Estimation of fractional myocardial blood volume and water exchange using ferumoxytol‐enhanced magnetic resonance imaging. Issue 6 (31st December 2020)

• Record Type: Journal Article
• Authors: Colbert, Caroline M.; Thomas, Michael A.; Yan, Ran; Radjenovic, Aleksandra; Finn, J. Paul; Hu, Peng; Nguyen, Kim‐Lien
• Is Part Of: Journal of magnetic resonance imaging. Volume 53:Issue 6(2021)
• Page Start: 1699
• Page End: 1709
• Publication Date: 2020-12-31
• Subjects: compartment modeling -- ferumoxytol -- swine model -- T1 mapping -- tissue blood volume; Magnetic resonance imaging -- Periodicals; 616
• DOI: 10.1002/jmri.27494
• Languages: English
• ISSNs: 1053-1807

Abstract

Abstract : Fractional myocardial blood volume (fMBV) estimated using ferumoxytol‐enhanced magnetic resonance imaging (MRI) (FE‐MRI) has the potential to capture a hemodynamic response to myocardial hypoperfusion during contrast steady state without reliance on gadolinium chelates. Ferumoxytol has a long intravascular half‐life and its use for steady‐state MRI is off‐label. The aim of this prospective study was to optimize and evaluate a two‐compartment model for estimation of fMBV based on FE‐MRI. Nine healthy swine and one swine with artificially induced single‐vessel coronary stenosis underwent MRI on a 3.0 T clinical magnet. Myocardial longitudinal spin–lattice relaxation rate (R1) was measured using the 5(3)3(3)3 modified Look‐Locker inversion recovery (MOLLI) sequence before and at contrast steady state following seven ferumoxytol infusions (0.125–4.0 mg/kg). fMBV and water exchange were estimated using a two‐compartment model. Model‐fitted fMBV was compared to simple fast‐exchange fMBV approximation and percent change in pre‐ and postferumoxytol R1. Dose undersampling schemes were investigated to reduce acquisition duration. Variation in fMBV was assessed using one‐way analysis of variance. Fast‐exchange fMBV and ferumoxytol dose undersampling were evaluated using Bland–Altman analysis. Healthy normal swine showed a mean mid‐ventricular fMBV of 7.2 ± 1.4% and water exchange rate of 11.3 ± 5.1 s −1 . There was intersubject variation in fMBV ( p < 0.05) without segmental variation ( p = 0.387). fMBV derived from eight‐dose and four‐dose sampling schemes had no significant bias (mean difference = 0.07, p = 0.541, limits of agreement −1.04% [−1.45, −0.62%] to 1.18% [0.77, 1.59%]). Pixel‐wise fMBV in one swine model with coronary artery stenosis showed elevated fMBV in ischemic segments (apical anterior: 11.90 ± 4.00%, apical septum: 16.10 ± 5.71%) relative to remote segments (apical inferior: 9.59 ± 3.35%, apical lateral: 9.38 ± 2.35%). A two‐compartment model based on FE‐MRI using the MOLLI sequence may enable estimation of fMBV in studies of ischemic heart disease. Level of Evidence: 2. Technical Efficacy Stage: 2.