Imaging neurodegeneration in Down syndrome: brain templates for amyloid burden and tissue segmentation

Abstract

The focus of Alzheimer’s disease (AD) neuroimaging research has shifted towards an investigation of the earliest stages of AD pathogenesis, which manifests in every young adult with Down syndrome (DS; trisomy 21) resulting from a deterministic genetic predisposition to amyloid precursor protein overproduction. Due to morphological differences in brain structure in the DS population, special consideration must be given to processing pipelines and the use of normative atlases developed for the non-DS population. Further, the use of typical MRI to MRI template spatial normalization is less desirable in this cohort due to a greater presence of motion artefacts in MRI images. The diffuse nature of PiB uptake and comparatively lower spatial resolution of the PET image permits the purposing of this modality as a template for spatial normalization, which can substantially improve the robustness of this procedure in the cases of MRI images with motion. The aim of this work was to establish standardized methods for spatial normalization and tissue type segmentation using DS specific templates in order to perform voxel-wise analyses. A total of 72 adults with DS underwent [ 11 C]PiB PET to assess brain amyloid burden and volumetric MRI imaging. A DS specific PiB template for spatial normalization and a set of DS specific prior probability templates were created with two-pass methods. With implementation of this DS specific PiB template, no participants were excluded due to poor spatial normalization, thus maximizing the sample size for PiB analyses in standardized space. In addition, difference images between prior probability templates created from the general population and the DS population reflected known morphological differences, particularly in the frontal cortex. In conclusion, DS specific templates that account for unique challenges improve spatial normalization and tissue type segmentation, and provide a framework for reliable voxel-wise analysis of AD biomarkers in this atypical population.