Detecting cortical signatures of suspected non‐amyloid pathology using large harmonized datasets

Authors: Swati Rane

Published: 2020-12-07

Abstract

AbstractBackgroundSuspected non‐amyloid pathology (SNAP) MCI individuals constitute 7‐25% of the total MCI group and have a high risk of progression to AD. They are categorized by low amyloid accumulation but significant neurodegeneration as A‐T+[N] as opposed to A‐T‐[N] MCI individuals who have neither a high amyloid burden nor neurodegeneration. The low prevalence of SNAP makes it difficult to evaluate robust biomarkers for this group. In this work we leverage the GAAIN platform to harmonize anatomical imaging data from multiple repositories across the world to determine and compare cortical thickness differences and longitudinal rates of atrophy in SNAP‐MCI and A+T+[N] MCI groups.MethodADNI, EDSD, ARWiBo, and PharmaCog databases were queried using the GAAIN platform for individuals with T1 imaging and AD biomarker status. A total of 218 A+T+[N] MCI individuals, and 107 A‐T+[N] SNAP MCI individuals were identified. Cross‐ sectional differences in cortical thickness was evaluated using FreeSurfer. The differences were re‐evaluated following harmonization to remove any site related systematic differences. Next, longitudinal rates of atrophy were computed using FreeSurfer using the 2‐step model and a cluster wise threshold of p=0.05 with multiple comparisons correction using family‐wise error. Longitudinal data was available in 18 NC, 137 MCI, and 34 SNAP subjects. All comparisons were adjusted for age and gender.ResultBefore harmonizing across site‐related differences, cross‐sectionally, MCI individuals have lower cortical thickness with age in the superior frontal gyrus, caudal middle frontal gyrus, rostral middle frontal gyrus, precentral gyrus, and postcentral gyrus (Figure 1). The SNAP‐MCI individuals only had lower cortical thickness with age in the superior temporal gyrus and precentral gyrus. No group differences were observed. After harmonization, there was minimal difference in outcome. Comparing longitudinal rate of atrophy SNAP‐MCI individuals had a faster rate of atrophy in the superior and medial frontal gyri (Figure 2).ConclusionWe did not find cortical thickness differences between MCI and SNAP individuals. We however show that MCI subjects have more widespread cortical thinning with age compared to SNAP‐MCI subjects. Comparing longitudinal rates of cortical atrophy, SNAP‐MCI individuals have faster rates of atrophy than MCI individuals in the frontal cortex.