Global and Medial Temporal MRI Morphometry in Alzheimer's Disease and Cognitively Normal Adults: A Retrospective...

Background/Objectives: Alzheimer's disease is one of the most prevalent forms of dementia and is accompanied by progressive anatomical changes within the brain. However, large-scale analyses that compare global and local anatomical differences across age, sex, diagnosis, and repeated scans remain limited. Methods: We constructed a cohort of 1385 total participants, with 3593 scans, who were diagnosed as cognitively normal or with Alzheimer's at entry into the study. Global and medial temporal brain volumes from 3T MRI scans were normalized to intracranial volume, and we used a linear mixed-effects model to analyze the association between participants' age, sex, and diagnosis while also accounting for multiple follow-up scans. Diagnosis-dependent relationships between global and medial temporal structures, brain parenchymal and hippocampal fractions, and body mass index were also calculated. Results: Older age was associated with a decrease in normalized brain volume across each region of interest. In general, AD participants had significantly lower volumes than cognitively normal individuals, with a greater degree of difference in medial temporal regions. Global measures did have different degrees of separation, with gray matter showing greater separation than white matter. Exponential model equations showed that the AD group generally had a greater age-associated percentage decrease across-ROI/ICV association. BMI nor BMI x diagnosis was significantly associated with brain volumes within the model, while the exploratory pooled pTau217 analysis showed it was 4.6 times higher in AD and negatively associated with most morphometric areas. Conclusions: This large, integrated analysis supports the use of MRI morphometry as a valuable tool to directly compare AD-associated differences across global and medial temporal areas and also evaluate diagnosis-dependent relationships across anatomically related structures. Findings indicate that medial temporal measures provide a greater AD vs. CN separation than global measures, with GM being more sensitive than WM. Plasma pTau217 also gives biological context, which helps suggest that MRI results should be interpreted with multiple clinical and biological factors.