Comparison of ConvNeXt and Vision-Language Models for Breast Density Assessment in Screening Mammography

TL;DR

This study compares CNN-based and multimodal models for breast density classification in mammography, finding that fine-tuned CNN models outperform multimodal approaches, highlighting the importance of domain-specific adaptation.

Contribution

It provides a comparative analysis of ConvNeXt and BioMedCLIP models across different learning scenarios for mammogram density classification.

Findings

Fine-tuned ConvNeXt outperforms BioMedCLIP linear probe.

Zero-shot classification shows modest performance.

Full fine-tuning yields better results than linear probing.

Abstract

Mammographic breast density classification is essential for cancer risk assessment but remains challenging due to subjective interpretation and inter-observer variability. This study compares multimodal and CNN-based methods for automated classification using the BI-RADS system, evaluating BioMedCLIP and ConvNeXt across three learning scenarios: zero-shot classification, linear probing with textual descriptions, and fine-tuning with numerical labels. Results show that zero-shot classification achieved modest performance, while the fine-tuned ConvNeXt model outperformed the BioMedCLIP linear probe. Although linear probing demonstrated potential with pretrained embeddings, it was less effective than full fine-tuning. These findings suggest that despite the promise of multimodal learning, CNN-based models with end-to-end fine-tuning provide stronger performance for specialized medical imaging. The study underscores the need for more detailed textual representations and domain-specific adaptations in future radiology applications.