Amyloid Plaque Detection (YOLOv11)

Automated detection of amyloid plaques (β-amyloid/Aβ deposits) in histopathological images for Alzheimer's disease and related amyloidopathies.

Performance

• mAP:0.481
• Accuracy:56%
• Dataset: 37 WSI with 6500 plaques and 333 clusters expert annotations
• Training: Multi-class YOLOv11 classifier optimized for variable plaque 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/Amyloid_pla...",
    filename="best.pt"
)
model = YOLO(model_path)

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

Clinical Relevance

Detects amyloid plaques in:

• Alzheimer's Disease (AD)
• Cerebral Amyloid Angiopathy (CAA)
• Mixed Dementia
• Down Syndrome-associated AD

Key Features

✅ Standardized quantification across institutions
✅ Supports CERAD score approximation
✅ Detects both diffuse and neuritic plaques
✅ High-throughput processing for research

Use Cases

• Screening for amyloid pathology presence
• Quantitative plaque burden assessment
• Multi-center research standardization
• Clinical trial outcome assessment

Limitations

• Requires β-amyloid immunohistochemistry
• Cannot automatically distinguish diffuse vs. neuritic plaques
• Performance varies with antibody selection
• 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}
}