Ferumoxytol has gained interest as a positive MR-contrast agent in patients with renal failure, due to the absence of nephrotoxicity concerns [1]. Furthermore, ferumoxytol is an effective blood pool contrast agent with an intravascular half-life of 14-15h, ideal for MRA acquisitions in delayed phases [1,2]. However, care is required with dose and dilution due to thrombus-mimicking artifacts that can arise due to the high transverse relaxivity of the agent [3]. Indeed, different doses of ferumoxytol have been reported when used for MRA [1,3,4]. The purpose of this study was to: a) Qualitatively compare ferumoxytol-enhanced MRA with gadolinium-enhanced MRA, b) Perform a dose comparison/optimization for ferumoxytol with respect to image quality and SNR performance and c) Perform a flip-angle optimization for ferumoxytol-enhanced MRA.

After IRB approval and informed consent, eight healthy volunteers (4 male, 4 female, mean age 47y) were recruited for this study. Scanning was performed on a clinical 3T MRI system (MR750, GE Healthcare). Ferumoxytol (Feraheme, AMAG-pharmaceuticals)-enhanced MRA was performed one month after gadolinium (gadobenate-dimeglumine (Multihance), Bracco)-enhanced MRA, which was acquired as the reference. The one month delay was selected to ensure complete elimination of gadolinium from the body. For the ferumoxytol scan, volunteers were randomly assigned to either 4mg/kg (0.07 mmol/kg, n=4) or 2mg/kg (0.035 mmol/kg, n=4). First pass MRA using a standard spoiled-gradient-echo MRA-sequence was performed with a flip angle (FA) of 50° for both gadolinium- and ferumoxytol-enhanced MRA. After ferumoxytol administration, consecutive MR images were acquired in steady state with decreasing flip angles down to 10°, in increments of 10°, while TR, TE and receiver bandwidth were kept constant. Other parameters for the MRA acquisition included: TR/TE = 4.54/1.40ms, BW = +/- 127 kHz, FOV = 400 x 360 x 320mm3, 224 x 160 x 160 matrix, for true spatial resolution interpolated to 0.8 x 0.7 x 1.1 mm3 through zero-filling. Relative SNR was calculated by dividing the signal intensities of aorta and muscle (psoas muscle). SNR values for gadobenate dimeglumine and the different ferumoxytol-doses were measured and analyzed separately. First pass ferumoxytol- and gadolinium-enhanced MRA images were qualitatively evaluated in a blinded fashion with respect to vessel contrast, artifacts and image quality on a 5-point-scale (5=excellent-1=not interpretable) by two radiologists with 7 and 10 years of experience, respectively.Relative SNR values are summarized in Table 1 (in Figure 1). Relative SNR (aorta/muscle) of the first pass MRA acquisitions was significantly higher with gadobenate dimeglumine compared to both 4mg/kg (p=0.01) and 2mg/kg (p=0.007). 4mg/kg revealed significantly higher SNR than 2mg/kg (p=0.004). In the portal-venous phase, relative SNR was not significantly higher using 4mg/kg than gadobenate dimeglumine (p=0.7) and 2mg/kg (p=0.3). Relative SNR did not decrease with ferumoxytol in the late phase 5 min post injection whereas SNR declined significantly with gadobenate dimeglumine compared to 4mg (p=0.008) and 2mg/kg (p=0.01, Fig.2). Mean values for vessel contrast, artifacts and image quality are summarized in Table 2 (in Figure 1). Qualitative evaluation revealed no significant differences between ferumoxytol- and gadolinium-enhanced MRA for vessel contrast and image quality (Fig.2,3). Less artifacts were reported for ferumoxytol-enhanced MRA (p=0.07). No significant difference in vessel contrast (p=0.18), artifacts (p=0.4) and image quality (p=0.4) was found between different ferumoxytol doses. The flip angle optimization of ferumoxytol-enhanced MRA demonstrated the following results (Fig. 4): With 4mg/kg, relative SNR (aorta/muscle) showed to be highest with a FA=40°(mean: 7.6, StDev: 1.6), with 2mg/kg FA=40° and FA=30° revealed equally high values (mean: 5.2, StDev: 0.45(40°)/0.9(30°)). Ferumoxytol used as MRA-contrast agent based on its T1-shortening properties showed comparable qualitative performance to gadobenate dimeglumine (Fig. 3) and could be utilized without relevant qualitative penalty even with a dose of 2mg/kg. Furthermore, ferumoxytol is advantageous for imaging at later phases due to its long intravascular half-life. SNR with 4mg ferumoxytol/kg was higher than with 2mg ferumoxytol/kg. However, both compared equally well with gadolinium-enhanced MRA. Flip angle optimization revealed slightly different optima for 4mg ferumoxytol/kg (40°) and 2mg ferumoxytol/kg (30-40°).Ferumoxytol for MRA achieved equivalent image quality even with the lower dose of 2mg/kg in this study. In order to achieve optimal results, the flip angle should be adapted to dose.