Susceptibility Mapping

From Spin Help
Jump to: navigation, search

In this document, the detailed steps about how to process Susceptibility Mapping program are mentioned. Here, we use a 0.5 mm x 0.5 mm x 0.5 mm dataset as an example with an input matrix size of 512x384x88.

Creating Susceptibility Maps

There are seven steps in the whole process, as follows:

  1. High pass filter the phase images. The filter size can be 32x32 or 64x64.
    For the example dataset both filter sizes, size 32x32 and 64x64, have been tried. Filter size 64x64 did a better job. In general either 32x32 or 64x64 is appropriate for most datasets. You can process the dataset using both filter sizes and then compare the results. The more background phase noises are removed, the better susceptibility maps we can get.
  2. Remove noise regions
    1. Option 1. Use complex thresholding algorithm to remove noise regions
      1. Create new magnitude mask which removed the skull of the brain. This step can be done by using “Process” -> SWI Boundary Detection
      2. Adopt the new magnitude mask to the phase images to remove the skulls in the phase images. This step can be done by “Filter” -> Threshold with mask
      3. Pad phase images to be 512x512. For the example dataset, after step “3.2”, we got phase images with 512x384, we need to pad the images to be 512x512. This can be done by “Process”->Pad Images
      4. Change “0” inside the phase images to be “2048”. In step “3.3”, after padding images, the image size was expanded by filling zeroes. However, in phase images, zero should be 2048, therefore, we have to change those zeroes to be 2048. This can be done by “Filter” –> Filter Images by Threshold.
    2. Option 2. Isolate just the region on interest and set all other phase values to zero (2048 on Siemens data)
      1. Draw a region of interest (ROI) using the drawing tool (link)
      2. Select "Filters"->Change_Intensity_Inside/_Outside_Boundary
      3. Set values outside ROI to zero (2048 on Siemens data)
      4. Apply to whole series
      5. If ROI is significantly smaller than original images the image can be cropped to save processing time and space
      6. Select Preferences_Menu "Crop Entire Series"
      7. Select Preferences_Menu "Position Crop Area"
      8. Select "Process"->Crop_Area
      9. Position crop box on image and double click
  3. Choose the slices that cover the region of interest
    1. Include an additional 4 slices above and below this region (8 additional slices)
    2. Delete all other slices
  4. (Optional) Apply a Hanning filter in the Z direction to the first and last slices to smooth the sharp edges
    1. Select “Process”->Add Hanning Filter to Phase Images
    2. Choose either 4 or 8 slices
    3. Only apply this step if a large number of slices are being processed
  5. Zero Fill images in Z direction. Adds zero slices above and below the current slices to reduce noise aliasing
    1. Select “Process” -> Zero Filling Phase Images
    2. Fill to the next largest power of 2 (8,16,32,64,128,256,etc.)
  6. Run Susceptibility mapping program
    1. Select “Filter” -> 3D Simple Inverse Transform Filter.