WE‐C‐116‐06: Reducing the Background Field Variations Using the Geometry Information

Abstract

Purpose: To demonstrate a new approach for reducing background phase variations in the susceptibility weighted image (SWI)[1]. Methods: In order to perform this experiment, we acquired high resolution sagittal 3D SWI images of the left leg. Background phase variations were removed to significantly reduce the phase due to the geometry and improve the data quality. This was done using forward modeling approach to simulate phase generated only due to the geometry of the leg [2–5]. Results: The collected data indicated that the phase effects due to the background are very large, so we used forward modeling approach to remove the unwanted phase variations in the image background. This approach uses a kernel described by the Green's function in k‐space to generate point dipole effects across the tissues using its magnetic susceptibility properties. The data resulting from this technique gave us improved phase images and also helped in collecting water‐fat out of phase information using complex division of flow compensated (FC) 5.2ms and 6.5ms data sets. To further see the potential of this technique, we complex divided the FC 7.8ms and not FC 7.8ms data sets resulting in data with direct flow information. Conclusion: We have demonstrated that using a homodyne high‐pass filter alone does not remove the phase due to the geometry. Using the proposed novel technique we were able to significantly reduce the phase due to the geometry and improve the quality of the data for further analysis of water and fat separation as well as visualizing veins. The remaining weakness of this method is that the veins appear as fat in the muscle part of the image but the next step in this processing is to due susceptibility mapping which will then remove the veins and leave a pristine fat map. We are currently evaluating this approach.