# MamaDino: A Hybrid Vision Model for Breast Cancer 3-Year Risk Prediction

arXiv: https://arxiv.org/abs/2602.13930 · HTML: https://mlsanity.com/paper/2602.13930/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 28 references that are in the map.
- First posted on arXiv 14 Feb 2026.

## Builds on (most cited first, 10 of 10)

- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [DINOv3](https://mlsanity.com/paper/2508.10104/) (2025)
- [CoAtNet: Marrying Convolution and Attention for All Data Sizes](https://mlsanity.com/paper/2106.04803/) (2021)
- [ConViT: Improving Vision Transformers with Soft Convolutional Inductive Biases](https://mlsanity.com/paper/2103.10697/) (2021)
- [Inductive Biases for Deep Learning of Higher-Level Cognition](https://mlsanity.com/paper/2011.15091/) (2020)
- [OPTIMAM Mammography Image Database: a large scale resource of mammography images and clinical data](https://mlsanity.com/paper/2004.04742/) (2020)
- [Dino U-Net: Exploiting High-Fidelity Dense Features from Foundation Models for Medical Image Segmentation](https://mlsanity.com/paper/2508.20909/) (2025)
- [DisAsymNet: Disentanglement of Asymmetrical Abnormality on Bilateral Mammograms using Self-adversarial Learning](https://mlsanity.com/paper/2307.02935/) (2023)
- [Generalized Hadamard-Product Fusion Operators for Visual Question Answering](https://mlsanity.com/paper/1803.09374/) (2018)
- [STA-Risk: A Deep Dive of Spatio-Temporal Asymmetries for Breast Cancer Risk Prediction](https://mlsanity.com/paper/2505.21699/) (2025)

## Abstract (arXiv)

Breast cancer screening programmes increasingly seek to move from one-size-fits-all interval to risk-adapted and personalized strategies. Deep learning (DL) has enabled image-based risk models with stronger 1- to 5-year prediction than traditional clinical models, but leading systems (e.g., Mirai) typically use convolutional backbones, very high-resolution inputs (>1M pixels) and simple multi-view fusion, with limited explicit modelling of contralateral asymmetry. We hypothesised that combining complementary inductive biases (convolutional and transformer-based) with explicit contralateral asymmetry modelling would allow us to match state-of-the-art 3-year risk prediction performance even when operating on substantially lower-resolution mammograms, indicating that using less detailed images in a more structured way can recover state-of-the-art accuracy. We present MamaDino, a mammography-aware multi-view attentional DINO model. MamaDino fuses frozen self-supervised DINOv3 ViT-S features with a trainable CNN encoder at 512x512 resolution, and aggregates bilateral breast information via a BilateralMixer to output a 3-year breast cancer risk score. We train on 53,883 women from OPTIMAM (UK) and evaluate on matched 3-year case-control cohorts: an in-distribution test set from four screening sites and an external out-of-distribution cohort from an unseen site. At breast-level, MamaDino matches Mirai on both internal and external tests while using ~13x fewer input pixels. Adding the BilateralMixer improves discrimination to AUC 0.736 (vs 0.713) in-distribution and 0.677 (vs 0.666) out-of-distribution, with consistent performance across age, ethnicity, scanner, tumour type and grade. These findings demonstrate that explicit contralateral modelling and complementary inductive biases enable predictions that match Mirai, despite operating on substantially lower-resolution mammograms.