Associations between gray and white matter volumes from ex vivo imaging and Alzheimer's Disease Neuropathologic Change Score in the ACT autopsy cohort.

Background: Applications of ex vivo magnetic resonance imaging (MRI) of human brain tissue from autopsy has grown in recent years and can provide useful information that is complementary to data from neuropathological evaluations. We sought to evaluate relationship between gray matter (GM) and white matter (WM) volumes and Alzheimer's Disease Neuropathologic Change Score (ADNC) in the Adult Changes in Thought (ACT) autopsy cohort. Method: GM and WM volumetric data for 33 regions of interests (ROIs) were derived using FreeSurfer for each hemisphere. The volumes were separated into GM and WM with data provided from left hemisphere or right hemisphere or from both. We adjusted GM volume for each ROI by taking the ratio of total ROI volume and total GM and WM volume across both hemispheres. We adjusted appropriately for brains where we had data from any one hemisphere and standardized each GM ROI volume. We adjusted and standardized for WM ROI volumes similarly. We ran logistic elastic net to model the relationship between dichotomized ADNC score and GM and WM volumes adjusting for age at death, sex and brain volume. We performed additional analyses to determine if the results from primary analyses were driven by "Not" or "High" level of the ADNC score. Result: There were 143 brains with Freesurfer data that passed quality control. The mean age of death was 91 and 63% were female (Table 1). Odds ratios (OR) obtained from the elastic net regression are presented in Table 2 and they are visualized in Figure 1. Larger fusiform GM volume (OR = 0.8), frontal pole GM volume (OR = 0.89), and temporal pole GM volume (OR = 0.85) were associated with decreased risk of AD in our primary analyses. Cuneus GM volume (OR = 1.08) and precentral WM volume (OR = 1.13) had an inverse relationship in these models. Follow‐up analyses revealed that the fusiform GM volume (OR = 0.75) was significantly smaller for individuals with "High" ADNC score (Table 2). Conclusion: Post‐mortem ex vivo imaging may serve a complementary role in autopsy studies. Future efforts will be focused on evaluating relationship between brain volumes and various standard neuropathological variables in larger data sets. [ABSTRACT FROM AUTHOR]