A probabilistic framework for clinicopathological Alzheimer's disease using tau-PET

[¹⁸F]flortaucipir tau-PET detects neurofibrillary tangle (NFT) pathology in Alzheimer's disease (AD), one of the core pathological hallmarks of the disease. In clinical settings, tau-PET is usually interpreted qualitatively, and therefore quantitative estimates of how scan results alter the probability that AD explains a patient's symptoms are lacking. Here, we evaluate the probability of clinicopathological AD given a positive or negative tau-PET scan and examine how patient age and amyloid-PET status modulate this probability.

We computed positive and negative predictive values (PPV/NPV) of tau-PET for clinicopathological AD, defined as mild cognitive impairment or dementia with AD as the primary etiology. To account for potential pathological and clinicopathological mismatches, PPV and NPV were modeled using literature-derived sensitivity and specificity estimates of [¹⁸F]flortaucipir PET for postmortem Braak V/VI NFT pathology, together with literature-derived age-dependent tau-PET positivity rates in cognitively unimpaired individuals. We also considered hypothetical clinician-estimated pre-PET AD probabilities. PPV and NPV were calculated for tau-PET and for sequential amyloid- and tau-PET scenarios.

Tau-PET PPV for clinicopathological AD was generally high, particularly in individuals with higher clinician-estimated pre-PET AD probabilities, and showed minor age-related declines (e.g., 84% at ages 50-55 vs. 75% at ages 85-90, at 50% pre-PET probability). Tau-PET NPV was consistently higher than PPV and showed negligible age-related decline (e.g., 92% at ages 50-55 vs. 90% at ages 85-90, at 50% pre-PET probability). A positive tau-PET following positive amyloid-PET substantially increased PPV, particularly in older individuals (e.g., PPV increased from 56% to 83% at ages 75-80 and 30% pre-PET probability), whereas the corresponding gain in NPV after tau-PET following amyloid-PET was smaller.

Tau-PET demonstrates high PPV and NPV for clinicopathological AD. A positive tau-PET following positive amyloid-PET further increases the probability of clinicopathological AD, particularly in older adults. This underscores tau-PET's clinical utility and highlights the effect of age and pre-PET certainty on post-PET AD probabilities.