Mammo-FM: Breast-specific foundational model for Integrated Mammographic Diagnosis, Prognosis, and Reporting

[Paper] [Github]

Citation

Mammo-FM

@article{ghosh2025mammo,
  title={Mammo-FM: Breast-specific foundational model for Integrated Mammographic Diagnosis, Prognosis, and Reporting},
  author={Ghosh, Shantanu and Joshi, Vedant Parthesh and Syed, Rayan and Kassem, Aya and Varshney, Abhishek and Basak, Payel and Dai, Weicheng and Gichoya, Judy Wawira and Trivedi, Hari M and Banerjee, Imon and others},
  journal={arXiv preprint arXiv:2512.00198},
  year={2025}
}

Mammo-CLIP (MICCAI 2024)

@InProceedings{10.1007/978-3-031-72390-2_59,
author="Ghosh, Shantanu
and Poynton, Clare B.
and Visweswaran, Shyam
and Batmanghelich, Kayhan",
editor="Linguraru, Marius George
and Dou, Qi
and Feragen, Aasa
and Giannarou, Stamatia
and Glocker, Ben
and Lekadir, Karim
and Schnabel, Julia A.",
title="Mammo-CLIP: A Vision Language Foundation Model to Enhance Data Efficiency and Robustness in Mammography",
booktitle="Medical Image Computing and Computer Assisted Intervention -- MICCAI 2024",
year="2024",
publisher="Springer Nature Switzerland",
address="Cham",
pages="632--642",
abstract="The lack of large and diverse training data on Computer-Aided Diagnosis (CAD) in breast cancer detection has been one of the concerns that impedes the adoption of the system. Recently, pre-training with large-scale image text datasets via Vision-Language models (VLM) (e.g., CLIP) partially addresses the issue of robustness and data efficiency in computer vision (CV). This paper proposes Mammo-CLIP, the first VLM pre-trained on a substantial amount of screening mammogram-report pairs, addressing the challenges of dataset diversity and size. Our experiments on two public datasets demonstrate strong performance in classifying and localizing various mammographic attributes crucial for breast cancer detection, showcasing data efficiency and robustness similar to CLIP in CV. We also propose Mammo-FActOR, a novel feature attribution method, to provide spatial interpretation of representation with sentence-level granularity within mammography reports. Code is available publicly: https://github.com/batmanlab/Mammo-CLIP.",
isbn="978-3-031-72390-2"
}

License and Copyright

This model is released under a Custom Academic License for Model Weights.

Summary of Terms

The Model may be used solely for non-commercial academic research purposes.

Permitted users include:

• Academic institutions
• Non-profit research organizations

Allowed activities include:

• Research use
• Evaluation and benchmarking
• Fine-tuning or linear probing for academic research only

The following are strictly prohibited:

• Any commercial use
• Any use by for-profit entities without prior written approval
• Any use in products, services, or revenue-generating activities
• Any industry research, including internal evaluation or development, without explicit prior written permission
• Any clinical deployment
• Any diagnostic or treatment use
• Any use in medical decision-making systems
• Redistribution of the model weights, in whole or in part
• Hosting, serving, or making the model available through an API, SaaS platform, web service, or other remote access mechanism
• Distillation, compression, extraction, imitation, or transfer of knowledge from the model to create, train, initialize, or improve another model

Derivative models are allowed only if they:

• Are used solely for non-commercial academic research
• Are not shared publicly or redistributed
• Inherit this same license in full
• Are not used to circumvent restrictions in this license

Any use of the Model must include:

• Proper citation of the associated publication(s)
• Acknowledgment of the original authors and institution(s)

Example acknowledgment:

"This work uses Mammo-FM developed by the authors of Mammo-FM, Boston University."

The full license text is provided in the LICENSE file.

Copyright © Batman Lab, 2024

Contact

For commercial use, industry research, or other exceptions, contact:

Kayhan Batmanghelich Email: batman@bu.edu

For any queries related to the code, contact Shantanu Ghosh Email: shawn24@bu.edu