Three-dimensional MR Elastography Depicts Liver Inflammation, Fibrosis, and Portal Hypertension in Chronic Hepatitis B or C

Abstract

Background The value of measuring mechanical properties to categorize various pathophysiologic states of the liver is as yet undetermined in chronic hepatitis B (CHB) or C (CHC). Purpose To evaluate multiparametric three-dimensional (3D) MR elastography as a means of detecting early necroinflammation, distinguishing necroinflammation from fibrosis, and gauging the severity of portal hypertension (PH) in CHB or CHC. Materials and Methods From January 2015 to September 2019, participants with CHB or CHC were prospectively enrolled from a single institution and were divided into two groups: those with liver biopsy and no evidence of PH (group 1) and those with PH and a hepatic venous pressure gradient (HVPG) measurement (group 2). For group 3, healthy volunteers were separately recruited from a nearby community. Multiple viscoelastic parameters (shear stiffness [SS], storage modulus, loss modulus, and damping ratio [DR]) were determined at 3D MR elastography at 60 Hz, and multivariable logistic or linear regression analysis was used to assess associations of mechanical parameters with histologic scores and HVPG. Results A total of 155 participants (median age, 41 years [interquartile range, 32-48 years]; 85 women) were in group 1 (training set: n = 78, validation set: n = 77), 85 participants (median age, 57 years [interquartile range, 43-61 years]; 51 women) in group 2, and 60 healthy volunteers (median age, 49 years [interquartile range, 27-64 years]; 38 men) in group 3. The liver DR was higher in participants with necroinflammation (DR, 0.13 ± 0.03) versus those without (at liver fibrosis stage F0) (DR, 0.10 ± 0.02; P < .001). Liver DR and SS together performed well in the diagnosis of necroinflammation (area under the receiver operating characteristic curve [AUC], 0.88 [95% CI: 0.79, 0.96]) and the scoring of moderate to severe activity (AUC, 0.88 [95% CI: 0.81, 0.95]) in the validation data set. Liver DR (regression coefficient [β] = -30.3 [95% CI: -58.0, -2.5]; P = .03) and splenic SS (β = 2.3 [95% CI: 1.7, 2.9]; P < .001) were independently associated with HVPG. Conclusion Three-dimensional MR elastography may detect early necroinflammation, distinguish necroinflammation from liver fibrosis, and correlate with hepatic venous pressure gradient in chronic hepatitis B and C. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Reeder in this issue.

Graphical abstract

Figure 1:

Flowchart of study participants. 3D = three-dimensional, CHB = chronic hepatitis B, CHC = chronic hepatitis C, HVPG = hepatic venous pressure gradient, PH = portal hypertension, PVP = portal vein pressure, TIPS = transjugular intrahepatic portosystemic shunt.

Figure 2:

Representative three-dimensional MR elastography images in healthy controls and participants with chronic hepatitis B or C (liver outlined with dotted line). (A) Liver shear stiffness (SS) map in controls. (B) Liver SS maps in patients with liver fibrosis (LF) from stages F0 to F4. (C) Damping ratio (DR) map in controls. (D) DR maps in patients with LF from stages F0 to F4.

Figure 3:

Boxplots show liver shear stiffness (SS) and damping ratio (DR) determinations in participants with liver fibrosis (LF), assessed with three-dimensional MR elastography. The lower and upper borders of boxes correspond to the first and third quartiles (the 25th and 75th percentiles). The midline of boxes indicate medians. The lower and upper whiskers correspond to the 5th and 95th percentiles, respectively. Dots beyond the end of the whiskers indicate outliers. (A) SS increased according to LF stage (all P < .001; adjusted α = .005). (B) DR was consistent across stages F0–F2, with declines at stages F3 and F4 (F0 vs F3 [P = .002], F0 vs F4 [P < .001], F2 vs F4 [P < .001], and F2 vs F3 [P = .003]; adjusted α = .005).

Figure 4:

Boxplots show liver shear stiffness (SS) and damping ratio (DR) determinations in participants with liver fibrosis (LF) and necroinflammation, assessed with three-dimensional MR elastography. The lower and upper borders of the boxes correspond to the first and third quartiles (the 25th and 75th percentiles). The midline indicates the median. The lower and upper whiskers correspond to the smallest and largest values, respectively, no further than 1.5 box lengths from the median. Dots beyond the end of the whiskers indicate outliers. (A, B) Boxplots illustrate liver SS with (A) the A0 versus A1–A3 activity score and (B) A0–A1 versus A2–A3 activity score at each LF stage (P = .04 at A0 [vs A1–A3] for stage F0). (C, D) Boxplots illustrate liver DR with (C) the A0 versus A1–A3 activity score and (D) A0–A1 versus A2–A3 activity score at each LF stage (significance at stage F0 in A0 vs A1–A3 [P < .001] and A0–A1 vs A2–A3 [P = .007] and stage F1 in A0–A1 vs A2–A3 [P = .001]).

Figure 5:

Representative three-dimensional (3D) MR elastography images (liver and spleen outlined by dotted line) and scatterplots of portal pressure with 3D MR elastography–derived parameters. Maps of (A) shear stiffness and (B) damping ratio (DR) in patients with portal hypertension. Scatterplots of (C) hepatic venous pressure gradient (HVPG) with spleen stiffness and (D) HVPG with liver DR. Scatterplots of 3D MR elastography–derived parameters with portal vein pressure are shown in Figure E4 (online).