To develop a unified cross-modal framework that integrates mammography and histopathology for breast cancer diagnosis.
Key Findings:
The framework outperforms state-of-the-art unimodal, multimodal, and domain generalization baselines.
Achieved a mean AUC of 0.90 under rigorous leave-one-domain-out evaluation.
Demonstrated smaller domain gaps (0.03 vs. 0.06–0.10) compared to existing methods.
Predictions align with clinically meaningful features, enhancing transparency and trust.
Interpretation:
The study advances multimodal integration and cross-institutional robustness, supporting the deployment of AI systems for diagnostic decision support.
Limitations:
The model's performance may vary with different datasets not included in the evaluation.
Potential challenges in generalizing to clinical settings outside the studied benchmarks.
Conclusion:
This research represents a significant step toward clinically deployable AI systems for breast cancer diagnosis by enhancing integration, robustness, and explainability.
