Digital anthropomorphic phantoms of non-rigid human respiratory and voluntary body motions: A tool-set for investigating motion correction in 3D reconstruction

Patient respiratory and body motions occurring during emission tomography create artifacts in the images, which can mislead diagnosis. For example, in myocardial-perfusion imaging these artifacts can be mistaken for perfusion defects. Various software and hardware approaches have been developed to detect and compensate for motion. A practical way to test these methods is to simulate realistic motion with digital anthropomorphic phantoms. However, simulated motions often do not correspond to real patient motions. In this study, we are creating XCAT phantoms based on real body and respiratory motion data acquired from MR scans of volunteers. We are exploring different MRI acquisition methods to allow respiratory amplitude-binned modeling of both inspiration and expiration, which portrays both non-rigid motion and motion hysteresis during breathing. Simultaneous to MRI, the positions of reflective markers placed on the body-surface are tracked in 3D via stereo optical imaging. This enables correlation of the internal organ motion (e.g. heart, liver etc.) in our models with external marker-motion as would be observed during clinical imaging. Our digital anthropomorphic phantoms can serve as a realistic dataset with known truth for investigating motion correction in 3D iterative reconstructions.