Mammo-AGE: Deep Learning Estimation of Breast Age from Mammograms

Age is a well-known and pivotal factor relating to organ functions and health in the human body. For breast, biological aging of breast tissue manifests specific physiological changes distinct from chronological aging. This study introduces a deep learning model to estimate the biological age of the breast using healthy mammograms. Furthermore, we applied the model specifically for downstream breast cancer diagnosis and risk prediction.

Environment Setup

Use environment.yml to setup the environment.

git clone git@github.com:Netherlands-Cancer-Institute/Mammo-AGE.git
cd Mammo-AGE
conda env create --name Mammo-AGE -f environment.yml
conda activate Mammo-AGE

Mammo-AGE is implemented with following specification:

• Python version: 3.10.13
• PyTorch version: 2.4.0
• CUDA version: 12.1

Mammo-AGE Model

For detailed training and evaluation scripts, refer to the following Train-README

# Example training
sh /script/Mammo-AGE_Training/sbatch/train_example.sh

# Example predicting
sh /script/Mammo-AGE_Training/sbatch/predict_example.sh

Mammo-AGE Model finetune on downstream tasks

For detailed training and evaluation scripts, refer to the following Finetune-README

# Example training
sh /script/Downstream_Finetune/sbatch/train_example.sh

# Example predicting
sh /script/Downstream_Finetune/sbatch/predict_example.sh

The configs above are meant to specify exact implementation details and our experimental procedure and may need to be adjusted to your specific use case.

Disclaimer

This code and accompanying pretrained models are provided with no guarantees regarding their reliability, accuracy or suitability for any particular application and should be used for research purposes only. The models and code are not to be used for public health decisions or responses, or for any clinical application or as a substitute for medical advice or guidance.

Citation

If you use this code or models in your scientific work, please cite the following paper:

Coming soon!

Acknowledgements

This work is based on the following repositories: POEs, Global-LocalTransformer, Mean-Variance Loss, Mirai, and BreaCR.

License

The code is Apache 2.0 licensed, as found in the LICENSE file

Contact details

If you have any questions, please contact us.

Email: ritse.mann@radboudumc.nl (Ritse Mann); taotanjs@gmail.com (Tao Tan); x.wang@nki.nl (Xin Wang)

Links: Netherlands Cancer Institute, Radboud University Medical Center and Maastricht University