Artificial intelligence (AI) is increasingly being explored as a tool to support more precise and dynamic management in the diagnosis and treatment of hematologic malignancies. Unlike previous reviews focused on single disease types or isolated technological pathways, this paper provides a comprehensive overview of AI’s current applications and latest advancements in diagnosis, classification, prognosis assessment, and treatment decision-making for leukemia, lymphoma, multiple myeloma, and myelodysplastic syndromes. It encompasses key technical pathways, including morphology, imaging, flow cytometry, and multimodal data fusion, and further constructs an AI-driven dynamic diagnosis and treatment system along with its integrated deployment framework for electronic health records. This framework is intended to illustrate how multimodal data integration, dynamic risk assessment, and more coordinated longitudinal management could be supported within an integrated workflow. This integrated dynamic model provides a structured roadmap for intelligent, end-to-end management of hematologic malignancies and holds promise for advancing future intelligent clinical pathways. While AI demonstrates significant potential to enhance diagnostic consistency, optimize risk stratification, and enable personalized treatment, its development remains constrained by challenges such as data bottlenecks, insufficient cross-institutional model generalization, and ethical oversight. Future efforts should focus on advancing multicenter prospective validation, adhering to international standards like TRIPOD + AI, and refining data privacy, model interpretability, and ethical oversight systems. These advances may help support the future development of more personalized and dynamic patient management strategies within an evidence-based framework.
Li et al. (2026) studied this question.