Zero-shot anatomical entity recognition in English clinical and research text
What it does
This model detects anatomical structures and body parts in English clinical narratives and biomedical literature, with users supplying entity labels at inference time. It reports an F1 score of 0.863 on the AnatEM benchmark, using a GLiNER backbone with 220M parameters.
Best for
- Radiology and surgical note parsing
- Site-of-disease extraction from records
- Anatomy-aware analytics in healthcare
What it detects (1 labels)
Trained on the label below; you can pass any other label names at inference time.
- Anatomy
Quality
| Benchmark | AnatEM · all models on this benchmark |
|---|---|
| F1 | 0.863 |
| Precision | 0.832 |
| Recall | 0.896 |
| Accuracy | 0.976 |
| Source | Reported by the authors (test_results.json in the repository) |
Anatomical Entity Recognition - Anatomical structures and body parts
Use it
pip install gliner
from gliner import GLiNER
model = GLiNER.from_pretrained("OpenMed/OpenMed-ZeroShot-NER-Anatomy-Base-220M")
print(model.predict_entities("Patient with type 2 diabetes and hypertension.", ["disease"]))All model information shown here is public and belongs to its authors: use and download it freely under each model’s license. The selection, organization, search design and presentation of this catalog are © 2026 Clinical Corpus. Please link to these pages rather than copy them.