Validation of plasma <i>p</i> ‐tau217 as a Blood‐Based Biomarker for dementia diagnosis in Latin American population

Abstract Background Plasma biomarkers have immense potential to transform dementia diagnosis globally, with phosphorylated tau 217 ( p ‐tau217) emerging as the most accurate AT(N) biomarker for diagnosis of Alzheimer's Disease (AD). Despite their promise, AT(N) biomarkers remain under‐validated in genetically and environmentally diverse populations, including Latin America (LA), where genetic, lifestyle, and social exposomes add layers of complexity. Diagnostic challenges like disease heterogeneity, limited research infrastructure, and restricted access to diagnostic tools further complicate diagnosis. In LA, research on plasma AT(N) biomarkers is scarce, and no studies in the global south have comprehensively integrated p ‐tau217 with clinical or neuroanatomical correlates, leaving critical gaps in understanding its cognitive and brain‐specific associations. Leveraging the ReDLat cohort, comprising participants from six diverse LA countries (Chile, Argentina, Colombia, Peru, Mexico, and Brazil), this study aims to validate plasma p ‐tau217 as a reliable biomarker for dementia diagnosis in AD and Frontotemporal Lobar Degeneration (FTLD). We hypothesize that p ‐tau217, combined with cognitive assessments and neuroanatomical correlates, will serve as a robust diagnostic tool for dementia in LA populations. Method A total of 600 participants (AD, FTLD, and healthy controls‐HC) were enrolled from the ReDLat cohort. Plasma samples were collected in EDTA tubes, aliquoted, shipped on dry ice, and stored at ‐80°C. Plasma p ‐tau217 levels were quantified using the Lumipulse G600II system. Cognitive and neuroimaging data were used to evaluate associations with p ‐tau217. Machine learning algorithms integrating plasma biomarker, cognitive, and neuroimaging data were developed to generate classification models. Result p ‐tau217 levels were significantly elevated in both AD and FTLD groups compared to HC. Plasma p ‐tau217 demonstrated strong associations with cognitive domains, and its levels corresponded with structural and functional brain alterations. Machine learning models utilizing p ‐tau217 achieved high diagnostic accuracy, with performance further improved by integrating neuroimaging and cognitive data. Conclusion This multimodal study highlights the clinical utility of p ‐tau217 as a blood‐based biomarker for dementia diagnosis in underrepresented LA populations. By addressing the region's genetic, environmental, and socioeconomic diversity, these findings establish a foundation for equitable diagnostic approaches and pave the way for broader implementation of plasma biomarkers in dementia diagnosis across diverse LA populations.