Fig. 6: The LDS model improves diagnostic accuracy and complements p-tau biomarkers in detecting AD and AT+ individuals. | npj Digital Medicine

Fig. 6: The LDS model improves diagnostic accuracy and complements p-tau biomarkers in detecting AD and AT+ individuals.

From: A machine learning–enabled blood transcriptomic signature for digital diagnosis and subtyping of Alzheimer’s disease

Fig. 6

A Plasma expression levels of p-tau181 and p-tau217 in AD vs NC groups (n = 180 NC, n = 270 AD). Correlation between LDS scores and p-tau181 (B) or p-tau217 (C) in the Independent clinical plasma cohort. D ROC curves of LDS, p-tau181, and p-tau217 for discriminating against AD from NC. Confusion matrices showing classification performance of p-tau181 (E) and p-tau217 (F). G ROC curves of combined models integrating LDS with p-tau biomarkers for AD vs NC. H ROC curve of LDS for stratifying AT+ vs AT– individuals, indicating its potential for detecting tau pathology. Statistical analyses included Mann–Whitney U testing for expression, Spearman correlation for association analysis, and DeLong testing for ROC comparison. ***P < 0.001.

Back to article page