Multi-View and Multi-Scale Alignment for Contrastive Language-Image Pre-training in Mammography

TL;DR

This paper introduces MaMA, a novel multi-view and multi-scale alignment method for adapting CLIP to mammography, addressing data scarcity and high-resolution challenges, and achieving superior performance on real-world datasets.

Contribution

It presents a specialized supervision framework, a symmetric local alignment module, and a parameter-efficient fine-tuning approach for mammography CLIP adaptation.

Findings

Outperforms state-of-the-art baselines on multiple mammography tasks

Uses only 52% of the model size compared to the largest baseline

Effective in handling high-resolution images and data limitations

Abstract

Contrastive Language-Image Pre-training (CLIP) demonstrates strong potential in medical image analysis but requires substantial data and computational resources. Due to these restrictions, existing CLIP applications in medical imaging focus mainly on modalities like chest X-rays that have abundant image-report data available, leaving many other important modalities underexplored. Here, we propose one of the first adaptations of the full CLIP model to mammography, which presents significant challenges due to labeled data scarcity, high-resolution images with small regions of interest, and class-wise imbalance. We first develop a specialized supervision framework for mammography that leverages its multi-view nature. Furthermore, we design a symmetric local alignment module to better focus on detailed features in high-resolution images. Lastly, we incorporate a parameter-efficient fine-tuning approach for large language models pre-trained with medical knowledge to address data limitations. Our multi-view and multi-scale alignment (MaMA) method outperforms state-of-the-art baselines for three different tasks on two large real-world mammography datasets, EMBED and RSNA-Mammo, with only 52% model size compared with the largest baseline. The code is available at https://github.com/XYPB/MaMA