Biomarkers that can help diagnose, predict outcomes, or monitor therapeutic responses in human diseases are essential for the development and clinical assessment of new drugs. In many cases, this currently relies on invasive sampling (e.g. tissue biopsies) or expensive imaging procedures that require access to high-cost, access-constrained equipment and expert analysis. One important example is in neurological diseases, where even access to fluid biomarkers frequently involves lumbar puncture to obtain cerebrospinal fluid (CSF). Alzheimer’s Disease (AD) develops over several years prior to detection and mostly relies on CT/MRI imaging and invasive sampling of CSF for diagnosis.
A team from The Hong Kong University of Science and Technology used the Olink® Target platform to measure >1100 proteins in plasma samples from AD patients and controls, identifying over 400 differentially regulated proteins. Hierarchical clustering associated with a distinct biological process and unique cell-specific expression profiles identified 19 “hubs, 7 of which showed significant correlation to cognitive performance phenotypes, reflecting different AD endotypes. Performance analysis in discovery and validation cohorts of a profile composed of one representative protein from each of the 19 hubs achieved very high-accuracy AD classification with an AUC of 0.959.
Plasma proteomics identifies a 19-protein signature to classify Alzheimer’s Disease with an accuracy of ~96%.
Jiang Y, et al. (2021) Alzheimers & Dementia, DOI: 10.1002/alz.12369