Alzheimer's disease (AD) poses an unprecedented global health challenge. It exhibits a protracted preclinical phase spanning nearly 20 years, during which early interventions can effectively extend the asymptomatic period, highlighting the critical importance of precision stratification and risk assessment strategies. Current biomarker-based approaches face significant limitations including invasive procedures, high costs, and lack of standardized risk stratification systems, necessitating the development of novel, cost-effective screening technologies for high-risk population identification. In this study, we developed and validated an innovative method using Laser-Induced Breakdown Spectroscopy (LIBS) for screening high-risk populations with preclinical AD characteristics. By analyzing the elemental distribution in feces from 20 AD mice (10 female and 10 male) and 20 age-matched normal mice from 4 to 51 weeks of age, we demonstrated that LIBS can capture subtle changes in metal concentrations associated with AD progression. Principal Component Analysis (PCA) reveals comprehensive alterations in elemental distribution patterns, while a Random Forest (RF) model achieves a classification accuracy exceeding 90 % in most stages. Shapley Additive Explanations (SHAP) analysis identifies Ca II 396.85 nm and Mg 279.55 nm as the most critical spectral features for accurate discrimination, both contributing 14 % to classification accuracy. These findings highlight the potential of LIBS-based fecal analysis as a non-invasive, rapid, and cost-effective tool for early AD screening. This study validated the feasibility of LIBS + machine learning (ML) for screening high-risk AD populations through animal experiments. Nevertheless, extensive studies and technological iterations on the novel LIBS+ML method are required to achieve clinical application standards.
扫码关注我们
求助内容:
应助结果提醒方式:
