Multimodal data fusion is transforming the landscape of medical imaging by integrating information from diverse imaging modalities—such as MRI, CT, PET, ultrasound, and spectral imaging—along with non-image data sources like biosignals and clinical records. This integrative approach enables more comprehensive, accurate, and actionable insights for diagnostic, prognostic, and therapeutic applications. However, challenges remain in harmonizing heterogeneous data, developing robust and interpretable computational models, and translating fused analytics into clinically meaningful outcomes. Overlapping data distributions, discrepancies in spatial and temporal resolution, and the complexity of high-dimensional data all pose enduring obstacles. While deep learning and other AI methods have made headway, open problems persist, particularly in generalizing models across sites, populations, and modalities, and in achieving clinically validated impact.

This Research Topic aims to accelerate innovation in multimodal data fusion and cross-disciplinary analytics for medical imaging. It seeks to gather pioneering work addressing the theoretical, methodological, and translational challenges inherent in synthesizing information from a variety of medical imaging and data sources.

Key goals include:

- Developing new computational frameworks or algorithms for integrating heterogeneous imaging and non-imaging data.
- Showcasing machine learning and AI-driven fusion strategies spanning beyond traditional deep learning paradigms.
- Advancing methods that improve interpretability, generalizability, and clinical utility.
- Encouraging benchmarking studies, comparative validation, and open access to resources for reproducibility and wider adoption.

Interdisciplinary submissions from medical imaging, biomedical engineering, clinical informatics, computer vision, and related fields are welcome. Relevant themes and topics include (but are not limited to):

- Multimodal and cross-modal fusion techniques: Algorithms for integrating two or more imaging modalities and/or combining image and non-image data.
- Novel AI and machine learning approaches: Hybrid models, interpretable AI, and transfer learning tailored for fused datasets.
- Imaging system design for multimodal acquisition: New technologies enabling streamlined, simultaneous, or sequential data capture from multiple sources.
- Computational analytics for diagnostics and therapy: Approaches enhancing clinical decision-making, risk stratification, or patient-specific predictions through data fusion.
- Evaluation and benchmarking: Protocols, metrics, and datasets for systematic validation and comparative studies of fusion methods.
- Clinical and translational applications: Deployment and assessment of multimodal fusion pipelines in real-world healthcare or translational research settings.
- Open datasets, frameworks, and reproducibility resources.

This Research Topic welcomes contributions that provide new knowledge, impactful insights, and practical advancements in multimodal data integration and cross-disciplinary analytics for medical imaging. By bridging fundamental innovations with clinical applications, the aim is to foster collaboration and knowledge transfer across diverse scientific and professional communities, ultimately supporting progress on SDG 9: Industry, Innovation, and Infrastructure and improving global health outcomes.

Impact

Frontiers in Imaging

Image Acquisition and Processing