Deciphering the role of p‐tau, iron and MAOB in neurobiological basis of F18‐flortaucipir signal with voxel‐to‐voxel histology to PET analysis in different tauopathies

Abstract

AbstractBackgroundFlortaucipir F18 was the first FDA approved PET tracer to detect tau pathology. Previous studies show that Flortaucipir is more sensitive to Alzheimer’s disease (AD) tau than four‐repeat tauopathies. Furthermore, the neurobiological basis of low Flortaucipir signal remain unclear, especially in basal ganglia. In this study, we aimed to identify the contributions of p‐tau (Ser 202), iron and MAOB to Flortaucipir signal in different tauopathies by performing voxel‐to‐voxel correlations between histological inclusion density maps and ante‐mortem flortaucipir PET uptake.MethodCoronal slides of ∼1cm thickness from four tauopathy cases (AD, progressive supranuclear palsy, corticobasal degeneration, and FTLD due to MAPT mutation P305S) and one FTLD‐TDP type‐A case were processed following a computational/histopathology pipeline (Ushizima, Chen, et al. 2021), allowing for whole 3D reconstruction of the histological maps at microscopical level, point‐by‐point aligned to their corresponding MRI T1 images (Figure 1). Serial 120 µm‐thick histological slides underwent immunohistochemistry to p‐tau (Ser 202, CP‐13) and MAOB, and Perl’s iron staining. Density of pathological changes per voxel (1.22×1.22×120um³) was measured using convolution neural network algorithm. Flortaucipir images were thresholded at 1.2 SUVR. Spearman’s bivariate and multiple linear regressions were used to compare Flortaucipir PET uptake to histological CP13/iron/MAOB counterparts. We examined both areas with high Flortaucipir PET uptake and high histological CP13/iron/MAOB burden.ResultTable 1 show patient features. Figure 2 depicts voxel‐to‐voxel correlations result in scatter plot and bivariate/multiple linear regression analysis.ConclusionThe correlation of CP13 (p‐Tau) to Flortaucipir was higher in AD than other tauopathies. Iron deposits contributed to Flortaucipir uptake in regions both with or without histopathologically confirmed p‐tau, including basal ganglia, whereas MAOB contribution was almost null in gray matter. The time gap between PET and autopsy and lower PET resolution compared to histology may explain the moderate regression coefficients, but other unexplored contributors cannot be discarded. Voxel‐to‐voxel histology to neuroimaging correlations allow for investigating multiple components of the neurological basis of PET. We are finalizing the regression analysis based on areas of high histological burden.