Abstract
An important aspect of positron emission tomography (PET) imaging in a clinical application is the localization and detection of tumors and lesions by administering a predetermined amount of radiotracer. This allows detailed 3D imaging of a wide range of molecular processes in the human body. The quality of the PET image is dependent on the amount of radiotracer administrated and the patient's body parameters. As the amount of injected radiotracer increases, an overall improvement in the quality of the reconstructed PET images and lesion detectability is expected, but it is accepted that any radiotracer doses are associated with the risk of radiation and it could be harmful to the patient if essential PET imaging is not performed because of the fear of radiation risk. To ensure the highest-quality diagnosis and the smallest radiation risk, the patient should receive the smallest amount of radiotracer that provides an image of sufficient quality. Our study proposed a PET simulation tool to predict the smallest amount of radiotracer that allows for a reliable diagnosis based on patients' significant body parameters (weight, age) within a fixed total scan time to improve diagnostic processes for detecting and localizing tumors. We built a model of a particular PET scanner and patient, based on real MRI images and a digital anthropomorphic phantom of the brain. We performed Monte Carlo simulations of PET data acquisitions. A dataset of 60 patients was used, and 11 independent dose prediction simulations were performed for each patient. We concluded that our simulator estimated injected radiotracer doses 28% smaller than the standard clinical doses that yielded PET images of clinically acceptable quality. We also found that the total injected radiotracer dose for adult patients was affected by considering the patient's weight rather than age.