BioPARR Documentation

Installation

Dependencies

The following programs must be installed to run the BioPARR suite:

3D Slicer (tested with 4.4.0)
Paraview (tested with version 4.3.1)
Abaqus (tested with version 6.14)

These programs have been shipped with the BioPARR binaries but your local version could also be used:

Gmsh (tested with 2.11.0)
ACVD

Installation procedure

Install the required dependencies.
Clone the BioPARR GitHub repository or download the precompiled binaries for Windows.
If using the source code, compile the binaries by following the compilation instructions.
Edit the AAA_Configure file and set up the paths to 3D Slicer and Paraview based on your installation.

Compilation

The steps to compile your own binaries for the Slicer extensions are as follows:

Go to the build instructions for Slicer and follow the steps to install prerequisites and clone Slicer but do not build yet.
In the cloned file Slicer/CMakeLists.txt find Slicer_VTK_COMPONENTS and the line
```
set(Slicer_VTK_COMPONENTS)
```
In the indented section below this line, add these lines:
```
    ${VTK_COMPONENT_PREFIX}FiltersVerdict
    ${VTK_COMPONENT_PREFIX}IOGeometry
```
Finish building Slicer according the instructions.
Change directory to the directory containing the extension you want to build. For example:
```
cd path/to/BioPARR/Software/SlicerModules/AAA_AverageStress
```

Run

cmake -DCMAKE_BUILD_TYPE:STRING=Debug -DSlicer_DIR:PATH=~/Applications/Slicer-SuperBuild-Debug/Slicer-build .
make

Binaries for this extension should now be available.

Analysis Configuration

Edit AAA_AnalyseCase.bat to change the analysis configuration. By default, both variable (if data for the variable thickness exists) and constant thickness geometries are created and analysed using three scenarios.

For example, if you don’t want variable thickness geometry, change

SET AAA_ANALYSE_VAR_THICKNESS=0

at the start of the script. If you don’t want the constant thickness scenario to be analysed, change

SET AAA_ANALYSE_CONST_THICKNESS=0

at the start of the script.

The scenarios to be analsed are configures in ./Software/Scripts/AnalyseAllCases.bat. If you want to remove some of these scenarios from analysis, change the definition of the variable

SET AAA_ANALYSIS_CASE=NoILT,ILTPressure,WallPressure

at the top of that file.

The actual configuration for each of the scenarios can be found in the Abaqus input file ./Software/Scripts/Abaqus/AAA.inp. You can modiy these scenarios or add additional scenarios that you want to analyse by creating additional subfolders and Abaqus input files.

The 3D Slicer modules used for analysis are found in ./Software/SlicerModules. You can run these with the --help parameters to find the different configuration options they have and then change their command line parameters. For exmaple, the module for creating the Abaqus input files corresponding to the wall and ILT parts is called from AAA_AnlyseCase.bat with the options --quad --hybrid which means it created hybrid tetrahedral elements with quadratic shape functions and 10 nodes. This may lead to very long analysis time so these options can be removed and linear tetrahedral elements used instead.

Case Analysis

Setup

Duplicate the folder CaseID and rename it with the ID of the case you are going to analyse. Once this has been created, change the working directory to this folder and run AAA_AnalyseCase.bat. This will set up the directory structure for you.

For example, if we were analysing Case11220:
```
cp -r CaseID Case11220
cd Case11220
AAA_AnalyseCase.bat
```
Check the data

Go to the shared data drive and copy the folders with the CT and MRI images you will use for this case. Open them in 3D Slicer and check if appropriate for analysis.
Cropping

Open the CT image in 3D Slicer, define the region of interest (ROI) and crop the volume (Converters -> Crop Volume) with isotropic output voxels. Save result in ./1_Segmentation_CT/CT_cropped.nrrd.

Open the MRI image in 3D Slicer, define the region of interest (ROI) and crop the volume (Converters -> Crop Volume) with isotropic output voxels. Save result in ./2_Segmentation_MRI/MRI_cropped.nrrd.

Note: The ROI should start just under the renal arteries nd end under the bifurcation of the iliac arteries (if visible in the images).
Segmentation
1. Lumen
  
  Open ./1_Segmentation_CT/CT_cropped.nrrd in 3D Slicer. Go into Volumes and under Window/Level select CT-abdomen. Go into Segment Editor and use the Threshold effect to create a rough segmentation of the blood channel. Use Islands to eliminate unconnected regions. Use Paint to make manual corrections (eliminate small veins, calcification the may be in contact with the blood channel, etc). The changes can be checked by using Surface Models -> ModelMaker to create a surface representation of your segmentation.
  
  Once you are satisfied with your manual corrections use Surface Models->Label Map Smoothing with a Gaussian smoothing parameter Sigma of 1 (set both input and output volumes as CT_cropped-label) to smooth out the segmentation. Save result to ./1_Segmentation_CT/CT_blood_label.nrrd.
  
  Note: Make sure the segmentation extends the entire height of the cropped CT image by checking the first and last red slices in 3D Slicer before saving the segmentation results.
2. AAA
  
  Use Paintto extend the blood channel label to cover the entire AAA. Use a different label to mark the region around the AAA. Use FastGrowCutEffect to segment the AAA. Once satisfied with the result stop FastGrowCutEffect, use ChangeLabelEffect to change the label used to mark the region around the AAA to background colour and use Paint to make manual corrections.
  
  Once you are satisfied with your manual corrections use Surface Models->Label Map Smoothing with a Gaussian smoothing parameter Sigma of 1 (set both input and output volumes as CT_cropped-label) to smooth out the segmentation. Save result to ./1_Segmentation_CT/CT_AAA_label.nrrd.
  
  Note: Make sure the segmentation extends the entire height of the cropped CT image by checking the first and last red slices in 3D Slicer before saving the segmentation results.
  
  Tip: Use a sphere in Paint to mark regions in several slices at once. Take care not to mark incorrect regions in the slices you don’t see.
3. AAA from MRI
  
  Use the same procedure as above to segment the AAA from the MRI. The lumen does not need to be segmented separately in this case. Save results to ./2_Segmentation_MRI/MRI_AAA_label.nrrd.
Wall extraction and MRI to CT registration

Run AAA_AnalyseCase.bat from the command prompt. The script will do this step automatically based on the segmentation data.
Wall thickness extraction

Open ./1_Segmentation_CT/CT_cropped.nrrd and ./2_Segmentation_MRI/MRI_aligned_to_CT.nrrd (created by the script) in 3D Slicer. If you set the as Background and Foreground images you can blend them together and check they are aligned. Look for areas where the wall is visible in the MRI (or calcifications in CT) and use the ruler to measure the wall thickness. Save measurements in ./3_Thickness/M.acsv.
Complete the analysis

Run AAA_AnalyseCase.bat from the command prompt. The script will create the AAA and ILT surfaces, mesh the AAA and ILT volumes, create the Abaqus input files, run the Abaqus analysis for the prescribed loading configurations, extract the stress in the wall and compute the rupture potential index (RPI).
Check the results

If there are no errors during the script execution, open some of the result files and check the output.

If there are errors, you must solve them so the analysis can progress further (this may involve cleaning up the segmentations or using different parameters for meshing). Then rerun AAA_AnalyseCase.bat

If you cannot solve your error on your own please feel free to submit an issue here.

Tutorials

Tutorials describing the different steps of the analysis can be found here.