A three-point visual reading of the amygdala on standard MRI, combined with existing amyloid and tau biomarkers, separated 469 patients at the Geneva Memory Center into four groups with different 30-month cognitive trajectories, according to a retrospective study published online in Neurology on December 22, 2025 by Fausto Roveta, Federica Ribaldi, Giovanni B. Frisoni and colleagues in Geneva, Turin and Verona. In linear mixed-effects models adjusted for baseline age and education, the 36 patients classified as LATE showed no measurable change in Mini-Mental State Examination scores over 30 months (beta -0.010, 95% CI -0.096 to 0.075, p = 0.813), a trajectory statistically indistinguishable from the 181 patients with neither Alzheimer nor LATE features. The 146 patients classified as Alzheimer disease (beta -0.136, 95% CI -0.184 to -0.088, p < 0.001) and the 106 classified as AD/LATE (beta -0.139, 95% CI -0.196 to -0.082, p < 0.001) declined faster.
Limbic-predominant age-related TDP-43 encephalopathy, or LATE, is an age-associated condition that closely resembles Alzheimer disease in clinic but has no validated in vivo biomarker. The amygdalar atrophy scale used here, published by the same Geneva group in European Radiology in December 2024, grades the amygdala on volumetric T1-weighted MRI as AAS0 (no atrophy), AAS1 (mild to moderate) or AAS2 (severe). Patients seen between 2014 and 2022 (mean age 71.2 years, 53 per cent female) were sorted by that score crossed with Alzheimer biomarker status, with amyloid PET given priority, then cerebrospinal fluid beta-amyloid 42/40 and phosphorylated tau 181, then tau PET Braak stage; cases with uncertain or discordant biomarker profiles were excluded before classification.
At baseline the pattern followed the same order. Mean MMSE was 27.9 in the No-AD/No-LATE group, 26.2 in the LATE group, 24.2 in the AD group and 22.5 in the AD/LATE group. On the Free and Cued Selective Reminding Test, a memory measure, the LATE and AD groups scored similarly (13.3 and 12.5), both above the AD/LATE group (10.6), with the No-AD/No-LATE group highest at 15.1. The LATE and AD/LATE groups had lower volumes and cortical thickness in regions linked to TDP-43 pathology than the AD and No-AD/No-LATE groups, which the authors read as internal support for the grouping.
The study is observational and retrospective, drawn from a single specialist centre, and no brain was examined after death, so no group label was verified against tissue. The authors list this plainly: neuropathologic confirmation was not available, the LATE group of 36 was small enough to limit statistical power, and other possible causes of amygdalar atrophy, including alpha-synuclein pathology and frontotemporal lobar degeneration with TDP-43, could not be ruled out. The scale's own autopsy comparison, in 96 cases from the Alzheimer's Disease Neuroimaging Initiative, found an association with LATE neuropathologic changes that did not reach conventional significance (p = 0.057 on a test for trend). The paper concludes that prospective and neuropathologic validation is warranted.