You need to agree to share your contact information to access this model
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
These weights are derived from clinical research data and are released for non-commercial research use only. They are not a medical device and must not be used for clinical decision making. By requesting access you agree to these terms.
Log in or Sign Up to review the conditions and access this model content.
CRLM desmoplastic status and survival prediction
Slide-level prediction of desmoplastic histopathological growth pattern (dHGP) status and overall survival (OS) for colorectal liver metastases (CRLM) from hematoxylin and eosin whole slide images (WSI).
Method
The WSI is compressed with the MTDP ResNet50 encoder at spacing 2.0 into a 2048-channel feature map (neural image compression). A gated-attention network then predicts a slide-level output (attention-based multiple instance learning).
Contents
paper_models/: 5-fold desmoplastic classifier (No-HGP vs HGP), validation AUC 0.94 to 0.95, plus the shared encoderresnet50-mh-best-191205-141200.pth.os_models/: 5-fold overall survival model, validation c-index 0.63 to 0.67.
Both ensembles share the single MTDP ResNet50 encoder bundled in paper_models/.
Intended use
Research use only. Not a medical device. Not for clinical decision making.
Links
- Code: https://github.com/DIAGNijmegen/pathology-liver-survival
- Grand Challenge: https://grand-challenge.org/algorithms/colorectal-liver-metastases-survival-prediction/
Citation
Confirm the correct paper and MTDP encoder citation with the authors before use.