🩸 cAItomorph: Transformer-Based Hematological Malignancy Prediction from Peripheral Blood Smears in a Real-Word Cohort

cAItomorph is a deep learning model designed for the cytomorphological analysis of peripheral blood smears. This project leverages transformer-based architectures to identify hematological malignancies from real-world clinical data.

📄 Preprint: arXiv:2509.20402
💾 Model Weights: (Coming Soon)

---

🌟 Overview

cAItomorph leverages DinoBloom-B hematology foundation model to encode singel cell image representations.

---

🔑 Key Messages

• Real-World Dataset: We assembled the first real-world dataset of peripheral blood smears for hematological malignancy diagnosis.
• Foundation Model Backbone: Built upon DinoBloom, a hematology foundation model, enabling robust and generalizable cytomorphological feature learning.
• Strong Diagnostic Performance: Achieves good performance on acute leukemias and myeloproliferative neoplasms.
• Clinical Relevance: Supports human experts by providing disease probabilities and cell level attentions, guiding downstream diagnostics.

---

⚙️ Installation

Follow the steps below to set up the environment and install dependencies.

conda create -n caitomorph python=3.10
conda activate caitomorph
pip install torch torchvision torchaudio
pip install numpy pandas h5py transformers Pillow einops

💾 Model weights

After clonning this repo, follows the steps:

cd cAItomorph
wget -O weights.zip ""
unzip weights.zip
rm weights.zip

🚀 Demo

See demo to start using our model.

Dataset

• AML_Hehr [5]: Patient-level single-cell images from 189 subjects, including four genetic AML subtypes and controls.
  https://doi.org/10.7937/6ppe-4020 -> Download from the official website manually.

Explore patients and model output

Visit HematoVis, an interactive tool to visualize single cells, model predictions and more...

Cite us if you use the model and data:

@article{dasdelen2025caitomorph,
  title={cAItomorph: Transformer-Based Hematological Malignancy Prediction from Peripheral Blood Smears in a Real-Word Cohort},
  author={Dasdelen, Muhammed Furkan and Kukuljan, Ivan and Lienemann, Peter and Sadafi, Ario and Hehr, Matthias and Spiekermann, Karsten and Pohlkamp, Christian and Marr, Carsten},
  journal={arXiv preprint arXiv:2509.20402},
  year={2025}
}

✉️ Contact

Prof. Dr. Carsten Marr
📧 [email protected]
🏛️ Institute of AI for Health, Helmholtz Munich