Substantial progress in Alzheimer?s Disease (AD) biomarker development over the past decade has resulted in improved understanding of early stages of the disease. One of the earliest ?18?F-labeled PET radioligands for tau, AV1451 or FTP, has been followed recently by newer ligands, including ?18?F-MK6240. These radioligands permit the ?in vivo? detection of tau pathology and thus open the door for investigation of intra-subject evolution of tauopathy from an isolated phenomenon of normal aging to a full-blown, widespread component of AD. Recently available data suggest that tau PET might allow identification of a set of identifiable stages of regional tau binding, from normal aging through asymptomatic amyloidosis to early and advanced AD dementia. The ?in vivo? staging of AD tauopathy could represent a breakthrough technology for AD therapy development, but will require the careful analysis of serial tau PET in order to succeed. Traditional methods of PET quantification are limited to the analysis of isolated time points of PET, and are not capable of leveraging the richer structure inherent in a sequence of intra-subject PET scans. Much of the serial tau PET data being acquired in ongoing therapeutic or observational studies is quantified by the SUVR parameter, which suffers from bias when applied in either cross-sectional or serial study designs. The serial PET data that is becoming more prevalent in present day neuroimaging research and in clinical trials requires methods of analysis tailored to its specific structure. We propose to develop new reference-tissue-based models of serial PET that, by encompassing all tissue regions and all serial time-points, will yield estimates of ligand binding and binding rates-of-change that are more accurate and precise than those based on binding quantified by SUVR. Our models will also improve kinetic parameter estimation from fully-dynamic PET data of reduced scan durations (e.g., 60 min or less). We will develop and test our models based on legacy serial tau and amyloid PET data, and also data from a cohort of 20 subjects currently undergoing serial study with the MK6240 tau ligand as part of an independent study. Legacy data will include 18 subjects with 3 or more time points of late-time FTP data, and 111 subjects with 3 or more time points of fully-dynamic PiB. We also propose to acquire a 4th serial time point to supplement the pre-existing, fully-dynamic (120 min) serial MK6240 tau PET data in a cohort of 20 subjects. Fully-dynamic data will be crucial in validating the proposed models for application to radioligands with varying kinetics, and in identifying optimal, ligand-specific reference tissues.
The proposed project will develop new methods of measuring change in the amount of an abnormal protein known as tau in living humans. More precise measurement of tau accumulation in humans has the potential to shorten clinical trials of disease-modifying drugs and more generally, to increase our understanding of the course and development of Alzheimer's disease.
