Journal: eClinicalMedicine
Article Title: Development and validation of machine learning models with blood-based digital biomarkers for Alzheimer’s disease diagnosis: a multicohort diagnostic study
doi: 10.1016/j.eclinm.2025.103142
Figure Lengend Snippet: ATR-FTIR spectra comparison and diagnostic performance between AD, MCI and HC participants. (A) ATR-FTIR average spectra of patients with AD and HCs in Cohort 1 and the ATR-FTIR spectra of purified GFAP, p-Tau, Aβ42; AD spectra digital biomarkers were marked on the spectra of corresponding AD indicator biomarkers; (B) Box diagram of the six AD spectra digital biomarkers with significantly different levels between AD and HC participants in Cohort 1 ( p < 0.05); (C) ROC curve of the discrimination power of each AD spectra digital biomarker and all six AD spectra digital biomarkers on the test set in Cohort 1; (D) ROC curve of diagnostic performance for AD vs. HC in Cohort 2; (E) ROC curve of the discrimination power of plasma biomarkers for AD vs. HC in Cohort 1; (F) Second derivative spectra of MCI patients in Cohort 3 and AD patients/HCs in Cohort 1, spectral digital biomarkers for distinguishing MCI and HC and for distinguishing MCI and AD were marked on the second derivative spectra with different colors; (G) ROC curve of performance for distinguishing MCI from HC and AD participants in Cohort 1; (H) Relationship between spectra digital biomarkers for distinguishing MCI and MMSE scores in AD of Cohort 1 and Cohort 3. AD, Alzheimer’s disease; HC, healthy control; GFAP, glial fibrillary acidic protein; AUC, area under curve; MCI, mild cognitive impairment; MMSE, mini-mental state examination; ROC, receiver operating characteristic.
Article Snippet: Spectra of plasma samples were obtained using an Alpha FTIR spectrometer with attenuated total reflection (ATR) attachment (Bruker Optics Ltd.), operated with OPUS 5.5 software.
Techniques: Comparison, Diagnostic Assay, Purification, Biomarker Discovery, Clinical Proteomics, Control
