[Update November 2022] This source code has been updated to work with current version of fastMONAI

Pulmonary nodule classification in lung cancer from 3D thoracic CT scans

This code is written by Alexander Selvikvåg Lundervold and Satheshkumar Kaliyugarasan .

Project Organization

├── figures            <- Generated figures
│   
├── notebooks          <- Jupyter notebooks 
│   
├── src                <- Source code for use in this project
│   
├── .ignore            <- Local files and folder to be ignored 
│   
├── README.md          <- The top-level README for developers using this project
│
└── environment.yml    <- Config for conda and pip  

---

Steps to run the experiment

1. Download the processed LIDC_IDRI Version 2 data used in this project from: https://wiki.cancerimagingarchive.net/display/DOI/Standardized+representation+of+the+TCIA+LIDC-IDRI+annotations+using+DICOM

2. Run the following command to create a new conda environment from yml file:

conda env create --file environment.yml
conda activate lung-ct

[Optional] Run the following command with your conda environment activated:

conda env update --file environment.yml

3. Run:

python prepare_images.py <IMAGE_PATH>

4. Go through the notebook: 1.0-classification.ipynb.

[Note] If conda environment is not showing up in Jupyter Notebook run the following lines:

python -m ipykernel install --user --name <ENVIRONMENT> --display-name "Python (lung-ct)"

Acknowledgement

Our work was supported by the Trond Mohn Research Foundation through the project “Computational medical imaging and machine learning – methods, infrastructure and applications" at the Mohn Medical Imaging and Visualization Center, grant number BFS2018TMT07.