Profiling tau accumulation with SPReAD: Sub-stages for propagation regions in Alzheimer's disease.

Abstract

Monitoring pathophysiological stages of Alzheimer's Disease (AD) is important for understanding, forecasting and ultimately treating the disease. One prominent example of such staging is the spread of tau protein manifested in the stage-wise accumulation of neurofibrillary tangles [1]. With the development of high-affinity tau PET tracers, it is possible to track the protein accumulation in vivo [3]. Quantification of the spatial continuity of such spread longitudinally could help disentangle the tau propagation mechanism. We introduce "Sub-stages for Propagation Regions in Alzheimer's Disease" ("SPReAD") as a method to represent Braak stage related tau tracer uptake continuously as a spatial series. To achieve such mapping, we projected tau uptake data to surfaces [2] with discrete Braak stage parcels (Fig. 1, top left), and, by assigning vertex weights recursively according to the neighbourhood relationships of adjacent stages' regions, we made a sub-stage map (Fig. 1, bottom left); then, we sampled average uptake values within these sub-stages to obtain each subject's uptake profile. We validated SPReAD using [18F]MK6240 standardized uptake value ratio (SUVR) data from 20 AD subjects at 2 visits; 5 AD subjects at 3 visits; 2 AD subjects at 4 visits, with approximately 12-month inter-visit interval, all amyloid positive; Braak stage discordant cases were excluded. The resulting sub-stage profiles appeared highly correlated across visits for AD patients (r=0.9; p<0.01). The differences between baseline and follow-up SUVRs were generally less prominent at the early stages, with especially significantly greater SUVR (p<0.05, two-sample t-test) at the follow-up at the latest Braak stage for a given subject (Fig. 1, right shows two such subjects with Braak stage VI). The Braak sub-stage profile representation using SPReAD helps identify changes within a given Braak stage and across the stages, both in terms of magnitude and spatial progression. It reduces surface manifold data into more interpretable spatial series, at the same time providing information on subtle between-stage transitions which represents a novel systematic way to quantify tau propagation patterns.