King's College London

Research portal

Robustness of MR Elastography in the Healthy Brain: Repeatability, Reliability, and Effect of Different Reconstruction Methods

Research output: Contribution to journalArticlepeer-review

Siri F. Svensson, José De Arcos, Omar Isam Darwish, Jorunn Fraser-Green, Tryggve H. Storås, Sverre Holm, Einar O. Vik-Mo, Ralph Sinkus, Kyrre E. Emblem

Original languageEnglish
Pages (from-to)1510-1521
Number of pages12
JournalJournal of Magnetic Resonance Imaging
Volume53
Issue number5
DOIs
Accepted/In press2021
PublishedMay 2021

King's Authors

Abstract

Background: Changes in brain stiffness can be an important biomarker for neurological disease. Magnetic resonance elastography (MRE) quantifies tissue stiffness, but the results vary between acquisition and reconstruction methods. Purpose: To measure MRE repeatability and estimate the effect of different reconstruction methods and varying data quality on estimated brain stiffness. Study Type: Prospective. Subjects: Fifteen healthy subjects. Field Strength/Sequence: 3T MRI, gradient-echo elastography sequence with a 50 Hz vibration frequency. Assessment: Imaging was performed twice in each subject. Images were reconstructed using a curl-based and a finite-element-model (FEM)-based method. Stiffness was measured in the whole brain, in white matter, and in four cortical and four deep gray matter regions. Repeatability coefficients (RC), intraclass correlation coefficients (ICC), and coefficients of variation (CV) were calculated. MRE data quality was quantified by the ratio between shear waves and compressional waves. Statistical Tests: Median values with range are presented. Reconstruction methods were compared using paired Wilcoxon signed-rank tests, and Spearman's rank correlation was calculated between MRE data quality and stiffness. Holm–Bonferroni corrections were employed to adjust for multiple comparisons. Results: In the whole brain, CV was 4.3% and 3.8% for the curl and the FEM reconstruction, respectively, with 4.0–12.8% for subregions. Whole-brain ICC was 0.60–0.74, ranging from 0.20 to 0.89 in different regions. RC for the whole brain was 0.14 kPa and 0.17 kPa for the curl and FEM methods, respectively. FEM reconstruction resulted in 39% higher stiffness than the curl reconstruction (P < 0.05). MRE data quality, defined as shear-compression wave ratio, was higher in peripheral regions than in central regions of the brain (P < 0.05). No significant correlations were observed between MRE data quality and stiffness estimates. Data Conclusion: MRE of the human brain is a robust technique in terms of repeatability. Caution is warranted when comparing stiffness values obtained with different techniques. Level of Evidence: 1. Technical Efficacy Stage: 1.

View graph of relations

© 2020 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454