Automation of the Logan Plot Based PiB-PET Quantification Over Multiple Subjects and Multiple Reference Regions

摘要

Though with known estimation bias concern, Logan plot with the time activity curve from a reference region as the input function is a simple and robust graphical analysis method for quantifying distribution volume ratio (DVR) of [11C] Pittsburgh compound B (PiB), which reflects β amyloid deposit in human brains, a cardinal neuropathological feature of Alzheimer's disease (AD). Various parts of the cerebellum have been used as the reference tissue region due to the fact that it shows no specific binding to PiB. Manual drawing has been the primary means to obtain the tracer time activity curve from different sub-regions of the cerebellum. The objective of this study was to implement a program to batch the Logan quantification procedure over multiple subjects and over multiple time intervals in the standard brain template coordinate space, utilizing a pre-defined anatomical automated labeling (AAL) cerebellum region in the SPM5 toolbox. Additionally, we compared the effects of using different cerebellum AAL sub-regions on our ability to distinguish patients with AD and normal controls. Using data from our normal subjects, we tested the batch program on different time frames and different AAL sub-regions. For each given combination of settings, the results were identical to those obtained when the Logan analysis was performed separately for each individual subject. The PiB quantification batch program provides an efficient and applicable tool that will be useful in PiB-PET studies of AD and AD-related risk factors.