Reconstruction from truncated projections using constrained total-variation minimization applied to PET for hadron-therapy monitoring

Hadron-therapy exploits ions to treat tumors by maximizing the dose released to the target and sparing healthy tissues. With hadrons, the dose distribution rises sharply at the end of the range, providing the characteristic Bragg peak, and drops quickly thereafter to a negligible value. During hadron-therapy, short-lived β⁺-emitters are produced along the beam path. Following positron annihilation, two photons are emitted, which can be detected using a PET scanner. The low yield of β⁺-emitters and the washout from the target region make the use of PET a few minutes after hadron-irradiation a challenging application. In-beam PET represents a potential candidate to measure the produced β⁺-emitters during irradiation, at the cost of truncation effects and degraded image quality due to the partial-rings of the PET scanner. Time-of-Flight (ToF) information can potentially be used to enhance image contrast and to compensate for truncation effects. However, the highly demanding timing performance that detectors require to be used in ToF-PET makes this option very costly. Alternatively, including total variation (TV) as prior information in the MAP reconstruction produces images with low noise, without degrading spatial resolution. In this work we compare ML-EM with ToF information and MAP using TV as prior from data acquired with a complete and a partial-rings of Gemini TF, of simulated proton-beams interacting in a PMMA target. Results show that MAP-TV, in the absence of ToF information, produces lower noise and higher correlation coefficients compared to ML-EM with ToF information (in the order of 400-600 ps). Moreover, MAP-TV outperforms MLEM with a ToF of 200 ps in some cases.