A0672
Title: Statistical modeling for positronium lifetimeimage reconstruction using time-of-flight positron emission tomography
Authors: Hsin-Hsiung Huang - University of Central Florida (United States) [presenting]
Abstract: Positron emission tomography (PET) has been widely used to diagnose serious diseases, including cancer and Alzheimer's disease, based on the uptake of radiolabeled molecules that target certain pathological signatures. Recently, a novel imaging mode known as positronium lifetime imaging (PLI) has also been shown to be possible with time-of-flight (TOF) PET. PLI is also of practical interest because it can provide complementary disease information reflecting conditions of the tissue microenvironment via mechanisms that are independent of tracer uptake. However, for the present practical systems with a finite TOF resolution, the PLI reconstruction problem has yet to be fully formulated to develop accurate reconstruction algorithms. This paper addresses this challenge by developing a statistical model for the PLI data and deriving from it a maximum-likelihood algorithm for reconstructing lifetime images alongside the uptake images. Using realistic computer simulation data shows that the proposed algorithm can produce quantitatively accurate lifetime images for a 570 ps TOF PET system. The recent findings about the parameter effects for the reconstruction bias and variance analysis and the real-world PLI image with mixed exponential distributions are also presented.
