This application provides a fully automatic segmentation of lung nodules and prediction of survival and nodal failure risks as a three step workflow[1].

REQUIREMENTS:

CT image
PET image (co-registered to the CT image)
Additional requirements for each of the three steps(Lung Field Segmentation, Lung Nodule Segmentation, Prognostic Modeling) as described in Usage.

USAGE:

Step 1: Lung Field Segmentation

Load the required co-registered CT and PET images.
Load the optional mask image (for example, body mask) within which the lung field will be extracted.
Launch the segmentation step using the 'Applications -> Lung Field Segmentation' menu.
The lung field is automatically generated and displayed.

This application is also available as with a stand-alone CLI for data analysts to build pipelines around, using the following example commands:
```
${CaPTk_InstallDir}/bin/SBRT_LungField.exe -p C:/PET.nii.gz -c C:/CT.nii.gz -o C:/outputBasename -m C:/foregroundMask.nii.gz # optional mask
```

It will generate the lung field mask image with name C:/outputBasename_lf.nii.gz. The mask image will contain 2 labels, label 3 for foreground and label 2 for lung field.

Step 2: Lung Nodule Segmentation

Load the required co-registered CT and PET images.
Load the required mask image(for example, lung field mask or a region containing the nodule within lung field with default label value of 2) within which the lung nodule will be extracted.
Load the optional seed image containing foreground seeds(for nodule) and background seeds for nodule segmentation.The foreground seeds should be with label 2, background seeds with label 1, and others with label 0.
Specify the label value for the foreground seeds(default value of 2).
Launch the segmentation step using 'Applications -> Lung Nodule Segmentation' menu.
The nodule is automatically generated and displayed.

This application is also available as with a stand-alone CLI for data analysts to build pipelines around, using the following example commands:
- The following command will generate two output images (seed image for nodule segmentation and nodule mask image) with names C:/outputBasename_seeds.nii.gz and C:/outputBasename_segmentation.nii.gz:
```
${CaPTk_InstallDir}/bin/SBRT_Nodule.exe -p C:/PET.nii.gz -c C:/CT.nii.gz -m C:/mask.nii.gz -o C:/outputBasename
```
- The following command will generate the nodule mask image with name C:/outputBasename_segmentation.nii.gz:
```
${CaPTk_InstallDir}/bin/SBRT_Nodule.exe -p C:/PET.nii.gz -c C:/CT.nii.gz -m C:/mask.nii.gz -o C:/outputBasename -s C:/seedImage.nii.gz
```

"Label_value" indicates the label of lung field in the input mask image, default value is 2.

Step 3: Prognostic Modeling

Load the required PET image.
Load the required nodule mask generated from step 2.
Supply the model directory. Note: A model trained on PENN data can be downloaded from ftp://www.nitrc.org/home/groups/captk/downloads/models/SBRT.zip
Launch the modeling step using 'Applications -> Prognostic Modeling' menu.
The predicted risks for survival and nodal failure are automatically calculated and displayed.[Note: The prediction models were obtained based on PET images with spatial resolution 4mm x 4mm x 4mm.]

This application is also available as with a stand-alone CLI for data analysts to build pipelines around, using the following example commands:
- The following command will calculate and print the predicted risks regarding survival and nodal failure:
```
${CaPTk_InstallDir}/bin/SBRT_Analysis.exe -i C:/PET.nii.gz -m C:/mask.nii.gz -l 1 -D C:/Model
```
- The following command will calculate and print the predicted risks regarding survival and nodal failure, it will also save the radiomic features used for the prediction into the assigned file:
```
${CaPTk_InstallDir}/bin/SBRT_Analysis.exe -i C:/PET.nii.gz -m C:/mask.nii.gz -l 1 -o C:/outputFile -D C:/Model
```

References:

H.Li, M.Galperin-Aizenberg, D.Pryma, C.Simone, Y.Fan, "Predicting treatment response and survival of early-stage non-small cell lung cancer patients treated with stereotactic body radiation therapy using unsupervised two-way clustering of radiomic features", Int. Workshop on Pulmonary Imaging, 2017.