First 4D In-beam PET Measurement for Beam Tracking of a Moving Phantom with a Scanned Carbon Ion Beam

More than 10 years of clinical operation of in-beam PET at GSI Darmstadt have proven its positive impact on quality assurance of carbon ion therapy, mostly for head-and-neek sites. Due to the promise of ion beam therapy for indications such as lung and liver tumors which are influenced by respiratory motion we started to investigate the potential of time-resolved, 4D in-beam PET. 4D in-beam PET is expected to facilitate in-vivo assessment of tumor miss or unwanted involvement of nearby critical structures in the presence of organ motion. In a first experiment performed at GSI, in-beam PET was used in combination with beam tracking. A homogeneous PMMA phantom was placed in the center of the field of view of the PET camera and moved parallel to the two detector heads (left-right in beam's eye view, amplitude: 3 cm peak-to-peak, period: similar to 3 s). Dose was delivered by beam tracking as a two-dimensionally spread-out Bragg-peak of 5 cm x 5 cm extension, centered at a depth of 10 cm in the central plane of the phantom. The dynamic PET acquisition was performed during the 6 min of beam delivery and for 25 min after irradiation. The data stream was synchronized with the time course of the dynamic beam application and with the phantom motion. Additional 18 min of decay were acquired with the activated phantom kept steady, for comparison with a separate measurement of the same treatment field without motion. Reconstruction of the data taken with a steady phantom yielded comparable activation patterns after beam tracking and stationary irradiation. For the acquisition under phantom motion with beam tracking, motion phase-sorted 4D PET reconstruction with 4D attenuation correction has been implemented. Summation of the phase-sorted distributions co-registered to the motion phase of the stationary reference irradiation shows the feasibility of 4D in-beam PET for recovery of the volumetric extension of fields delivered to moving targets. As for conventional 3D in-beam PET, counting statistics is a critical issue for the achievable accuracy in treatment verification.