Cognitive correlations with amyloid, tau and neurodegeneration PET across the Alzheimer's disease age spectrum.

Background: In Alzheimer's disease (AD), cognitive symptoms have varying associations with tau, amyloid and neurodegeneration. Patients with early‐onset AD (EOAD, age<65) have more non‐amnestic symptoms, greater AD pathological burden, and less co‐pathologies than late‐onset AD (LOAD, age≥65). It is unknown whether the association between cognitive symptoms and Alzheimer's disease biomarkers differs with age, and this has implications for clinical trials. The goal of this study was (1) to evaluate the association between domain‐specific cognitive performance and PET‐based pathology in AD, and (2) to compare whether theses associations differ in EOAD versus LOAD. Method: Amyloid‐positive patients meeting criteria for mild cognitive impairment or AD dementia (Total n=71; EOAD n=33, mean age 58±4, MMSE 20±7, 61% female; LOAD n=38, mean age 74±6, MMSE 24±5, 45% female) underwent neuropsychological testing, 18F‐flortaucipir (FTP‐PET), 11C‐PIB‐PET (PIB‐PET), and a subset underwent 18F‐fluorodeoxyglucose‐PET (FDG‐PET, n=57). Composite scores were calculated for episodic memory, semantic memory, language, executive function and visuospatial domains. Voxelwise regressions (thresholded at uncorrected p<0.001 voxel‐level, FWE p<0.05 cluster‐level) evaluated PET/cognition correlations, and EOAD versus LOAD differences were tested with a PET*group interaction. Mediation analyses assessed direct and FDG‐PET‐mediated associations between FTP‐PET and cognitive performance in regions‐of‐interest based on the voxelwise regressions. Results: FTP‐PET and FDG‐PET (lesser extent), but not PIB‐PET, showed significant domain‐specific voxelwise associations with cognitive performance that aligned with known brain‐behavior relationships (episodic memory/medial temporal lobe, semantic memory/anterior temporal, language/perisylvian, executive/frontal, visuospatial/occipital) (Figures 1‐3). There were no significant interactions between age and PET correlations across all modalities. Mediation analysis revealed that FDG‐PET mediated the relationship between FTP‐PET and cognition in all domains except episodic memory (Table 1). FTP‐PET also had a direct effect on executive function (p<0.001) and visuospatial (p=0.02) performance. Conclusion: In AD, cognitive performance correlated with tau and neurodegeneration, but not amyloid, in similar patterns across the age spectrum. The relationship between tau and cognition was in part mediated by glucose metabolism. These findings provide support for inclusion of EOAD patients in clinical trials, for tau PET as a biomarker that captures molecular specificity and aligns with clinical performance, and for the shared potential of disease‐modifying therapies across the AD age spectrum. [ABSTRACT FROM AUTHOR]