Statistical Method for Identification of Alzheimer Disease With Multimodal Predictive Markers Mild Cognitive Impairment.

Abstract

Introduction: Predicting the progression from mild cognitive impairment (MCI) to Alzheimer's disease (AD) is crucial for early intervention. Identifying reliable predictive markers can enhance diagnostic accuracy and improve clinical decision-making. This study aimed to explore multimodal predictive markers to distinguish stable MCI (sMCI) from progressive MCI (pMCI) to AD using statistical analysis.

Methods: We analyzed data from the Alzheimer's disease neuroimaging initiative (ADNI), categorizing 487 individuals as sMCI and 348 as pMCI. The study incorporated multiple assessment modalities, including demographics, positron emission tomography (PET), genotyping, magnetic resonance imaging (MRI), and neurocognitive tests. A rigorous data preprocessing approach was applied, including cleaning and feature selection. The area under the curve (AUC) and the Wilcoxon test were used to evaluate the discriminative power of predictive markers.

Results: Our findings showed the strong predictive potential of PET, particularly florbetaben (FBB), which achieved an AUC of 0.84. Neurocognitive tests, including the Alzheimer's disease assessment scale (ADAS13), ADNI-modified preclinical Alzheimer cognitive composite (mPACCtrailsB and mPACCdigit), logical memory delayed recall total (LDELTOTAL), and ADAS cognitive subscale question 4 (ADASQ4), also demonstrated high discriminatory power with AUC values ranging from 0.82 to 0.83. These results indicated that a combination of neuroimaging and cognitive assessments can significantly differentiate between sMCI and pMCI.

Conclusion: The results emphasize the importance of multimodal assessments, particularly PET imaging and neurocognitive tests, in distinguishing sMCI from pMCI. These findings contribute to early AD diagnosis strategies and personalized intervention planning..