Respiratory Motion Correction in Emission Tomography Imaging

During an emission tomography exam of lungs, respiratory motion blurs the reconstructed image, which may lead to misinterpretations and imprecise diagnosis. Solutions like respiratory gating, correlated dynamic PET techniques, listmode data based techniques and others have been tested with improvements over the spatial activity distribution in lungs lesions, but with the disadvantages of either requiring extra hardware or more expensive scanner systems, or discarding part of the acquired data. The objective of this study is to include a respiratory motion model in the image tomographic reconstruction process, which will allow to consider all the acquired data, without any additional requirement. To this end, we propose an extension of the MLEM reconstruction algorithm, in which we extend the probability matrix to take into account both the motion and the deformation of the voxels to be reconstructed. We present results from synthetic simulations incorporating real respiratory motion data, demonstrating the potential benefits of such an approach.