Center-for-Computational-Neuropathology/Lewy_body

Lewy Body Detection (YOLOv11)

Automated detection of Lewy bodies (α-synuclein aggregates) in histopathological images for Parkinson's disease and related synucleinopathies.

Performance

• mAP: 0.535 (superior balanced performance)
• F1-Score: 0.59 at optimal thresholds
• Training: Stable convergence over 196 epochs
• Dataset: 5 WSI with ~1000 expert annotations

Quick Start

from ultralytics import YOLO
from huggingface_hub import hf_hub_download

# Download and load model
model_path = hf_hub_download(
    repo_id="Center-for-Computational-Neuropathology/Lewy_body",
    filename="best.pt"
)
model = YOLO(model_path)

# Run inference
results = model.predict("synuclein_stained_image.jpg", conf=0.25, imgsz=640)

Clinical Relevance

Detects Lewy bodies in:

• Parkinson's Disease (PD)
• Dementia with Lewy Bodies (DLB)
• Parkinson's Disease Dementia (PDD)

Key Features

✅ Best overall performance among four pathology types
✅ Stable training convergence
✅ Balanced precision-recall trade-off
✅ Eliminates inter-observer variability

Limitations

• Requires α-synuclein immunohistochemistry staining
• Performance varies with staining protocols
• May not distinguish Lewy bodies from Lewy neurites
• Requires expert validation for clinical use

Citation

@article{neuropath_yolo_2025,
  title={Automated Detection of Neurodegenerative Pathology Using YOLOv11},
  author={[Authors]},
  journal={[Journal]},
  year={2025}
}