Iterative reconstruction with a rotating open-ring PET scanner for proton therapy range verification

Objective. In-beam positron emission tomography (PET) is a leading contender for non-invasive monitoring of proton therapy. The design constraints for in-beam PET have led to the development of an open-ring scanner as part of the PETITION project, which provides the unique opportunity to image immediately following the delivery of each treatment field. This study introduces computational techniques for the reconstruction of data recorded between delivery of each treatment field by a clinically implementable rotating open-ring design. Our proposed design enables fully-3D imaging of activity induced during proton therapy for verification on a field-by-field basis. Approach. We introduce a modification of the maximum likelihood expectation maximisation (MLEM) algorithm which accounts for imaging a source consisting of multiple isotopes with a rotating open-ring PET system. Pre-calculated system matrices are used to perform timely reconstructions for inter-field and post irradiation imaging for verification. We show the capabilities of our system by simulating a Derenzo-like phantom, as well as the treatment of a superatentorial neoplasm with 3 fields. We also show a proof of principle experimental measurement of a single field delivered to a CIRS 731-HN phantom. Main results. We show an increase in image quality when compared to fixed position imaging with an open-ring scanner, making the rotating open-ring design comparable to a similar full-ring design. The normalised root mean square error (NRMSE) was a factor of 1.6-2.2 better in comparison to imaging in a fixed position with an open-ring scanner. The new MLEM implementation was capable of assigning range to within an average of 0.6 mm along the beam direction for all fields, allowing for range verification of multiple fields. Significance. We have introduced a novel MLEM algorithm for imaging with a rotating open-ring PET device, opening the way for a clinically feasible implementation of fully-3D PET imaging of all treatment fields for verification of proton therapy.