AI-Driven Multimodal and Multi-Dimensional Medical Data Intelligence Across the Surgical Care Continuum

Modern perioperative care is increasingly driven by multimodal and multidimensional data. A clinical team managing a surgical patient draws on spatial information from imaging, cellular evidence from pathology, molecular insights from genomics, and temporal context from serial laboratory tests and clinical notes. Each surgical patient presents a distinct combination of anatomical, histological, molecular, and physiological risk, multimodal models offer a route to patient-specific risk estimates and treatment selection rather than population-average predictions. Integrating these heterogeneous sources is therefore essential for surgical risk stratification, intraoperative decision support, and postoperative outcome prediction. Yet most existing Artificial Intelligence (AI) systems process each modality in isolation and fall short of the integrative reasoning that high-quality perioperative care requires.Recent advances in large language models and vision-language models increasingly support cross-modal data transformation through shared representational spaces. Methods for aligning and fusing imaging, pathology, genomic, and text data are maturing. Combined with spatiotemporal modeling across static preoperative scans, dynamic intraoperative signals, and longitudinal postoperative follow-up, they provide a technical foundation for multimodal perioperative AI. These methods extend AI from static pattern recognition toward dynamic prediction of intraoperative adverse events, postoperative complications, and functional recovery trajectories. Challenges remain, including limited annotation for training, robustness across the full perioperative continuum, and generalizability across institutions, monitoring devices, and patient populations. These gaps point to a need for rigorous, clinically grounded multimodal AI research that mirrors the integrative, time-sensitive, and multidisciplinary nature of real-world perioperative decision-making. This Research Topic aims to bring together an interdisciplinary collection of studies on the design, validation, and clinical application of AI methods that jointly model multimodal and multidimensional data in the perioperative setting. We seek contributions that move beyond unimodal, single-task benchmarks to address the complexity of clinical reasoning across the entire surgical care continuum.The Topic will highlight approaches for integrative analysis and fusion of heterogeneous data types, methods for modeling information across spatial and temporal dimensions, and techniques for cross-modal data transformation and real-time decision support. We are particularly interested in work that addresses data efficiency under limited annotation, robustness under real-world acquisition variability, and generalization across centers, devices, and populations. By connecting researchers across medical image computing, computational pathology, clinical informatics, and machine learning, this collection aims to advance AI systems for individualized perioperative care.We welcome submissions that report original empirical findings, novel methods, or critical syntheses related to multimodal and multidimensional medical data intelligence in perioperative clinical management. Acceptable article types include, but are not limited to, Original Research, Systematic Review, Methods, Brief Research Report, Perspective, and Clinical Trial, in line with Frontiers in Medicine guidelines.Topics of interest include, but are not limited to:• Integrative analysis and fusion of multimodal perioperative data (imaging-pathology, imaging-genomics, imaging-text, and related combinations)• Generative models for cross-modal data transformation, automated surgical reporting, and real-time intraoperative decision support• Multimodal predictive tasks for intraoperative events and postoperative complications under limited annotation• AI methods spanning the entire surgical care pathway: preoperative assessment, surgical planning, intraoperative guidance, and postoperative monitoring• Dynamic prediction of postoperative complications and functional recovery from multimodal data• Generalization, robustness, and fairness across institutions, devices, procedures, and populations• Integration of histopathology, immunohistochemistry, and gene expression or sequencing data with imaging for individualized surgical risk stratification and treatment selection Topic editors: Heng Li, Shuai Zhang, Na Guo, Ke Niu, Tianyu Fu Source: https://www.frontiersin.org/research-topics/85059/ai-driven-multimodal-and-multi-dimensional-medical-data-intelligence-across-the-surgical-care-continuum