# Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift

arXiv: https://arxiv.org/abs/2607.10358 · HTML: https://mlsanity.com/paper/2607.10358/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 35 references that are in the map.
- First posted on arXiv 11 Jul 2026.

## Builds on (most cited first, 22 of 22)

- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [On the Opportunities and Risks of Foundation Models](https://mlsanity.com/paper/2108.07258/) (2021)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [DINOv3](https://mlsanity.com/paper/2508.10104/) (2025)
- [BiomedCLIP: a multimodal biomedical foundation model pretrained from fifteen million scientific image-text pairs](https://mlsanity.com/paper/2303.00915/) (2023)
- [MedGemma Technical Report](https://mlsanity.com/paper/2507.05201/) (2025)
- [Exploring scalable medical image encoders beyond text supervision](https://mlsanity.com/paper/2401.10815/) (2024)
- [VinDr-Mammo: A large-scale benchmark dataset for computer-aided diagnosis in full-field digital mammography](https://mlsanity.com/paper/2203.11205/) (2022)
- [UniMed-CLIP: Towards a Unified Image-Text Pretraining Paradigm for Diverse Medical Imaging Modalities](https://mlsanity.com/paper/2412.10372/) (2024)
- [OPTIMAM Mammography Image Database: a large scale resource of mammography images and clinical data](https://mlsanity.com/paper/2004.04742/) (2020)
- [Mammo-CLIP: A Vision Language Foundation Model to Enhance Data Efficiency and Robustness in Mammography](https://mlsanity.com/paper/2405.12255/) (2024)
- [Foundation Models in Radiology: What, How, When, Why and Why Not](https://mlsanity.com/paper/2411.18730/) (2024)
- [Risk of Bias in Chest Radiography Deep Learning Foundation Models](https://mlsanity.com/paper/2209.02965/) (2022)
- [Advancing human-centric AI for robust X-ray analysis through holistic self-supervised learning](https://mlsanity.com/paper/2405.01469/) (2024)
- [Detection of masses and architectural distortions in digital breast tomosynthesis: a publicly available dataset of 5,060 patients and a deep learning model](https://mlsanity.com/paper/2011.07995/) (2020)
- [Multi-View and Multi-Scale Alignment for Contrastive Language-Image Pre-training in Mammography](https://mlsanity.com/paper/2409.18119/) (2024)
- [A Versatile Foundation Model for AI-enabled Mammogram Interpretation](https://mlsanity.com/paper/2509.20271/) (2025)
- [Mammo-FM: Breast-specific foundational model for Integrated Mammographic Diagnosis, Prognosis, and Reporting](https://mlsanity.com/paper/2512.00198/) (2025)
- [Full Field Digital Mammography Dataset from a Population Screening Program](https://mlsanity.com/paper/2411.02710/) (2024)
- [Bias and Generalizability of Foundation Models across Datasets in Breast Mammography](https://mlsanity.com/paper/2505.10579/) (2025)
- [GLAM: Geometry-Guided Local Alignment for Multi-View VLP in Mammography](https://mlsanity.com/paper/2509.10344/) (2025)

## Built on by (most cited first, 1 of 1)

- [Dataset-Specific Encoding of Image Content Underlies Cross-Dataset Failure in Brain MRI Foundation Models](https://mlsanity.com/paper/2610.13027/) (2026)

## Abstract (arXiv)

Foundation models are increasingly used as image feature extractors for mammography, but their robustness under external domain shift remains unclear. We benchmark 15 foundation-model backbones across breast density, BI-RADS severity, and cancer status using a unified frozen-backbone linear-probe protocol, training on 3 source datasets and evaluating on 12 task-compatible out-of-distribution (OOD) datasets after label harmonization. Mammography-specific vision-language models (Mammo-FM and MaMA) provide the strongest mean OOD performance, but robustness is not explained by mammography exposure alone. DINOv3 remains a competitive vision-only baseline, and mammography-adapted pretraining does not consistently improve generalization. Dataset-level analysis further shows that even leading models show heterogeneous performance across datasets. Feature-space inspection reveals that useful representations can preserve clinical signal while retaining dataset and acquisition structure. These findings highlight dataset-level OOD evaluation as a central criterion for assessing mammography representations. Our code is publicly available: https://github.com/biomedia-mira/mammo-ood.