Meeting Banner
Abstract #4367

Consistency Conditions in MRI-Based Attenuation Correction for Hybrid PET/MRI Scanners

Zhengyi Yang1, Viktor Vegh1, Ian Turner2, David Reutens1

1Center for Advanced imaging, The University of Queensland, Brisbane, Queensland, Australia; 2School of Mathematical Sciences, Queensland University of Technology, Brisbane, Queensland, Australia

Attenuation correction in PET/MRI imaging is a challenging task because the attenuation map is difficult to be calculated due to the lack of information on tissue classification in MRI image. Consistency conditions on PET emission data contain useful information on the attenuation map and can be employed to regularize the creation of MRI-based attenuation map. In this study, the efficacy of consistency conditions were validated on simulated images of normal brain and brain with lesion. The results indicated these conditions are insensitive to brain lesions in finding the attenuation coefficients of each tissue class.

Abstract PDF E-Poster Video

Keywords

accuracy accurate accurately additional advanced affect appropriate assisted attenuation azimuthal bone brain calculation cancer capabilities challenging class classification clinics coefficient coefficients combination combinations combined component computed computer conditions consistency constant constraints construction containing corrected correction create cross currently dataset deep derived described detection development directly discontinuity distance drawbacks earlier efficacy embedding emission entries errors estimating example exist exposure extensive finding fixed formulation forward functional global ground hurdle hybrid implying inaccurate inadequate inconsistency index indicate intensity ionizing journal leading lesion lesions linked mathematical maximizing measure measured measures medical metabolic metabolically minimized moment moments much object occur organized patient patients periodic photon physics positron projection projections radial radiation radiology radon reality relies representing response robust rotation rules sampling satisfy scanners school section sections sensitivity serious significantly simulated slice soft space staging structural studies subjects suffers task technology threshold tissue tomography transform transmission truth turner understood uptake varies various vary violation width yang zero