Longitudinal cortical atrophy in cases of AT concordance and discordance

Abstract Background Although β‐amyloid and tau PET positivity (A+T+) has been related to neurodegeneration and cognitive decline in Alzheimer’s disease (AD), the driving force of neurodegeneration in discordant AT cases remains controversial. We investigated the impact of AT status on longitudinal rates of cortical atrophy and cognitive decline. Method A subset of 349 individuals (cognitively unimpaired [CU; n=230], cognitively impaired [CI; n=119]) with β‐amyloid and tau PET (a priori baseline), longitudinal MRI (interval; Mean=4.5±3.3 years, scans/individual; Median=3, IQR=[2, 8]), and a cognitive follow‐up ≥2 years, was selected from the ADNI dataset. We classified the individuals based on their AT PET status (A+; global β‐amyloid ≥20 CL, T+; composite temporal region ≥1.34 SUVR) and clustered them based on their annual rates of cognitive decline on the ADAS‐Cog13 (fast vs. slow decliners [FDs vs. SDs]). Baseline AD‐signature cortical thickness was compared across AT groups. We employed a linear mixed‐effects model to assess the interaction between AT status and longitudinal cortical thickness across all cortical regions. Results Demographic characteristics were similar across AT groups (A‐T‐; n=192, A‐T+; n=5, A+T‐; n=106, A+T+; n=46). Most A+T+ individuals (70%) were CI, whereas most A+T‐ (71%), A‐T+ (80%), and A‐T‐ (71%) individuals were CU. Similarly, most A+T+ individuals (85%) were characterized as FDs, whereas most A+T‐ (74%), A‐T+ (60%), and A‐T‐ (88%) individuals were characterized as SDs. A+T+ individuals showed significantly lower AD‐signature cortical thickness (Mean A+T+ =2.78±0.19 mm) compared to A+T‐ (Mean A+T‐ =2.89±0.15 mm; P=0.0023) and A‐T‐ (Mean A‐T‐ =2.89±0.15 mm; P<0.001) individuals, and the highest Aβ burden (Mean=85.3±30.6 CL; P<0.0001) at baseline. AT status significantly influenced longitudinal cortical thickness. A+T+ individuals showed significantly steeper cortical thinning rates compared to A‐T‐ individuals, primarily in temporoparietal areas. Despite similar group‐level cortical thinning rates with A‐T‐ individuals, A+T‐ individuals showed heterogeneous patterns of cortical atrophy at the individual level. These findings remained consistent even when conducting separate analyses for CU and CI individuals. Conclusion Unlike AT discordance, concordant AT positivity exhibited homogeneous cortical atrophy patterns consistent with AD. Considering the concurrent high prevalence of mixed pathologies in neuropathological AD, we further aim to explore the underlying neurodegenerative force in discordant AT cases.