Serum miRNA Detection-based Alzheimer’s disease prediction regression model

Abstract

With the deepening of Alzheimer’s disease (AD) research, serum miRNA has attracted widespread attention as a potential biomarker. Traditional diagnostic methods for AD have certain limitations, such as reliance on clinical symptoms and neuroimaging examinations, which lack sensitivity (Sen) and specificity (Spe) for early diagnosis. Therefore, this article aimed to explore the expression levels of serum miRNA in AD patients and its clinical significance, to construct an AD prediction regression model based on serum miRNA detection. This article found no statistical differences in gender, underlying diseases, age, triglycerides (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) between the control group (healthy individuals) and the AD group, but obvious distinctions were observed in Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA), Alzheimer’s Disease Assessment Scale-cognitive part (ADAS-cog), and Activities of Daily Living (ADL) scores. Further analysis revealed obvious distinctions in miR-31, miR-93, miR-124-3p, miR-143, miR-146a, and miR-218-5p between the two groups, with miR-124-3p showing the best diagnostic effect, followed by miR-218-5p. Based on these findings, this article constructed an AD prediction regression model, and the experimental results indicated that the model has high Sen, Spe, and accuracy (Acc) in the early diagnosis of AD, reducing the error rate of subsequent diagnoses and providing new ideas and methods for the early diagnosis of AD.