Enhanced T1 sensitivity and reduced artifacts in ferumoxytol-enhanced CMRI at 0.55T compared to 3T: a preclinical in vivo study

Background: Ferumoxytol-enhanced (FE) magnetic resonance imaging (MRI) has gained increasing interest due to its versatility across applications such as MR angiography, 4D flow imaging, tumor characterization, and myocardial assessment^1-6. Despite its diagnostic potential, clinical adoption remains limited by factors including high cost and the risk of hypersensitivity reactions, leading to strategies that reduce contrast dose without compromising image quality^7,8. Prior in vitro studies suggest that iron oxide nanoparticles like ferumoxytol may exhibit enhanced relaxivity at lower magnetic field strengths, potentially improving efficacy at reduced doses^9-11. However, direct in vivo comparisons of ferumoxytol performance across different field strengths are lacking. In this study, we compare the T1 sensitivity to ferumoxytol and artifact severity at 3T vs 0.55T using large animal models and commercially available MRI platforms.

Methods: We conducted a preclinical study using domestic swine scanned at 3T (n=18) and 0.55T (n=23) on commercial MRI systems under a protocol approved by local IACUC. Standard MOLLI T1 maps were acquired before and after slow infusion of ferumoxytol at doses up to 4 mg/kg. We defined T1 sensitivity as “the percent change in T1 after ferumoxytol infusion” (for blood and myocardium, separately). For each dose, myocardial and blood regions of interest (ROIs) were analyzed for T1 sensitivity. The susceptibility/banding artifact severity on the T1 maps was independently scored by two expert readers blinded to the field strengths using a 0–3 scale (0=no artifacts, 3=severe artifacts). The T1 sensitivity and artifact severity at two field strengths were statistically compared using student’s t-test.

Results: As shown with the representative cases in Figure 1, T1 maps showed progressive T1 shortening with increasing ferumoxytol dose at both field strengths, as expected. However, 0.55T demonstrated significantly higher T1 sensitivity (ΔT1) in both myocardium and blood for the same ferumoxytol dose. As shown in Figure 2a, myocardial T1 sensitivity to ferumoxytol was linear and nearly 2.5-fold greater at 0.55T compared to 3T (p < 0.01). Blood T1 sensitivity showed a nonlinear trend at both field strengths but was also significantly higher at 0.55T than 3T (p < 0.01) as demonstrated in Figure 2-b. Artifact severity scores were significantly lower at 0.55T (0.10 ± 0.35 vs. 1.64 ± 1.11 at 3T, p < 0.01), with reduced susceptibility/banding artifacts evident in raw T1-weighted images and the T1 maps (Figure 3).

Conclusion: FE CMR at 0.55T offers superior T1 sensitivity and reduced susceptibility artifacts compared to 3T. These findings suggest that low-field MRI can enable lower contrast doses while maintaining diagnostic performance, potentially reducing cost and improving safety. This work supports broader adoption of low-field CMRI for ferumoxytol-based cardiovascular imaging.

*HB Unal and S Zeynali contributed equally to this work.

Figure 1: Representative in vivo data comparing myocardial and blood T1 sensitivity to ferumoxytol at 3T (top row) and 0.55T (bottom row). T1 maps acquired at baseline and after two different ferumoxytol doses illustrate the quantitative impact of contrast on T1 values in both tissue types. Numerical annotations indicate T1 values for myocardium (blue ROIs) and blood (green ROIs), along with calculated T1 sensitivity (ΔT1) for each dose. At 1.0 mg/kg, myocardial T1 sensitivity was 15% and blood T1 sensitivity was 71% at 3T, versus 28% and 88% respectively at 0.55T. At 1.7 mg/kg, 3T yielded 21% myocardial and 80% blood T1 sensitivity, while 0.55T achieved 39% and 94% respectively. These results highlight the stronger T1 response at 0.55T for equivalent doses, with more pronounced changes in both myocardial and blood T1 values compared to 3T.
Figure 2: Overview of in vivo findings comparing myocardial and blood T1 sensitivity across all subjects (0.55T: n = 23; 3T: n = 18). Each point represents T1 sensitivity at different ferumoxytol doses, with two doses per animal (in addition to the native baseline). (a) Myocardial T1 sensitivity exhibited a linear dose-dependent increase at both field strengths, with 0.55T (red) showing a significantly steeper slope—over twice that of 3T (blue) (p < 0.01). (b) Blood T1 sensitivity followed an exponential trend with dose at both field strengths, and again, 0.55T demonstrated markedly higher sensitivity than 3T for equivalent doses (p < 0.01).
Figure 3: Evaluation of susceptibility and banding artifacts in MOLLI T1 maps at 0.55T versus 3T. (a) Artifact severity was scored by two blinded readers across 108 images using a 0–3 scale (0 = no artifacts, 3 = severe artifacts). The average scores were significantly lower at 0.55T (0.10 ± 0.35) compared to 3T (1.64 ± 1.11), indicating reduced artifact burden at the lower field strength (p < 0.01). (b) Sample T1-weighted bSSFP images (left) and corresponding MOLLI T1 maps (right) acquired at 3T (top) and 0.55T (bottom) illustrate this difference. Yellow arrows highlight susceptibility and banding artifacts visible at 3T but absent at 0.55T, underscoring the advantage of low-field imaging in minimizing artifact-related degradation.