학술논문

Validation of the Performance of MRE for the Detection of Advanced Fibrosis due to NASH across Multiple Clinical Trials
Document Type
Article
Source
춘·추계 학술대회(The Liver Week). Aug 13, 2020 2020(1):249
Subject
Magnetic resonance elastography
MRE
Advanced fibrosis
NASH
Language
Korean
Abstract
Aims: Magnetic resonance elastography(MRE) is a quantitative imaging biomarker for the detection of advanced fibrosis due to NASH. Our aim was to validate the performance of MRE for detection of advanced fibrosis using data from multiple clinical trials. Methods: Baseline data were pooled on 296 subjects with NASH from seven randomized, phase 2 and 3 trials including ATLAS and STELLAR-3/-4 trials. All subjects underwent 2D MRE and liver biopsy with staging of fibrosis according to NASH CRN classification. Associations between MRE-stiffness, fibrosis stage, noninvasive tests (NITs) of fibrosis (ELF, FibroTest, FIB-4, NAFLD Fibrosis Score [NFS]), and NASH activity (NAFLD Activity Score ≥5 vs <5) were determined. The discrimination of MRE for advanced fibrosis (F3-F4 vs F0-F2) and cirrhosis (F4 vs F0- F3) was evaluated using areas under receiver operating characteristic (AUROC) curves, and cutoffs from literature-based MRE thresholds (3.64 and 4.67 kPa, respectively) and optimal thresholds(defined by the maximal sum of sensitivity and specificity) were determined. Results: Among 296 subjects, fibrosis stages were F0-1(6%), F2(11%), F3(44%), and F4(40%); median MRE-stiffness was 4.71 kPa(IQR 3.52, 6.35). MRE-stiffness was correlated with fibrosis stage(Spearman ρ=0.60), and other NITs of fibrosis(ρ =0.47-0.52; all P<0.05). The AUROCs(95% CI) of MRE-stiffness for detecting advanced fibrosis and cirrhosis were 0.85(0.80, 0.90) and 0.81(0.76, 0.86), respectively. In general, cutoffs from the literature and optimal cutoffs derived from this dataset had similar performance for classification of fibrosis by 2D MRE. Among subjects with F0-F2 fibrosis on biopsy, those with MRE-stiffness ≥3.64 kPa (potential misclassification) had higher ELF and FibroSure than those with MRE-stiffness <3.64 kPa (both P<0.05). Conversely, among subjects with F3-F4 fibrosis on biopsy, those with MRE-stiffness <3.64 kPa had lower ELF, FibroSure, FIB-4, and NFS compared with those with MRE-stiffness ≥3.64 kPa(all P<0.05). Conclusions: This multi-center, multi-study validation demonstrates the clinical utility of 2D MRE for the detection of advanced fibrosis due to NASH.

Online Access