A Chronological Overview of Using Deep Learning for Leukemia Detection: A Scoping Review

Area of Science:

• Medical Diagnostics
• Artificial Intelligence in Medicine
• Hematology

Background:

• Leukemia, a rare but fatal blood cancer, necessitates prompt diagnosis for effective treatment and improved patient outcomes.
• Traditional diagnostic methods for leukemia face limitations in accuracy and speed.
• Deep learning (DL) models present a potential solution for enhancing leukemia diagnosis.

Purpose of the Study:

• To systematically review and report on the published literature concerning the application of deep learning (DL) models for leukemia diagnosis.
• To analyze the evolution and current state of DL technologies in detecting leukemia.

Main Methods:

• A systematic literature search was conducted following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.
• Databases searched included Embase, Ovid MEDLINE, and Web of Science for articles published between 2010 and 2023.
• Keywords used were "leukemia" AND "deep learning" OR "artificial neural network" OR "neural network" AND "diagnosis" OR "detection."

Main Results:

• Twenty articles met the eligibility criteria and were included in the review.
• Early studies established foundational methods, evolving towards more generalized DL approaches for leukemia detection.
• Recent DL models demonstrate integrated architectures, significantly improving diagnostic accuracy and efficiency.

Conclusions:

• Deep learning models show considerable promise as a tool to aid physicians in diagnosing leukemia.
• The continuous refinement of DL techniques emphasizes simplicity and efficiency, positioning it as a valuable method for leukemia detection.
• Further research in real-world settings is needed to validate the transformative impact of DL on leukemia diagnosis and patient prognosis.