Biomarkers

Background: An urgent need exists for minimally invasive testing for accurate detection of Alzheimer's disease (AD). Circulating microRNAs (miRNAs) have been investigated as a promising candidate biomarker for AD diagnosis and prediction because of their involvement in multiple brain signaling pathways in both health and disease. This study developed and validated a serum miRNA panel in discriminating clinically diagnosed AD from age-matched cognitively healthy controls.

Method: 383 serum samples (194 AD, 189 cognitively healthy controls) were divided into three cohorts: discovery (n=59), training (n=126), and validation (n=198). In the discovery cohort, 49 miRNAs curated from literature databases were verified using individual serum sample via reserve transcriptase-quantitative Polymerase chain amplification (RT-qPCR). A logistic regression model was built with 11 differentially expressed miRNAs using the training cohort, and the final panel comprising 7 miRNAs with superior diagnostic performance was established. The diagnostic efficacy of the 7-miRNA panel was further evaluated in the validation cohort by the receiver operating characteristic (ROC) analysis.

Result: Of the initial 49 screened serum miRNAs, 11 differentially expressed miRNAs were selected for logistic regression model construction based on their potential for detecting AD patients (AUC ≥ 0.7). After model optimization and validation via RT-qPCR, a 7-miRNA panel (miR-146a-5p, let-7i-5p, miR-21-5p, miR-29c-3p, miR-92a-3p, let-7f-5p, and miR-1285-5p) was identified with area under the curve (AUC) of 0.970 and 0.932 in the training and validation cohorts, respectively. The sensitivity of 7-miR test was 88%, and the specificity was 85% in the validation cohort.

Conclusion: These findings suggest that the 7-miRNA signature in serum serves as a novel noninvasive tool for the adjunctive diagnosis of AD. The panel shows promise for clinical application, setting the stage for future studies across diverse populations.