Distinct patterns of gray matter atrophy in hippocampal sclerosis of aging and Alzheimer’s disease neuropathology

Abstract

AbstractBackgroundHippocampal sclerosis of aging (HS) and Alzheimer’s Disease Neuropathology (ADNP) both contribute to dementia in the elderly but are difficult to distinguish in vivo. Currently, there are no established biomarkers for HS, and few studies have investigated the distinction between HS‐ and ADNP‐related patterns of brain atrophy. We hypothesized that HS‐related atrophy would be restricted to the hippocampus and surrounding structures, while ADNP‐related atrophy would be most pronounced in the lateral temporal and parietal lobes.MethodWe used data from the National Alzheimer’s Coordinating Center (NACC) (09/2005‐03/2017) and identified 134 participants with suitable 3D T1w MRI and autopsy data. HS pathology was available as present/absent and we defined ADNP as present by a combination of high Braak stage (V/VI) and CERAD neuritic plaque score (3). We used the Computational Anatomy Toolbox (CAT12) to segment the MRI scans into tissue types and generate gray matter density maps. We examined the following regions of interest (ROIs): hippocampus, parahippocampus, amygdala, temporal and parietal regions, and global tissues measures. Additionally, we performed a voxel‐based morphometry (VBM) analysis. We studied the relationship between volumes/densities and neuropathologies using multiple linear regression models while adjusting for sex, age, education, dementia status, years from MRI to death, and intracranial volume.Result Table 1 depicts participant characteristics. We found that HS, but not ADNP, was associated with the volumes of the hippocampus (HS P<0.001, ADNP P=0.4), parahippocampus (anterior only, HS P=0.007, ADNP P=0.8), and amygdala (HS P=0.009, ADNP=0.2, Table 2). ADNP was significantly associated with lateral temporal lobe volumes, most strongly the inferior/middle temporal gyri. We found a trend of association between ADNP and gray matter atrophy (P=0.051). HS was significantly associated with white matter atrophy (P=0.001). We display the VBM results in Figure 1, which largely confirm the ROI results. Additionally, we found HS was associated with bilateral thalamic and insular regions and ADNP with the left amygdala and hippocampal tail atrophy.ConclusionWhile HS is strongly associated with atrophy of the hippocampus and surrounding structures, ADNP is associated with atrophy of lateral temporal lobe regions. These diverging atrophy patterns are potential biomarkers for differentiating these two clinically indistinguishable conditions during life.