# Toward a robust lesion detection model in breast DCE-MRI: adapting foundation models to high-risk women

arXiv: https://arxiv.org/abs/2509.02710 · HTML: https://mlsanity.com/paper/2509.02710/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 4 of its 18 references that are in the map.
- First posted on arXiv 2 Sep 2025.

## Builds on (most cited first, 4 of 4)

- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [KAN: Kolmogorov-Arnold Networks](https://mlsanity.com/paper/2404.19756/) (2024)
- [Foundation Models in Radiology: What, How, When, Why and Why Not](https://mlsanity.com/paper/2411.18730/) (2024)
- [Medical Slice Transformer: Improved Diagnosis and Explainability on 3D Medical Images with DINOv2](https://mlsanity.com/paper/2411.15802/) (2024)

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

- [Transformer Classification of Breast Lesions: The BreastDCEDL_AMBL Benchmark Dataset and 0.92 AUC Baseline](https://mlsanity.com/paper/2509.26440/) (2025)

## Abstract (arXiv)

Accurate breast MRI lesion detection is critical for early cancer diagnosis, especially in high-risk populations. We present a classification pipeline that adapts a pretrained foundation model, the Medical Slice Transformer (MST), for breast lesion classification using dynamic contrast-enhanced MRI (DCE-MRI). Leveraging DINOv2-based self-supervised pretraining, MST generates robust per-slice feature embeddings, which are then used to train a Kolmogorov--Arnold Network (KAN) classifier. The KAN provides a flexible and interpretable alternative to conventional convolutional networks by enabling localized nonlinear transformations via adaptive B-spline activations. This enhances the model's ability to differentiate benign from malignant lesions in imbalanced and heterogeneous clinical datasets. Experimental results demonstrate that the MST+KAN pipeline outperforms the baseline MST classifier, achieving AUC = 0.80 ± 0.02 while preserving interpretability through attention-based heatmaps. Our findings highlight the effectiveness of combining foundation model embeddings with advanced classification strategies for building robust and generalizable breast MRI analysis tools.