arXiv:2611.18793v1 Announce Type: new
Abstract: Purpose: Multimodal medical imaging fusion (MMIF) is an increasingly important approach to breast cancer detection (BCD), as different imaging modalities rely on different physical principles and provide complementary information. No single modality can capture all the structural, functional, and biological information needed for accurate diagnosis and characterization. Methods: This review examines more than 15 imaging modalities used in breast cancer detection, including conventional techniques such as mammography, ultrasound, and magnetic resonance imaging (MRI), as well as emerging approaches such as thermal imaging, photoacoustic imaging, microwave imaging, and multimodal microscopy. Results: Mammography, ultrasound, and MRI remain among the most widely used modalities, providing complementary structural and functional information. Recent AI advances have enabled fusion at the data, feature, and decision levels, with applications in detection, classification, segmentation, prognosis, and personalized cancer management. Multimodal approaches increasingly combine imaging data with clinical, histopathological, and genomic information. However, important challenges remain, including limited availability of large and diverse multimodal datasets, variations in imaging protocols, missing modalities, limited external validation, and the need for interpretable and clinically reliable AI models. Conclusion: MMIF offers promising opportunities to improve breast cancer detection and characterization by combining complementary imaging and non-imaging information. Future research should prioritize robust fusion strategies, standardized and diverse datasets, multicenter validation, explainable and privacy-preserving AI, and the integration of imaging with clinical and molecular data to facilitate reliable translation into clinical practice.
Abstract: Purpose: Multimodal medical imaging fusion (MMIF) is an increasingly important approach to breast cancer detection (BCD), as different imaging modalities rely on different physical principles and provide complementary information. No single modality can capture all the structural, functional, and biological information needed for accurate diagnosis and characterization. Methods: This review examines more than 15 imaging modalities used in breast cancer detection, including conventional techniques such as mammography, ultrasound, and magnetic resonance imaging (MRI), as well as emerging approaches such as thermal imaging, photoacoustic imaging, microwave imaging, and multimodal microscopy. Results: Mammography, ultrasound, and MRI remain among the most widely used modalities, providing complementary structural and functional information. Recent AI advances have enabled fusion at the data, feature, and decision levels, with applications in detection, classification, segmentation, prognosis, and personalized cancer management. Multimodal approaches increasingly combine imaging data with clinical, histopathological, and genomic information. However, important challenges remain, including limited availability of large and diverse multimodal datasets, variations in imaging protocols, missing modalities, limited external validation, and the need for interpretable and clinically reliable AI models. Conclusion: MMIF offers promising opportunities to improve breast cancer detection and characterization by combining complementary imaging and non-imaging information. Future research should prioritize robust fusion strategies, standardized and diverse datasets, multicenter validation, explainable and privacy-preserving AI, and the integration of imaging with clinical and molecular data to facilitate reliable translation into clinical practice.
