Implementation and Assessment of Tau Thresholds in Non-Demented Individuals as Predictors of Cognitive Decline in Tau Imaging Studies.

Background: Tau accumulation in Alzheimer's disease is associated with short term clinical progression and faster rates of cognitive decline in individuals with high amyloid-ß deposition. Defining an optimal threshold of tau accumulation predictive of cognitive decline remains a challenge. Objective: We tested the ability of regional tau PET sensitivity and specificity thresholds to predict longitudinal cognitive decline. We also tested the predictive performance of thresholds in the proposed new NIA-AA biological staging for Alzheimer's disease where multiple levels of tau positivity are used to stage participants. Methods: 18F-flortaucipir scans from 301 non-demented participants were processed and sampled. Four cognitive measures were assessed longitudinally. Regional standardized uptake value ratios were split into infra- and suprathreshold groups at baseline using previously derived thresholds. Survival analysis, log rank testing, and Generalized Estimation Equations assessed the relationship between the application of regional sensitivity/specificity thresholds and change in cognitive measures as well as tau threshold performance in predicting cognitive decline within the new NIA-AA biological staging. Results: The meta temporal region was best for predicting risk of short-term cognitive decline in suprathreshold, as compared to infrathreshold participants. When applying multiple levels of tau positivity, each subsequent level of tau identified cognitive decline at earlier timepoints. Conclusions: When using 18F-flortaucipir, meta temporal suprathreshold classification was associated with increased risk of cognitive decline, suggesting that abnormal tau deposition in the cortex predicts decline. Likewise, the application of multiple levels of tau clearly predicts the distinctive cognitive trajectories in the new NIA-AA biological staging framework.

Biostatistical Estimation of Tau Threshold Hallmarks (BETTH) Algorithm for Human Tau PET Imaging Studies.

A methodology for determining tau PET thresholds is needed to confidently detect early tau deposition. We compared multiple threshold-determining methods in participants who underwent either 18F-flortaucipir or 18F-MK-6240 PET scans. Methods: 18F-flortaucipir (n = 798) and 18F-MK-6240 (n = 216) scans were processed and sampled to obtain regional SUV ratios. Subsamples of the cohorts were based on participant diagnosis, age, amyloid-ß status (positive or negative), and neurodegeneration status (positive or negative), creating older-adult (age ≥ 55 y) cognitively unimpaired (amyloid-ß-negative, neurodegeneration-negative) and cognitively impaired (mild cognitive impairment/Alzheimer disease, amyloid-ß-positive, neurodegeneration-positive) groups, and then were further subsampled via matching to reduce significant differences in diagnostic prevalence, age, and Mini-Mental State Examination score. We used the biostatistical estimation of tau threshold hallmarks (BETTH) algorithm to determine sensitivity and specificity in 6 composite regions. Results: Parametric double receiver operating characteristic analysis yielded the greatest joint sensitivity in 5 of the 6 regions, whereas hierarchic clustering, gaussian mixture modeling, and k-means clustering all yielded perfect joint specificity (2.00) in all regions. Conclusion: When 18F-flortaucipir and 18F-MK-6240 are used, Alzheimer disease-related tau status is best assessed using 2 thresholds, a sensitivity one based on parametric double receiver operating characteristic analysis and a specificity one based on gaussian mixture modeling, delimiting an uncertainty zone indicating participants who may require further evaluation.

Image-Based 2D Re-Projection for Attenuation Substitution in PET Neuroimaging.

PURPOSE: In dual modality positron emission tomography (PET)/magnetic resonance imaging (MRI), attenuation correction (AC) methods are continually improving. Although a new AC can sometimes be generated from existing MR data, its application requires a new reconstruction. We evaluate an approximate 2D projection method that allows offline image-based reprocessing. PROCEDURE: 2-Deoxy-2-[18F]fluoro-D-glucose ([18F]FDG) brain scans were acquired (Siemens HR+) for six subjects. Attenuation data were obtained using the scanner's transmission source (SAC). Additional scanning was performed on a Siemens mMR including production of a Dixon-based MR AC (MRAC). The MRAC was imported to the HR+ and the PET data were reconstructed twice: once using native SAC (ground truth); once using the imported MRAC (imperfect AC). The re-projection method was implemented as follows. The MRAC PET was forward projected to approximately reproduce attenuation-corrected sinograms. The SAC and MRAC images were forward projected and converted to attenuation-correction factors (ACFs). The MRAC ACFs were removed from the MRAC PET sinograms by division; the SAC ACFs were applied by multiplication. The regenerated sinograms were reconstructed by filtered back projection to produce images (SUBAC PET) in which SAC has been substituted for MRAC. Ideally SUBAC PET should match SAC PET. Via coregistered T1 images, FreeSurfer (FS; MGH, Boston) was used to define a set of cortical gray matter regions of interest. Regional activity concentrations were extracted for SAC PET, MRAC PET, and SUBAC PET. RESULTS: SUBAC PET showed substantially smaller root mean square error than MRAC PET with averaged values of 1.5 % versus 8.1 %. CONCLUSIONS: Re-projection is a viable image-based method for the application of an alternate attenuation correction in neuroimaging.