MR–Consistent Simultaneous Reconstruction of Attenuation and Activity for Non–TOF PET/MR

Abstract

Attenuation correction (AC) is required for accurate quantification of the reconstructed activity distribution in positron emission tomography (PET). For simultaneous PET/magnetic resonance (MR), however, AC is challenging, since the MR images do not provide direct information on the attenuating properties of the underlying tissue. Standard MR–based AC does not account for the presence of bone and thus leads to an underestimation of the activity distribution. To improve quantification for non–time–of–flight PET/MR, we propose an algorithm which simultaneously reconstructs activity and attenuation distribution from the PET emission data using available MR images as anatomical prior information. The MR information is used to derive voxel–dependent expectations on the attenuation coefficients. The expectations are modeled using Gaussian–like probability functions. An iterative reconstruction scheme incorporating the prior information on the attenuation coefficients is used to update attenuation and activity distribution in an alternating manner. We tested and evaluated the proposed algorithm for simulated 3D PET data of the head and the pelvis region. Activity deviations were below 5% in soft tissue and lesions compared to the ground truth whereas standard MR–based AC resulted in activity underestimation values of up to 12%.