Prognostic value of regional tau PET measures in Alzheimer’s Disease

Abstract

AbstractBackgroundRecent findings (Ossenkoppele et al, JAMA Neurology, 2021; Pontecorvo et al, Brain, 2019) suggest that tau‐positron emission tomography (PET) may better predict cognitive change in AD compared to amyloid‐PET and MRI‐based cortical thickness. However, the relationship between region‐based standardized uptake value ratio (SUVR) metrics and cognitive change in terms of prognostic value is not clearly understood. In this study, we evaluated the potential utility of region‐based tau‐PET metrics to stratify patients by cognitive decline risk and amyloid status in the brain.MethodsWe analyzed a subgroup of N=100 symptomatic (24 MCI and 76 AD) subjects from the ADNI database that had longitudinal cognitive scores and flortaucipir and florbetapir scans acquired within ±6 months of baseline cognitive scores. We quantified tau burden using global MUBADA SUVR (Devous et al, JNM, 2018) and AAL atlas‐based temporal and parietal regions using a modified cerebellar gray matter reference region (Pontecorvo et al, Brain, 2017). For every global and region‐based tau SUVR, patients were dichotomized into tau groups of high (T+) and low (T‐) by using a series of thresholds (1.15‐1.45). For all identified tau groups, average Mini‐Mental State Exam (MMSE) score, average MUBADA SUVR, percentage of amyloid‐positive individuals and change in 13‐item Alzheimer’s Disease Assessment Scale‐Cognitive subscale (ADAS‐Cog13) scores were assessed. Signal‐to‐noise ratios (SNRs) and mean annual cognitive changes were measured using ADAS‐Cog13 scores that were obtained in a 2‐to‐3‐year period per patient.ResultsFor every global and region‐based flortaucipir measure, an increase in SUVR threshold resulted in lower average MMSE score, larger average global tau burden, higher percentage of amyloid‐positive patients and increased mean cognitive decline for patients in T+ group. SUVR thresholds that identified subgroups with high percentage (≥90%) of amyloid‐positive (Table 1) and high (>1) SNRs (Table 2), were observed to be lower for brain regions identified later in the tau pathologic cascade: inferior temporal (1.45), lateral temporal (1.35), lateral parietal (1.20).ConclusionsRegion‐based tau‐PET analysis showed that utilizing temporal and parietal SUVR values could help stratify patients by risk of cognitive decline and amyloid status. Clinical trial eligibility algorithms based on region‐based tau metrics, might enable enrolling patients into customized AD trials.