چکیده:
Predicting progression from Mild Cognitive Impairment (MCI) to Alzheimer's Disease (AD) is challenging, requiring reliable predictive markers for intervention. This study identifies predictive markers across various assessments, including demographics, Positron Emission Tomography (PET), genotyping, Magnetic Resonance Imaging (MRI), and neurocognitive tests from statistical methods. The primary goal is to discern markers effectively distinguishing stable MCI (sMCI) from progressive MCI (pMCI) to AD. We use the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset which individuals are meticulously categorized into sMCI (487 individuals) and pMCI to AD (348 individuals). In this study, we employ an innovative methodology that involves comprehensive data cleaning and engineering on the ADNI dataset. Area Under the Curve (AUC) and Wilcoxon testing as robust statistical methods are used to reveal distinct cognitive patterns between sMCI and pMCI. With a focus on 50 features, our findings highlight the discriminatory potential of PET, specifically Florbetaben (FBB), boasting an AUC value of 0.84. Neurocognitive tests, including Alzheimer’s Disease Assessment Scale 13 items (ADAS13), ADNI-modified Preclinical Alzheimer Cognitive Composite with Trails B (mPACCtrailsB), ADAS cognitive subscale question 4 (ADASQ4), Logical memory delayed recall Total score (LDELTOTAL), and ADNI-modified Preclinical Alzheimer Cognitive Composite with Digit Symbol Substitution (mPACCdigit), exhibit substantial discriminatory power, each with AUC values of 0.83, 0.83, 0.82, 0.82, and 0.82, respectively. Consequently, diverse features, analyzed via robust methods, reveal cognitive markers for differentiating stable and progressive MCI, offering insights for improved early AD diagnosis and intervention.
| بازنشر اطلاعات | |
 |
این مقاله تحت شرایط Creative Commons Attribution-NonCommercial 4.0 International License قابل بازنشر است. |
