/radiomics-volume-effects

Code for the 2020 Physica Medica paper.

Primary LanguageJupyter NotebookBSD 2-Clause "Simplified" LicenseBSD-2-Clause

Machine learning helps identifying volume-confounding effects in radiomics

Authors: Alberto Traverso, Michal Kazmierski, Ivan Zhovannik, Mattea Welch, Leonard Wee, David Jaffray, Andre Dekker, Andrew Hope

This repository contains the analysis code for the 2020 Physica Medica publication. Refer to the manuscript. for more information.

Setup

  1. Install the required packages:
pip install -r requirements.txt
  1. Launch Jupyter in the project directory:
jupyter notebook
  1. Open the notebook file in the browser window.

Notes

The original Lung1 and HN1 data, including the images and contours can be found at https://xnat.bmia.nl. This repository only includes the clinical data and the extracted image features. The HeadNeckMR dataset is not yet available publically. The features were extracted using PyRadiomics with the PyRex extension to handle DICOM inputs. The extraction parameters can be found in the PyRadiomics params file. The code was tested under Python 3.6.