Open-source · SINTEF Medical Image Analysis
Raidionics automatically segments central nervous system tumors from MRI and produces a standardized clinical report — as a desktop app or a 3D Slicer plugin.
What it does
Built for neurosurgical and radiological research teams working with pre- and post-operative MRI, where volumes and locations are measured by hand today.
Voxel-wise tumor masks from a single click, trained with an Attention U-Net architecture on partner-hospital data.
Every mask becomes the same set of numbers, so cases and cohorts can be compared directly.
Pre- and post-operative scans are reported together, quantifying what the surgery achieved.
How it works
Four steps, fully automatic. The desktop app and the 3D Slicer plugin are front ends over the same open-source backend libraries.
01
Pre- or post-operative scans, as raw DICOM or volume formats such as NIfTI.
02
In-house trained models outline the tumor compartments: contrast-enhancing, surrounding non-enhancing, FLAIR hyperintensity.
03
Volumes, diameters, laterality and cortical/subcortical profiles, mapped onto reference atlases.
04
A structured clinical report ready for review, following the EANO/RANO 2.0 guidelines.
Runs locally
Inference happens on your own machine through a graphical interface. No data leaves it, and no internet connection is needed to process a patient.
Scriptable backends
The segmentation and reporting libraries are usable on their own via Python, the command line, or Docker — for batch studies and reproducible pipelines.
Open source
BSD-2-Clause licensed, developed in the open, with trained models released alongside the peer-reviewed papers describing them.
See it in action
Use it as a stand-alone desktop app, or inside 3D Slicer if that's already part of your workflow.
A self-contained desktop app — install it and go, no other software required.
View on GitHub →
Runs the same segmentation and reporting pipeline inside 3D Slicer.
View on GitHub →Model training data provided by partner hospitals across Europe and the US
Get the software
Free installers for Windows, macOS, and Ubuntu Linux. Prefer 3D Slicer? Use the plugin instead.
Latest release · v1.3.2
Prefer 3D Slicer? Get the Raidionics-Slicer plugin instead.
Research software. Raidionics is intended for research use. It is not a certified medical device, it has no CE mark or FDA clearance, and its outputs must not be used as the sole basis for diagnosis or treatment decisions. All results should be reviewed by a qualified clinician.
Research
The methods behind Raidionics, and related clinical research using the platform.
If Raidionics contributed to your work, please cite the software paper. Reporting the version you used (visible in the app's About dialog) helps others reproduce your results.
@article{bouget2023raidionics,
title = {Raidionics: an open software for pre- and postoperative
central nervous system tumor segmentation and standardized reporting},
author = {Bouget, David and Alsinan, Demah and Gaitan, Valeria and
Helland, Ragnhild Holden and Solheim, Ole and Reinertsen, Ingerid},
journal = {Scientific Reports},
volume = {13},
number = {1},
pages = {15570},
year = {2023},
doi = {10.1038/s41598-023-42048-7}
}
Who's behind it
Built by SINTEF Medical Image Analysis with clinical partners at St. Olavs Hospital and NTNU.
