## Functionality > Specialized Apps

# Survival Prediction Index

This tool extracts and employs distinctive imaging biomarkers predictive of <i>de novo</i> glioblastoma patients’ survival via multi-parametric MRI pattern analysis which might assist in personalized treatment [5-7].

<b>REQUIREMENTS:</b>

-# T1-weighted (T1)
-# Post-contrast T1-weighted (T1-Gd)
-# T2-weighted (T2)
-# T2-weighted Fluid Attenuated Inversion Recovery (T2-FLAIR)
-# Diffusion Tensor Imaging (DTI) images: AX, FA, RAD, TR
-# Dynamic susceptibility contrast-enhanced MRI (DSC-MRI)
-# Segmentations for the following tissues:
  - Ventricles  - Peritumoral Edema
  - Non-Enhancing tumor core
  - Enhancing tumor core

<b>USAGE:</b>

Training process:
- As inputs, user needs to select TrainingDirectory and OutputDirectory.
- In the TrainingDirectory, the data of each subject should be in separate folders such as : <br>
/AAAA, /AAAB, /AAAC

- In each subject folder, there should be the following sub-folders
  - AAAA/SEGMENTATION (one file for segmentation of the tumor)
  - AAAA/DTI (AX, FA, RAD, TR)
  - AAAA/PERFUSION (RCBV, PSR, PH)
  - AAAA/T1 (T1 image)
  - AAAA/T2 (T2 image)
  - AAAA/T1CE (Post-contrast T1-weighted image)
  - AAAA/FLAIR (Flair image)
Each sub-folder must hold images with filenames
that include the corresponding modality, such as t1, t1ce for T1 and T2 images and
"labels" tag, in the name for segmentation, as in "AAAC0_t1ce_pp.nii"

- There will be two model files generated as output; one for 6 months
prediction and the other for 18 months prediction. The application will also write "mean.csv" and "std.csv" files to be used to z-score test subjects.  

Testing process:
- As inputs, user needs to select TestDirectory and ModelDirectory.
- The data in TestDirectory should be organized the same way as in TrainingDirectory. ModelDirectory will have the files written as output of the training phase.
- A csv file having SPI indices of all the test subjects is written as output.
NOTE: Currently, the user only has the option to train a new classifier based on their data and do testing on that trained classifier. In the future, a classifier trained on a large cohort will be provided.