Preprint | Artifacts | License | Citation
This repository contains the source code from the study Explainable histomorphology-based survival prediction of glioblastoma, IDH-wildtype.
Artifacts from the study, including the model weights of the trained sparse autoencoder, are released at Zenodo.
Our study comprises multiple processing steps located in the experiments directory. Each step can be executed from the command line using experiment_tracking_cli.py in a dedicated development container located in the .devcontainer directory.
Example usage: python experiment_tracking/experiment_tracking_cli.py multiple_instance_learning experiments/multiple_instance_learning/parameters.json
The study results are aggregated and plotted in individual Jupyter notebooks located in the analysis directory. These notebooks can be executed in the analysis development container.
If you use our work, please consider citing our preprint:
Redlich, J.-P. et al. Explainable histomorphology-based survival prediction of glioblastoma, IDH-wildtype. Preprint at https://doi.org/10.48550/arXiv.2601.11691 (2026).
The source code is under BSD 3-Clause Clear License. Please refer to the license file for details.
The source code is intended for research purposes only and is not qualified for use as a medical product or as part thereof. It is provided "as is" without specific verification or validation.
Jan-Philipp Redlich, [email protected], Fraunhofer Institute for Digital Medicine MEVIS, Germany.