Machine Learning Driven by Magnetic Resonance Imaging for the Classification of Alzheimer Disease Progression:...

Area of Science:

• Neurology
• Artificial Intelligence
• Public Health

Background:

• Alzheimer disease (AD) diagnosis relies on cognitive impairment severity.
• Current understanding lacks specific etiological factors for AD.

Purpose of the Study:

• To comprehensively assess Alzheimer disease prevalence across different stages using machine learning (ML).
• To provide insights into AD prevalence patterns for future research.

Main Methods:

• Systematic review and meta-analysis adhering to PRISMA 2020 guidelines.
• Inclusion of 24 relevant studies focusing on ML approaches for AD diagnosis.
• Prevalence data visualized using forest plots for 2, 3, 4, and 6 stages of cognitive impairment.

Main Results:

• Prevalence of cognitively normal (CN) and AD across 6 stages was 49.28%.
• Prevalence estimate for 3 stages (CN, mild cognitive impairment, AD) was 29.75%.
• Analysis of 4 stages found an overall prevalence of 13.13%, while 6 stages showed 23.75% prevalence.

Conclusions:

• Significant heterogeneity in AD prevalence estimates is influenced by demographic and setting characteristics.
• Machine learning effectively describes AD prevalence across different cognitive stages.
• This research offers valuable perspectives for future Alzheimer disease studies.