Cancer Imaging Phenomics Toolkit (CaPTk)  1.6.2.Alpha
Brain Cancer: Directionality Estimator

This application estimates the volumetric changes and their directionality for a given ROI across two timepoints [1] and also returns the projections of the boundary point with the maximum distance from a seedpoint (with label TU) in each of the 3 2D visualized planes.

REQUIREMENTS:

1. The image of timepoint 1 and a tissue seedpoint with label TU within the ROI at timepoint 1.
2. Two ROI files, one for each timepoint (note that the populated voxels of the ROIs should be of Label 1).

USAGE:

1. Load the image of timepoint 1 in CaPTk and initialize the (or load an already initialized) seedpoint with label TU defining the point within the ROI from where you want to estimate the expansion.
2. Launch the Directionality Estimator application using the 'Applications -> Directionality Estimator' menu option.
3. Specify the Input ROI files for the 2 timepoints and the Output directory where the output masks and text files will be saved.
4. Click on 'Confirm" and a pop-up window will be displayed showing the total volumetric change, as well as the change per 3D octant.
5. A parametric map is automatically loaded as an image in the visualization panels of CaPTk dividing the ROI of timepoint 1 in octants and assigning a value in the range [0,1] on each octant. Smaller and larger values represent the octants with smaller and larger relative volumetric change.
6. A multi-label map is also automatically loaded as an ROI in the visualization panels of CaPTk depicting with i) label 1 (red) and label 3 (yellow) the intersection of the timepoint 1 and timepoint 2 ROIs, ii) label 3 (yellow) the octant of maximum expansion, and iii) label 2 (green) the voxels of expansion in timepoint 2, i.e. the voxels that did not exist in the timepoint 1 ROI but only in the ROI of timepoint 2. Note that the visualization of the multi-label map can be controlled from the "Drawing" panel.
7. Tissue point table gets updated with 3 points marked "CSF" (which are correspond to the direction of max point along the different planes) and 1 point marked "RTN" (which is the actual 3D point of max distance from initialized seed point).
8. The results as well as the actual distances are automatically saved in a text file in the Output folder together with the multi-label ROI images.
9. The user can save both the actual and the projection points of max distance in a location of their choosing.
• This application is also available as with a stand-alone CLI for data analysts to build pipelines around, using the following example command:
DirectionalityEstimate.exe -l C:/labelMap.nii.gz -o C:/output.txt -i 40,25,60


OUTPUT: The following files are saved in the output folder:

• directionalityOutput.txt, which includes the total volumetric change and the change per 3D octant (as shown in the output window), as well as the coordinates of the point of the largest distance (i.e., RTN) from the TU seedpoint together with its projections in the 3 2D planes (i.e., CSF).
• roi_DE_octantImage.nii.gz, which describes the partitioned version of timepoint_2_ROI in octants with the corresponding numbering of octants accompanying the numbering included in the output text file.
• roi_DE_visualizationRatio.nii.gz, which describes the ratio of the timepoint_2_ROI over timepoint_1_ROI of the volumetric differences per octant.
• roiDE_visualizationProbability.nii.gz, which describes the parametric map dividing the ROI of timepoint 1 in octants and assigning a value in the range [0,1] on each octant. Smaller and larger values represent the octants with smaller and larger relative volumetric change. This is the same result as roi_DE_visualizationRatio.nii.gz but scaled between 0 and 1.
• roi_DE_visualizationIncrease.nii.gz, which describes the multi-label map depicting with i) label 1 (red) and label 3 (yellow) the intersection of the timepoint 1 and timepoint 2 ROIs, ii) label 3 (yellow) the octant of maximum expansion based on distance and not based on volumetric difference, and iii) label 2 (green) the voxels of expansion in timepoint 2, i.e. the voxels that did not exist in the timepoint 1 ROI but only in the ROI of timepoint 2. Note that the visualization of the multi-label map can be controlled from the "Drawing" panel.

Reference

1. M.E.Schweitzer, M.A.Stavarache, N.Petersen, S.Bakas, A.J.Tsiouris, C.Davatzikos, M.G.Kaplitt, M.M.Souweidane, "Modulation of Convection Enhanced Delivery (CED) distribution using Focused Ultrasound (FUS)", Neuro Oncol. 19(Suppl 6):vi272, 2017. DOI:10.1093/neuonc/nox168.1118