GEH121333 are preclinical-phase ultrasmall superparamagnetic iron oxide (USPIO) particles (GE Global Research, Niskayuna, NY, USA, supplied through GE Healthcare AS, Oslo, Norway) with a high r₁/r₂ relaxivity ratio compared to other USPIOs [1], opening the possibility of both T₁- and T₂(*) contrast-enhanced MRI. The aim of this study was to investigate the use of the GEH121333 particles for susceptibility contrast MRI to detect changes in blood vessel morphology (blood volume, vessel size and vessel density) in xenograft tumors after bevacizumab treatment. The results were compared to DCE-MRI data. In addition, R₁ maps were generated to explore dual mode use of the GEH121333 contrast agent.

All experimental procedures involving animals were approved by the institutional ethics committee and were in accordance with national, regional and institutional guidelines. TOV-21G ovarian cancer xenografts were grown on the hind leg of athymic mice (n=18). MRI was performed on a 7T Bruker Biospec with an 86mm volume resonator for RF transmission and a quadrature mouse brain surface coil for reception. Images of 4 sagittal slices with slice thickness=0.6mm, FOV=20×20mm² and matrix=64×64 were acquired for each tumor using the following sequences:
Multi-echo Spin Echo (MSE): TE=10.5ms, 32 echoes with 10.5ms echo spacing, TR=3s, 2 averages;
Multi-echo Gradient Echo (MGE): TE=2.5ms, 30 echoes with 2.5ms echo spacing, flip angle=30°, 1 average;
Variable TR RARE (VTR): TE_eff=13ms, TR=225/500/1500/3000/6000/12000ms, RARE factor=2, 1 average.
High resolution T₂w images with matrix=256×256 were acquired for drawing tumor ROIs and 3D RARE images were acquired for tumor volume measurements.
The timeline for the experiment is shown in Fig. 1. GEH121333 was injected i.v. at a dose of 10mg Fe/kg. Immediately after imaging on day 0 and again on day 3, the mice received 5mg/kg bevacizumab (Avastin®, Roche AG, Basel, Switzerland) (n=9) or saline (vehicle control, n=9). R₁, R₂ and R₂* maps were computed voxelwise from the VTR, MSE and MGE images, respectively. Pre-contrast R₁, R₂ and R₂* maps were subtracted from the post-contrast maps to obtain ΔR₁, ΔR₂ and ΔR₂* maps. ΔR₂ and ΔR₂* relaxation rates are measures for the microvascular and total blood volume fraction, respectively [2]. Estimates of blood vessel density (Q) and size (R) were calculated from ΔR₂ and ΔR₂* according to Q= ΔR₂/ (ΔR₂*)^2/3 [3] and R= ΔR₂*/ ΔR₂ [4]. On day 6, DCE-MRI was performed using gadodiamide (Omniscan™, GE Healthcare, Oslo, Norway) with the same scan geometry as described above and analyzed using the extended Tofts model as described previously [5]. The area under the enhancement curve after 1 minute (AUC_1min) and the volume transfer constant (K^trans) were computed.
From the manually drawn tumor ROIs a three-voxel-wide (~1mm) tumor rim was automatically extracted and parameter medians were computed for these tumor rims for statistical analysis. Two-tailed independent sample t-tests were performed to compare changes in susceptibility contrast MRI parameters and DCE-MRI parameters in control versus treated tumors. In cases of failed contrast agent injection, the mice were excluded from statistical analyses (number of animals are indicated in the figures).

Bevacizumab had no effect on tumor growth (p=0.68) during the treatment period, but caused a decrease in ΔR₂ (p=0.015), ΔR₂* (p=0.035) and Q (p=0.016), but not in R (p=0.312) (Fig. 2 and 3). Additionally, changes in vascular function after treatment were detected by DCE-MRI, which showed significantly lower AUC_1min (p=0.04) and K^trans (p=0.004) compared to the control group. GEH121333 injection resulted in a signal decrease in the tumor rim in T₂w images, but also a signal increase in T₁w images (Fig. 4), demonstrating the potential utility of these USPIOs for positive contrast-enhanced MRI. The decrease in ΔR₁ after treatment as shown in Fig. 2 is very similar to the decrease in ΔR₂. Also the longitudinal development of R₁ and R₂ (Fig. 5) was very similar. Already on day 1, the R₁ and R₂ were lower in treated compared to control tumors, which may reflect early treatment-induced differences in extravasation and accumulation of GEH121333. This trend was even more pronounced on day 6. Our data shows that the GEH121333 particles can be used for detecting changes in vascular morphology after anti-angiogenic treatment similar to other studies [6, 7]. The combination of DCE-MRI and USPIO-MRI allowed us to describe both changes in vascular function and morphology, thereby providing a more comprehensive assessment of anti-angiogenic treatment effects. The high r₁/r₂ relaxivity ratio of GEH121333 also shows promise for its use as a positive contrast agent for other applications such as combined susceptibility contrast MRI and DCE-MRI using only a single contrast injection.