Evaluation of liver fibrosis using diffusion-weighted virtual magnetic resonance elastography and ultrasound elastography

Abstract

Introduction: This study evaluates the effectiveness of virtual magnetic resonance elastography (VMRE), a new diffusion-weighted imaging (DWI)-based method, for detecting liver fibrosis, comparing it with the more accessible ultrasound elastography (USE).

Materials and methods: This prospective study enrolled patients with chronic liver disease who were referred for liver biopsy. Inclusion criteria were: Sepanlou (Lancet Gastroenterol Hepatol 5:245-266, 2020) clinical indication for liver biopsy and Rinella (Journal of Hepatology 79:1542-1556, 2023) eligibility for MRI. Exclusion criteria included: Sepanlou (Lancet Gastroenterol Hepatol 5:245-266, 2020) MRI contraindications, Rinella (Journal of Hepatology 79:1542-1556, 2023) hepatic iron overload, D'Amico (Hepatology International 12:34-43, 2018) clinical or laboratory evidence of acute hepatitis or cholestasis, and Terrault (Hepatology (Baltimore, Md) 63:261, 2016) inadequate image quality (motion artifacts, low signal-to-noise ratio). All patients underwent 3T VMRE (b = 200/1500 s/mm²) and two-dimensional shear wave elastography (2D-SWE). VMRE was analyzed by two blinded readers; USE by a single radiologist. Using METAVIR staging (F0-F4) as reference, fibrosis was categorized as F0-1 vs. F2-4 and F0-2 vs. F3-4. Statistical analyses included ICC, Bland-Altman, Kruskal-Wallis with Bonferroni-corrected Dunn tests, and ROC analysis. An HBV subgroup (n = 33) and a non-HBV group (n = 16; including metabolic dysfunction-associated steatotic liver disease (MASLD), autoimmune, and toxic hepatitis) were analyzed separately to assess VMRE performance across different etiologies.

Results: Initially, 59 patients were enrolled. After excluding 10 patients due to MRI contraindications, hepatic iron overload, or inadequate image quality, 49 patients were included in the final analysis (mean age 48.2 ± 14.9 years; 28 males, 21 females; 67% HBV-positive). VMRE demonstrated significant limitations in clinical utility, failing to discriminate fibrosis stages in the overall cohort (AUC 0.45-0.51, p > 0.05). While HBV-infected patients showed some promise with an overall significant variation across stages (p = 0.004), post-hoc analysis revealed VMRE could only distinguish between the extreme ends of the fibrosis spectrum (F0 vs. F4: adjusted p = 0.0058). This restricted diagnostic capability was reflected in the HBV subgroup's modest AUC values of 0.75-0.76, which remained below clinical acceptability thresholds. In striking contrast, ultrasound elastography exhibited robust performance across all analyses. It achieved excellent diagnostic accuracy (AUC 0.86-0.95) with highly significant p-values (< 0.001) for all fibrosis classifications, along with clinically practical threshold values (8.85-10.1 kPa). Inter-rater agreement for VMRE was excellent (ICC = 0.972), and intra-rater agreement for USE was good (ICC = 0.756).

Conclusion: VMRE demonstrates insufficient diagnostic accuracy for fibrosis staging in both HBV and non-HBV populations, with only limited ability to distinguish extreme fibrosis stages (F0 vs. F4) in HBV patients. While showing excellent technical reproducibility (ICC = 0.972), its poor discriminative performance (AUC 0.45-0.76 across groups) and inability to differentiate intermediate stages preclude clinical utility. In contrast, USE achieved consistently superior diagnostic accuracy (AUC 0.86-0.95) with practical threshold values, supporting its preference over VMRE especially in centers lacking access to standard MRE. Further VMRE development requires technical optimization and larger validation studies.

Keywords: Chronic liver diseases; Diffusion magnetic resonance imaging; Elasticity imaging techniques; Fibrosis.