Depiction of SPIO in atherosclerotic plaque using true resolution SGM

Introduction: Detection of atherosclerotic plaque with iron oxide contrast agents (e.g. SPIO) that induce a susceptibility effect in MRI is sometimes hindered by competing sources of negative contrast. There is also the characteristic blooming artefact associated with dephasing in the presence of SPIO. Solutions utilising MR sequences to convert the hypo-intensities into a positive contrast often require refinement of the scan parameters and are in addition to any negative contrast scans. Utilising the off-resonance effect of an iron oxide contrast agent, saturation of the on-resonant signal using the IRON technique has been used in detection of atherosclerotic plaque [1]. Similarly the local magnetic field gradients in the presence of iron oxide have been exploited by GRASP to indentify plaque with positive contrast [2]. Although these positive contrast methods aid in detection, hyper-intensities are often over-emphasised and may thus fail to accurately depict the plaque. Post-processing methods also exist to produce positive contrast images by mapping the effect of susceptibility-like field gradients in gradient-echo acquisitions [3-4]. The technique of susceptibility gradient mapping (SGM) uses a “short term” FT i.e. a FT using a small finite extent in k-space resulting in a map that arguably has a lower resolution [3]. Positive contrast images have more recently been calculated at the same resolution by looking at the signal intensity drop associated with successive truncation of k-space lines to determine any echo-shifting effect [4-5]. However such a truncation and consequent zero filling results in Gibb’s ripple artefacts that may affect mapping of the echo-shift. We present a modification of the method outlined in [4] that uses a filter to provide a consistent truncation line-by-line. This method can be seen not only to provide a more robust shift determination since it does not require additional reconstruction algorithms, but also producing positive contrast images for better localisation of susceptibility sources as it retains the original resolution of the acquired data. Theory: Susceptibility induced field inhomogeneities alter the imaging gradients applied during a gradient echo acquisition. Such local field inhomogeneities are also created in the presence of SPIO and thus result in a displacement of the associated echoes in k-space. Determining the shift in k-space thus allows a parametric map to be created that relates the local susceptibilityinduced field gradient to its corresponding pixel. SGM utilises short-term FTs at each pixel across its neighbours to determine the shift. It therefore requires a compromise between the spatial resolution and the accuracy to which the shift is expressed. Our alternate technique for locating this shift involves applying a filter based on a 1-Lorentzian function applied across, for instance, the kx-direction. The Lorentzian function (i.e. of the form 1/[(1/T2 )+(kx-kxo)]) used to form the filter is based on a relatively long T2 * value representing minimal line broadening and then applied in a sliding window fashion so that only a single line of defined kx-value is completely at zero. This continues line-by-line moving from –kmax to +kmax so that only a single line is completely at zero and thus producing (2*kmax)+1 images. Plotting the intensity of each pixel from these images against the kx line that is truncated to zero shows a drop at the k-space centre. In the case of voxels influenced by SPIO that have their echo-shifted along the kxdirection, the intensity minima provides a measure of the shift (Fig.1). For a 3D dataset, this procedure is repeated along all spatial dimensions to produce a parameter map of each pixel’s related shift in k-space. A magnitude image of this map therefore positively highlights susceptibility changes within the original dataset. Materials and Methods: ApoE-KO mice were fed a high-fat diet for 8 weeks to induce a model for atherosclerotic plaque around the brachiocephalic artery [6]. At day 0 and day 1 they