This is the implementation of the submitted paper “Biology-guided Deep Learning Predicts Prognosis and Cancer Immunotherapy Response”

A set of sample was stored in the folder "Data"

This code has been tested On Ubuntu 18.04
The required 3rd libraries are listed as follow：
Python =3.6
TensorFlow =1.10
cudatoolkit =9.0
cudnn =7.6.5
imgaug =0.4.0
numpy =1.19.2
scikit-learn =0.24.1
simpleitk =2.0.2
opencv-python =4.5.1.48
xlrd
pydicom

1、Run the “ROI_Extract.py” to pre-process the patient data and convert the dicom into npy.
2、Run the "Train.py" to re-train the well-designed model.