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

Area of Science:

• Hematology
• Computational Biology
• Oncology

Background:

• Acute myeloid leukemia (AML) presents significant challenges with high relapse rates and poor survival.
• Current treatments for AML have limitations, creating an unmet need for improved long-term survival and reduced toxicity.
• Artificial intelligence and machine learning (AI/ML) are emerging as powerful tools to address these clinical challenges in AML.

Purpose of the Study:

• To systematically review the application of AI/ML in acute myeloid leukemia.
• To describe the evolution of AI/ML tools and their clinical relevance in AML.
• To highlight the potential of contemporary AI/ML algorithms in addressing AML-related problems.

Main Methods:

• A systematic narrative review of 426 publications.
• Publications were selected based on their focus on the intersection of AML and AI/ML.
• The review period spanned from January 1, 2010, to July 30, 2024.

Main Results:

• The evolution of AI/ML tools in AML was analyzed, distinguishing between early and contemporary algorithms (e.g., generative adversarial networks, transformer-based algorithms).
• Contemporary AI/ML algorithms are being utilized for diagnostic challenges, molecular risk stratification, and clinical outcome prediction in AML.
• The review demonstrates the increasing integration and impact of AI/ML in various aspects of AML management.

Conclusions:

• AI/ML represents a promising frontier for addressing clinical challenges in AML.
• Further utilization of AI/ML is recommended, particularly in the context of allogeneic stem cell transplantation.
• AI/ML holds potential for improving patient outcomes and treatment strategies in acute myeloid leukemia.