Population Graph Model to Predict the Clinical Benefits of Immunotherapy in Patients with Non-small Cell Lung Cancer

Project Structure

Step 1. GCN-master

• Constructs population graph based on the patients' baseline clinical characteristics
• Utilizes nine baseline features to build patient similarity graphs before immunotherapy

Step 2. GCN_HGNN.py

• Implements DHGN prognostic model construction, training and external test
• Generates prognostic scores for immunotherapy survival outcome prediction in patients with NSCLC

Important Note

Before running the project:

• Update the median survival time in the code with your own dataset (marked with comments)

Materials for readers:

• MSK_CT_Features_1221.csv: CT features of the 136 patients in the MSK test dataset
• MSK_Clinical_Baseline_Features.csv: Clinical Baseline Characteristics of the 136 patients in the MSK test dataset

Acknowledgments

We gratefully acknowledge the contributions of Huang, Yongxiang and Chung, Albert CS to the PAE model.