Motion estimation and correction for simultaneous PET/MR using SIRF and CIL

Abstract

SIRF is a powerful PET/MR image reconstruction research tool for processing data and developing new algorithms. In this research, new developments to SIRF are presented, with focus on motion estimation and correction. SIRF’s recent inclusion of the adjoint of the resampling operator allows gradient propagation through resampling, enabling the MCIR technique. Another enhancement enabled registering and resampling of complex images, suitable for MRI. Furthermore, SIRF’s integration with the optimization library CIL enables the use of novel algorithms. Finally, SPM is now supported, in addition to NiftyReg, for registration. Results of MR and PET MCIR reconstructions are presented, using FISTA and PDHG, respectively. These demonstrate the advantages of incorporating motion correction and variational and structural priors. This article is part of the theme issue ‘Synergistic tomographic image reconstruction: part 2’.

Core Imaging Library - Part I: a versatile Python framework for tomographic imaging

“…The purpose of this example was to give proof of principle of prototyping new reconstruction methods for PET with SIRF, using the generic algorithms of CIL, without needing to implement dedicated new algorithms in SIRF. Another example with SIRF for PET/MR motion compensation employing CIL is given in [ 19 ] within this issue.…”

“…Multi-channel functionality (e.g. for dynamic and spectral CT) is presented in the part II paper [ 18 ] and a use case of CIL for PET/MR motion compensation is given in [ 19 ], both within this same issue; further applications of CIL in hyperspectral X-ray and neutron tomography are presented in [ 20 , 21 ].…”

Synergistic motion compensation strategies for positron emission tomography when acquired simultaneously with magnetic resonance imaging

“…More information is provided by the collaborative computational project in synergistic image reconstruction for biomedical imaging (CCP SyneRBI: https://www.ccpsynerbi.ac.uk/), which includes open source datasets (e.g. among others: https://www.isd.kcl.ac.uk/pet-mri/simulated-data/, [101]) and software for PET and MRI image reconstruction, and motion estimation and compensation [100,102,103]. What the scientific community lacks is the availability of reliable physical test objects with realistic motion capabilities mimicking human tissue deformations for all different types of motion that can be controllable externally and can be scanned with MRI, PET and CT producing realistic acquisition data and motion artefacts.…”

Core Imaging Library -- Part II: Multichannel reconstruction for dynamic and spectral tomography

“…In order to overcome this problem, we employ the SPDHG algorithm, where the above operations are applied to a randomly selected subset of the data in every iteration. It has been used to different clinical imaging applications, such as PET, [24] and motion estimation/correction in PET/MR, [25], and produce significant computational improvements compared to the PDHG algorithm. The setup of the SPDHG algorithm is very similar to the PDHG algorithm.…”