Kwon · Journal of neurology 2026 · cross-sectional diagnostic cohort study · n=185

Blood-based multimodal biomarker models for differentiating early-onset Alzheimer's disease from frontotemporal dementia: a longitudinal study of early-onset dementia and family members (LEAF) study.

Cited 0 times in the scientific literature.

Level 3 - non-randomized controlled study

Non-randomized observational diagnostic cohort study

PubMed 42472733 · doi:10.1007/s00415-026-13996-7 · record verified 2026-08-28

What was done

Researchers evaluated the diagnostic ability of plasma biomarkers to differentiate early-onset Alzheimer's disease (EOAD) from early-onset frontotemporal dementia (EOFTD). They analyzed 185 patients aged 65 years or younger (150 with EOAD, 35 with EOFTD) enrolled in the LEAF study between 2021 and 2023. Plasma concentrations of p-tau217, p-tau181, neurofilament light (NfL), and glial fibrillary acidic protein (GFAP) were measured using immunoassays alongside APOE ε4 status.

What was found

Plasma p-tau217 achieved an AUC of 0.831 and p-tau181 achieved an AUC of 0.862 for discriminating EOAD from EOFTD, with no statistically significant difference between the two p-tau markers. P-tau levels were higher in EOAD, whereas NfL levels were higher in EOFTD and elevated in EOAD patients with severe hippocampal atrophy. Combining p-tau with NfL, GFAP, and APOE ε4 status improved overall diagnostic accuracy (numerical AUC for the multimodal model was not provided in the abstract).

Why it matters

Blood-based biomarkers offer an accessible, minimally invasive method to help resolve the challenging clinical differential diagnosis between early-onset Alzheimer's disease and frontotemporal dementia.

Limits

The sample was heavily unbalanced, with more than four times as many EOAD cases (n = 150) as EOFTD cases (n = 35). The abstract does not provide confidence intervals, sensitivity/specificity metrics, or numerical performance data for the combined multimodal models, and the findings require external validation in independent cohorts.

Cited by