Mild cognitive impairment cases affect the predictive power of Alzheimer's disease diagnostic models using routine...

Area of Science:

• Neurology
• Gerontology
• Biomedical Informatics

Background:

• Current diagnostic models for Alzheimer's disease (AD) using routine primary care data have limited predictive accuracy.
• The influence of mild cognitive impairment (MCI) on the performance of these diagnostic models is not fully understood.
• Accurate early identification of AD is essential for timely intervention and patient management.

Purpose of the Study:

• To investigate the effect of mild cognitive impairment (MCI) on the predictive performance of Alzheimer's disease (AD) diagnostic models.
• To assess the utility of routine clinical variables for classifying individuals into healthy control, MCI, and AD groups.
• To identify specific MCI subgroups that most significantly affect diagnostic model accuracy.

Main Methods:

• Utilized data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort.
• Employed the CatBoost machine learning algorithm for multiclass classification.
• Assessed the predictive value of common clinical variables (hematological tests, blood tests, medical history) accessible in primary care settings.

Main Results:

• Mild cognitive impairment (MCI) demonstrably affected the predictive performance of Alzheimer's disease (AD) diagnostic models.
• Identified three distinct subgroups within the MCI classification.
• A specific subgroup of MCI patients, likely representing prodromal AD, was identified as the primary driver of performance impact.

Conclusions:

• Accurate classification of MCI subtypes, particularly distinguishing prodromal AD, is critical for enhancing AD diagnostic models.
• Future research should prioritize differentiating MCI from prodromal AD to improve translational diagnostic tools for primary care.
• Moving beyond binary AD vs. control classifications towards nuanced MCI categorization is essential for advancing AD diagnostics.