Back to Projects List
PyRadiomics Models
Key Investigators
1 Department of Radiation Oncology, Dana-Farber Cancer Institute, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA.
2 Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA.
3 Department of Radiology, Netherlands Cancer Institute, Amsterdam, The Netherlands.
4GROW-School for Oncology and Developmental Biology, Maastricht University Medical Center, Maastricht, The Netherlands.
5 Kitware, Inc.
6 Isomics, Inc.
Project Description
Objective
  - Develop framework to easily apply and share radiomic models using PyRadiomics and it’s 3D slicer extension SlicerRadiomics.
 
  - In addition to the current lesion-based extraction, add a voxel-based extraction.
 
  - Update the SlicerRadiomics extension to accept the PyRadiomics parameter files.
 
Approach and Plan
Objective #1
  - Develop configuration file to store a radiomic model.
 
  - Add a pyradiomics module or a separate github python repository that can read/write the config file and use radiomics to apply it.
 
  - Add commandline interface to use new functionality.
 
  - Develop new Slicer module to interface with this new pyradiomics module to make this new functionality directly available in slicer.
 
Objective #2
  - Update radiomics base module to allow voxel-based extraction
 
  - Add new feature extractor module to provide interface for voxel-based extraction
 
  - Add Commandline interface for voxel based extraction
 
  - Add examples (settings, usage) to show new functionality
 
Objective #3
  - Update the SlicerRadiomics widget to have a toggle button for manual customization (UI elements in Slicer) or parameter file customization (path line edit to select file)
 
  - Update the SlicerRadiomics logic to allow an extraction using the parameter file customization
 
  - Add a test to SlicerRadiomicsTest to test this new functionality
 
Progress and Next Steps
Objective #1
Related PR
  - Designed a model definition configuration file and validation schemas
 
TODO:
  - Implement functionality to apply models
 
  - Add several model types (e.g. linear regression, logistic regression)
 
Objective #2
Related PR
  - Added voxel based extraction functionality
 
  - Updated current feature extractor module to optionally allow voxel based extraction
 
  - Added examples to show new functionality
 
TODO:
  - Add commandline interface
 
Objective #3
Related PR
  - Update widget
 
  - Update Logic
 
  - Update Test
 
Illustrations
Objective #2


Objective #3


Background and References
Other related references:
  - Image biomarker standardisation initiative: https://arxiv.org/abs/1612.07003
 
  - Radiomics Ontology: http://www.radiomics.org/RO/01000, http://bioportal.bioontology.org/ontologies/RO