Risk prediction models for dementia constructed by supervised principal component analysis using miRNA expression...

Area of Science:

• Biomarkers and Diagnostics
• Neuroscience
• Genetics and Genomics

Background:

• Dementia, including Alzheimer's disease (AD), Vascular Dementia (VaD), and Dementia with Lewy Bodies (DLB), poses a significant health challenge.
• MicroRNAs (miRNAs) are emerging as promising novel biomarkers for various diseases, including dementia.
• Accurate and early diagnosis of dementia subtypes is crucial for effective management and treatment.

Purpose of the Study:

• To investigate serum miRNA expression profiles for identifying novel dementia biomarkers.
• To construct and validate subtype-specific miRNA-based risk prediction models for dementia.
• To assess the potential clinical utility of miRNA biomarkers in prospective dementia risk prediction.

Main Methods:

• Serum samples from 1,601 Japanese individuals were analyzed for miRNA expression.
• Supervised principal component analysis (PCA) logistic regression was employed to build risk prediction models.
• Models were validated on a separate cohort to assess predictive accuracy for AD, VaD, and DLB.

Main Results:

• Risk prediction models demonstrated high accuracy: 0.873 for AD (78 miRNAs), 0.836 for VaD (86 miRNAs), and 0.825 for DLB (110 miRNAs).
• This study represents the first application of miRNA-based risk prediction models in a dementia prospective cohort.
• The developed models showed effectiveness in prospective disease risk prediction.

Conclusions:

• Serum miRNAs can serve as effective biomarkers for predicting the risk of dementia subtypes.
• The developed miRNA-based risk prediction models show promise for future clinical application in dementia diagnosis.
• Further refinement of these models could significantly contribute to early and accurate dementia detection.