The Virtual Pediatric Airways Workbench (VPAW) is an application, based upon 3D Slicer, for creating and visualizing anatomically accurate, three-dimensional airway models for geometric and computational fluid dynamics analyses. Here we describe two of its modules, VPAW Model and VPAW Visualize.

The VPAW Model module provides a graphical user interface so that the user can easily install, configure, and run the pipelines from the pediatric_airway_atlas repository. The pipeline starts with one CT image and one set of landmarks per patient and produces a variety of additional images, landmarks, segmentations, areas, centerlines, and Laplace transforms. These outputs can be used in comparisons to atlas data.

The VPAW Visualize module facilitates the visualization of results produced by the pediatric_airway_atlas pipelines. This module allows researchers to visualize and overlay intermediate and final outputs for a patient.

The VPAW Model module

When you start VPAW you start on the Home screen. To run the Pediatric Airway Atlas pipeline, first click on the "Go to VPAW Model module" button.

You need to specify the directory containing the source code for the Pediatric Airway Atlas model. Click on the "..." button that is to the right of "Pediatric Airway Atlas source directory."

Use the dialog box to select the directory that contains your source code. This is the top-level directory of the uncbiag/pediatric_airway_atlas git repository on your local computer. Click the "Choose" button.

You are returned to the VPAW Model screen. Click on the "Link to Pediatric Airway Atlass codebase and install dependencies" button to tell VPAW that you want to use the selected source code and install any required Python packages. Note that these packages need to be installed only once and they are installed to a Python environment that is internal to VPAW; they are not installed to Python environments that you use outside of VPAW.

If there are missing Python dependencies, you see a dialog box like the following. Click the "Yes" button.

Upon completion of the installation of the Python dependencies, you are alerted to the version numbers of the successfully installed Python packages. Click the "OK" button.

Whether or not there were dependencies that needed to be installed, you are notified that VPAW is now connected to the Pediatric Airway Atlas source code that you supplied. Click the "OK" button.

Next you need to specify the directory that contains your inputs to be processed by the pipeline. This directory has a Microsoft Excel file called FilteredControlBlindingLogUniqueScanFiltered.xls and two subdirectories, images and landmarks. The images subdirectory has one PPPP\_INPUT.nrrd file per patient, where PPPP is a four-digit patient number. The landmarks subdirectory has one PPPP\_LANDMARKS.fcsv file for each of the same patient numbers.

The FilteredControlBlindingLogUniqueScanFiltered.xls file of clinical data by patient and the image files must be supplied. The landmarks files must be pre-computed prior to running these pipelines. The directory we are using in this example has inputs for two patients, 1127 and 1129.

Click on the "..." button that is to the right of "Input/Output Root Directory."

Use the dialog box to select the directory that contains your inputs. Click the "Choose" button.

You are returned to the VPAW Model screen. Next you need to specify the directory that contains the model parameters that the pipeline should use when processing the input data. The directory should contain a single file with a name like "116(158.10-38.AM.24.Mar).pth".

Click on the "..." button that is to the right of "Models Directory."

Use the dialog box to select the directory that contains your model file. Click the "Choose" button.

You are returned to the VPAW Model screen.

Optionally specify a patient number. If you leave this blank, all patients are processed. In this example, we specify patient 1127. Whether or not you specify a patient number, click on the "Run Pediatric Airway Atlas" button.

After the Pediatric Airway Atlas pipeline has completed, you see a dialog box indicating that. Click on the "OK" button.

You are returned to the VPAW Model screen. An examination of the input/output directory now shows both the inputs and the newly generated outputs.

The VPAW Visualize module

To look at the inputs and outputs, we use the VPAW Visualize screen. Click on the "Go to VPAW Visualize module" button near the upper left corner.

You are taken to the VPAW Visualize screen. To visualize the data just analyzed by the pipeline or any previous dataset, we enter the location of the dataset. Click on the "..." button that is to the right of "Data directory."

Use the dialog box to select the directory that contains your data to be visualized. Click the "Choose" button.

You are returned to the VPAW Visualize screen.

Specify the patient number for the data you wish to visualize. In this example, we specify patient 1127. Click on the "Show" button.

The VPAW Visualize page now lists a number of data features that can be visualized. Click on the "Compute Isosurfaces" button to generate a model node consisting of isosurfaces of the Laplace solution image.

Notice the letters "S" (red), "A" (green), and "L" (yellow) to the right of the corresponding sliders in the main viewing panels, which indicate superior, anterior, and left. You adjust the red, green, and yellow sliders to specify which axial plane (inferior to superior), coronal plane (posterior to anterior), and sagittal plane (right to left) of the data are displayed. Toggle the visibility of each element in the scene by toggling the eye icon in the second column of the list of inputs and outputs.