AI and Machine Learning in Hematology: From Multimodal Data to Explainable and Personalized Clinical Intelligence

Artificial intelligence (AI) and machine learning (ML) are increasingly transforming hematology by enabling the analysis of complex and heterogeneous clinical and biomedical data. Advances in deep learning, computer vision, natural language processing and multimodal learning have created new opportunities for automated blood-cell and bone-marrow morphology analysis, disease classification, prognosis, treatment-response prediction and personalized patient management. At the same time, the integration of genomic, transcriptomic, proteomic, imaging, laboratory and longitudinal clinical data is opening new possibilities for precision hematology. Emerging approaches such as foundation models, generative AI, federated learning and privacy-preserving machine learning further expand the potential of AI-driven hematological research. However, challenges related to data quality, dataset heterogeneity, interpretability, bias, privacy, reproducibility, generalizability and clinical validation remain significant barriers to translating AI innovations into reliable and clinically applicable solutions.This Research Topic aims to bring together researchers, clinicians and computational scientists to explore the development and application of AI and ML for advancing hematological research and clinical practice. The goal is to highlight innovative, reliable and clinically meaningful computational approaches for diagnosis, disease classification, prognosis, risk stratification, treatment-response prediction and personalized clinical decision-making across hematological disorders. Particular emphasis will be placed on emerging approaches capable of integrating multimodal and longitudinal data, including laboratory investigations, blood and bone-marrow morphology, digital pathology, flow cytometry, medical imaging, genomic and other omics data, and electronic health records. The Research Topic will also examine explainable and trustworthy AI, foundation models, generative AI, federated and privacy-preserving learning, and AI-assisted clinical decision-support systems. By bringing together methodological advances and clinically relevant applications, the Topic seeks to identify opportunities and challenges for translating AI research into robust, reproducible and generalizable solutions for hematology.This Research Topic welcomes contributions addressing the development, evaluation and application of AI and ML across hematology.Relevant themes include:• AI for blood and bone-marrow morphology, digital pathology and flow cytometry.• Multimodal and multiview learning.• Genomic, transcriptomic, proteomic and single-cell data analysis.• Disease diagnosis and classification.• Prognosis and risk stratification.• Treatment-response and relapse prediction.• Hematopoietic stem-cell transplantation.• Transfusion medicine.• Drug discovery and treatment optimization.• Clinical decision support.Contributions on foundation models, generative AI, large language models, explainable and trustworthy AI, causal learning, federated and privacy-preserving learning, synthetic data and human–AI collaboration are particularly encouraged.Submissions should demonstrate clear relevance to hematology and, where appropriate, address validation, reproducibility, interpretability, fairness, uncertainty and clinical applicability. Topic editors: Shiv Naresh Shivhare, Dr. Urvashi Rahul Saxena, Himani Bansal, Akshay Sharma Source: https://www.frontiersin.org/research-topics/86498/ai-and-machine-learning-in-hematology-from-multimodal-data-to-explainable-and-personalized-clinical-intelligence