Machine Learning‐Enhanced Ultrasensitive Immuno‐CRISPR Array Facilitates Early Diagnosis of Alzheimer's Disease by Detecting Multiple Plasma Biomarkers

Jun 9, 2026Advanced science (Weinheim, Baden-Wurttemberg, Germany)

Using advanced machine learning and ultrasensitive tests to detect multiple blood markers for early Alzheimer's diagnosis

AI simplified

CRISPR Gene Editing on OpenScience ↗PubMed ↗DOI ↗OA ↗

Abstract

The ultrasensitive CRISPR-based multi-protein detection array (UCMDA) achieves a detection limit of 1 fg/mL, 10,000-fold more sensitive than conventional methods.

The UCMDA can concurrently detect six core Alzheimer's disease biomarkers.
Integration of antibody pair-based multiplex amplification with CRISPR detection improves diagnostic capabilities.
Clinical validation with 155 plasma samples shows enhanced diagnostic performance for Alzheimer's-related mild cognitive impairment and Alzheimer's disease.
The multi-biomarker model significantly outperforms single-biomarker approaches in diagnosing Alzheimer's disease.
This method offers a scalable, cost-effective, and minimally invasive strategy for early detection and monitoring of Alzheimer's disease.

AI simplified

Early and accurate diagnosis of Alzheimer's disease (AD) remains a significant challenge due to the multifactorial and dynamic nature of its pathology. Although plasma-based biomarkers such as amyloid-β (Aβ) and phosphorylated tau (p-tau) have shown promise as diagnostic indicators, current single-biomarker detection techniques lack the requisite sensitivity and specificity for early-stage diagnosis. Here, we present the development of an ultrasensitive CRISPR-based multi-protein detection array (UCMDA) capable of concurrently detecting six core AD biomarkers, including Aβ, Aβ, p-tau, p-tau, p-tau, and p-tau. By integrating antibody pair-based multiplex recombinase polymerase amplification (RPA) with spatially encoded CRISPR-Cas12a detection, the UCMDA achieves a detection limit of 1 fg/mL, which is 10 000-fold more sensitive than conventional ELISA. Clinical validation in a cohort of 155 plasma samples demonstrated that logistic regression (LR)-based integration of the six biomarkers significantly enhanced diagnostic performance, with the multi-biomarker model substantially outperforming single-biomarker approaches in diagnosing AD-MCI and AD. This platform offers a scalable, cost-effective, and minimally invasive strategy for early detection and disease monitoring. This work highlights the potential of CRISPR-based multiplex protein detection technologies combined with machine learning-assisted analysis to enhance the precision of diagnosing neurodegenerative disorders. 40 42 181 217 231 396,404