Improved Quantitative Methods for Multiple Neuropharmacological Non-Invasive Brain PET Studies.

Positron Emission Tomography;PET Pharmacokinetics;PET Kinetic Modeling;Neuropharamacological Measurements Using PET;Neuroimaging;PET for Clinical Diagnosis and Drug Discovery;Radiology;Health Sciences;Biomedical Engineering

Joshi, AniketNoll, Douglas C. ;

University of Michigan

关键词: Positron Emission Tomography;    PET Pharmacokinetics;    PET Kinetic Modeling;    Neuropharamacological Measurements Using PET;    Neuroimaging;    PET for Clinical Diagnosis and Drug Discovery;    Radiology;    Health Sciences;    Biomedical Engineering;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/61729/adjoshi_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Positron emission tomography (PET) is a medical imaging modality offering a powerful tool for brain research by mapping of in vivo neuropharmacological functions such as metabolism, enzyme activity, and neuroreceptor binding site density and occupancy. Quantification in brain-PET can be classified into: 1) accurate quantification of radiotracer distribution such that image values are proportional to the radiotracer concentration in tissue, and 2) accurate quantification of the pharmacological state of the system-of-interest. This thesis addresses both these aspects for functional neuroreceptor imaging studies of the living brain.Traditional brain PET studies have at least two primary limitations. First, they measure only a single neuropharmacological aspect in isolation, which is often insufficient for characterizing a neurological condition. Second, data acquisition is accompanied by the invasive arterial blood sampling for measuring the input function to the system-of-interest. The motivation for this thesis was to address both these limitations, which led to the development of quantitative methods for multiple neuropharmacological PET studies performed without blood sampling. One such experimental design investigated was a dual-measurement intervention study where the system-of-interest is perturbed with the intent of changing the subject’s pharmacological status and system parameters are estimated both pre- and post-intervention. Second was a dual-tracer study where two radiotracers targeting two different neuropharmacological systems were injected closely in time in the same study.A major challenge in analyzing multiple pharmacological PET studies is the statistical noise-induced bias and variance in the parameter estimates. Methods developed in this thesis reduced almost all the bias (>90%) in the intervention studies with a corresponding improvement in precision. Parameter estimates for dual-tracer studies were obtained with inter-subject regions-of-interest means within ±10% of those obtained from single-tracer scans without appreciable increase in variance.The thesis also addresses inter-scanner PET image variability, a major confound in multi-center studies used to investigate disease progression. Since various PET centers have scanners with different hardware and software, systematic differences exist in multi-center data. This thesis develops a framework to reduce the inter-scanner PET image variability before pooling multi-center data for analysis. The methods developed reduced variability in phantom scans from different sites by approximately 50%.