Toward Clinically Actionable Machine Learning and Artificial Intelligence Algorithms in Acute Leukemia: A Systematic Narrative Review

Jean M.G. Sabile, Ping Zhang, Anil V. Parwani, Boris Chobrutsiky, Arpita P. Gandhi, Andrew Srisuwananukorn

Abstract: Introduction: Acute myeloid leukemia (AML) is a heterogenous hematologic malignancy that maintains high relapse rates and poor survival despite ongoing treatment advances. There is critically unmet need for consistently providing long-term survival with minimal treatment toxicity for AML patients. Advances in artificial intelligence/machine learning (AI/ML) offer new approaches to addressing clinical challenges in AML. Methods: In this systematic narrative review, 426 publications focusing on the intersection of AML and AI/ML between January 1, 2010, and July 30, 2024, are reviewed. Results: The evolution of AI/ML tools over time is described from a clinically relevant perspective with a distinction between early epochs of AI/ML versus more contemporary algorithms, such as generative adversarial networks and transformer-based algorithms. This review highlights the utilization of contemporary AI/ML algorithms via addressing diagnostic challenges, molecular risk stratification problems, and clinical outcome prediction in the context of AML. Conclusion: Overall, AI/ML represents a promising new frontier in approaching clinical problems in AML, though there are still opportunities for utilization, particularly in the setting of allogeneic stem cell transplantation.

External IDs:doi:10.1159/000547532