Improved correction of post-injection transmission data for PET scanners without measured random correction

The most accurate attenuation correction in positron emission tomography (PET) is achieved by measuring a transmission scan with a rotating rod source in combination with a blank scan using the rod windowing technique. In clinical routine, however, transmission scans often have to be acquired after the administration of the radiotracer. The raw data recorded this way are biased by coincidences generated by the additional activity in the scanner's field of view. In a first approximation, these additional true coincidences can be removed by subtracting a decay-corrected emission scan from the post-injection transmission data. But there are also additional 'mixed' random events generated by the single photon fluxes from the patient activity and the rod source together. In this work, the correction for post-injection transmission data was extended to take into account the effect of the rod windowing and the 'mixed' randoms. A couple of experiments were performed to determine the reduced scanner sensitivity for the additional true and random events with active rod windowing. The resulting correction, which can be used for most tomographs without a measured random correction, was validated in a study with the EEC whole-body phantom.