shah VH, Therneau TM, Venkatesh SK, Mounajjed T, Larson JJ, Mara KC, Kellogg TA, Kendrick ML, McKenzie TJ, Greiner SM, Li J, Glaser KJ, Wells ML, Gunneson TJ, Ehman RL, Yin M. Multiparametric Magnetic Resonance Elastography Improves the Detection of NASH Regression Following Bariatric Surgery. Hepatol Commun 2020;4(2):185–192. doi: 10.1002/hep4.1446 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Charlton M, Krishnan A, Viker K, Sanderson S, Cazanave S, McConico A, Masuoko H, Gores G. Fast food diet mouse: novel small animal model of NASH with ballooning, progressive fibrosis, and high physiological fidelity to the human condition. Am J Physiol Gastrointest Liver Physiol 2011;301(5):G825–834. doi: 10.1152/ajpgi.00145.2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Vacca M, Kamzolas I, Harder LM, Oakley F, Trautwein C, Hatting M, Ross T, Bernardo B, Oldenburger A, Hjuler ST, Ksiazek I, Linden D, Schuppan D, Rodriguez-Cuenca S, Tonini MM, Castaneda TR, Kannt A, Rodrigues CMP, Cockell S, Govaere O, Daly AK, Allison M, Honnens de Lichtenberg K, Kim YO, Lindblom A, Oldham S, Andreasson AC, Schlerman F, Marioneaux J, Sanyal A, Afonso MB, Younes R, Amano Y, Friedman SL, Wang S, Bhattacharya D, Simon E, Paradis V, Burt A, Grypari IM, Davies S, Driessen A, Yashiro H, Pors S, Worm Andersen M, Feigh M, Yunis C, Bedossa P, Stewart M, Cater HL, Wells S, Schattenberg JM, Anstee QM, Investigators L, Tiniakos D, Perfield JW, Petsalaki E, Davidsen P, Vidal-Puig A. An unbiased ranking of murine dietary models based on their proximity to human metabolic dysfunction-associated steatotic liver disease (MASLD). Nat Metab 2024;6(6):1178–1196. doi: 10.1038/s42255-024-01043-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kaplan DE, Ripoll C, Thiele M, Fortune BE, Simonetto DA, Garcia-Tsao G, Bosch J. AASLD Practice Guidance on risk stratification and management of portal hypertension and varices in cirrhosis. Hepatology 2023. doi: 10.1097/HEP.0000000000000647 [DOI] [PubMed] [Google Scholar]
- 18.Pak W, Takayama F, Mine M, Nakamoto K, Kodo Y, Mankura M, Egashira T, Kawasaki H, Mori A. Anti-oxidative and anti-inflammatory effects of spirulina on rat model of non-alcoholic steatohepatitis. J Clin Biochem Nutr 2012;51(3):227–234. doi: 10.3164/jcbn.12-18 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Takeuchi-Yorimoto A, Noto T, Yamada A, Miyamae Y, Oishi Y, Matsumoto M. Persistent fibrosis in the liver of choline-deficient and iron-supplemented L-amino acid-defined diet-induced nonalcoholic steatohepatitis rat due to continuing oxidative stress after choline supplementation. Toxicol Appl Pharmacol 2013;268(3):264–277. doi: 10.1016/j.taap.2013.01.027 [DOI] [PubMed] [Google Scholar]
- 20.Li J, Lu X, Zhu Z, Kalutkiewicz KJ, Mounajjed T, Therneau TM, Venkatesh SK, Sui Y, Glaser KJ, Hoodeshenas S, Manduca A, Shah VH, Ehman RL, Allen AM, Yin M. Head-to-head comparison of magnetic resonance elastography-based liver stiffness, fat fraction, and T1 relaxation time in identifying at-risk NASH. Hepatology 2023;78(4):1200–1208. doi: 10.1097/HEP.0000000000000417 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Manduca A, Bayly PJ, Ehman RL, Kolipaka A, Royston TJ, Sack I, Sinkus R, Van Beers BE. MR elastography: Principles, guidelines, and terminology. Magn Reson Med 2021;85(5):2377–2390. doi: 10.1002/mrm.28627 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Shi Y, Qi YF, Lan GY, Wu Q, Ma B, Zhang XY, Ji RY, Ma YJ, Hong Y. Three-dimensional MR Elastography Depicts Liver Inflammation, Fibrosis, and Portal Hypertension in Chronic Hepatitis B or C. Radiology 2021;301(1):154–162. doi: 10.1148/radiol.2021202804 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, Ferrell LD, Liu YC, Torbenson MS, Unalp-Arida A, Yeh M, McCullough AJ, Sanyal AJ, Nonalcoholic Steatohepatitis Clinical Research N. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005;41(6):1313–1321. doi: 10.1002/hep.20701 [DOI] [PubMed] [Google Scholar]
- 24.Serai SD, Obuchowski NA, Venkatesh SK, Sirlin CB, Miller FH, Ashton E, Cole PE, Ehman RL. Repeatability of MR Elastography of Liver: A Meta-Analysis. Radiology 2017;285(1):92–100. doi: 10.1148/radiol.2017161398 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Koo TK, Li MY. A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J Chiropr Med 2016;15(2):155–163. doi: 10.1016/j.jcm.2016.02.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kostallari E, Wei B, Sicard D, Li J, Cooper SA, Gao J, Dehankar M, Li Y, Cao S, Yin M, Tschumperlin DJ, Shah VH. Stiffness is associated with hepatic stellate cell heterogeneity during liver fibrosis. Am J Physiol Gastrointest Liver Physiol 2022;322(2):G234–G246. doi: 10.1152/ajpgi.00254.2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Li J, Allen AM, Shah VH, Manduca A, Ehman RL, Yin M, Group NMIR. Longitudinal Changes in MR Elastography-based Biomarkers in Obese Patients Treated with Bariatric Surgery. Clin Gastroenterol Hepatol 2023;21(1):220–222 e223. doi: 10.1016/j.cgh.2021.10.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Barr RG, Ferraioli G, Palmeri ML, Goodman ZD, Garcia-Tsao G, Rubin J, Garra B, Myers RP, Wilson SR, Rubens D, Levine D. Elastography Assessment of Liver Fibrosis: Society of Radiologists in Ultrasound Consensus Conference Statement. Radiology 2015;276(3):845–861. doi: 10.1148/radiol.2015150619 [DOI] [PubMed] [Google Scholar]
- 29.Bravo AA, Sheth SG, Chopra S. Liver biopsy. N Engl J Med 2001;344(7):495–500. doi: 10.1056/NEJM200102153440706 [DOI] [PubMed] [Google Scholar]
- 30.Venkatesh SK, Wells ML, Miller FH, Jhaveri KS, Silva AC, Taouli B, Ehman RL. Magnetic resonance elastography: beyond liver fibrosis-a case-based pictorial review. Abdom Radiol (NY) 2018;43(7):1590–1611. doi: 10.1007/s00261-017-1383-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Aller MA, Blanco-Rivero J, Arias N, Santamaria L, Arias J. The Lymphatic Headmaster of the Mast Cell-Related Splanchnic Inflammation in Portal Hypertension. Cells 2019;8(7). doi: 10.3390/cells8070658 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Mathurin P, Gonzalez F, Kerdraon O, Leteurtre E, Arnalsteen L, Hollebecque A, Louvet A, Dharancy S, Cocq P, Jany T, Boitard J, Deltenre P, Romon M, Pattou F. The evolution of severe steatosis after bariatric surgery is related to insulin resistance. Gastroenterology 2006;130(6):1617–1624. doi: 10.1053/j.gastro.2006.02.024 [DOI] [PubMed] [Google Scholar]
- 33.Dzyubak B, Venkatesh SK, Manduca A, Glaser KJ, Ehman RL. Automated liver elasticity calculation for MR elastography. J Magn Reson Imaging 2016;43(5):1055–1063. doi: 10.1002/jmri.25072 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.