Prediction of Radioactive Injection Dosage for PET Image

Advanced computer and imaging techniques find extensive use in medicine. Medical imaging modalities such as positron Emission Tomography (PET) are becoming an increasingly important component of clinical applications and research oncology for diagnosis, treatment planning, and tumor monitoring in order to gather details about the process of the patient body whether it is a disease or normal physiological process. An important aspect of PET imaging in clinical application is the localization and detection of tumors and lesions by administering a predetermined amount of radiotracer. This allows, for example, a detailed view of what is going inside the patient body in cellular level. The quality of PET image is strongly dependent on the amount of administrated radiotracer and the patient’s body parameters. As the amount of injection radiotracer increases, the quality of resulting image increases and the lesion detection efficiency increases. The PET examiner society recognizes that any dose of radiotracer is associated with some possible radiation risks. It can harmful to the patient if essential PET imaging session is not made due to fear of radiation risk. In order to ensure the highest quality diagnosis and the smallest radiation risk, the patient should receive the smallest amount of radiotracer that provides image with sufficient quality. Our study is focused on proposing an efficient PET simulation tool that predicts the smallest possible amount of administrated radiotracer to provides the appropriate diagnostic information based on significant patient’s body parameters (weight, age) at fixed scanning to improve the clinical diagnostic process in term of tumor-detecting and localization. We have built a model of particular PET scanner and model of a patient based on real MRI image and digital anthropomorphic phantom of our region of interest (brain). We have performed Monte Carlo simulation for whole PET procedure with a special parameter. In validation stage, we have analyzed the system performance (in term of spatial resolution, sensitivity, and scatter fraction). In evaluating stage, a dataset of 60 patients is used and 11 independent dose prediction simulations for each patient are performed. We conclude that our simulator performs a desirable and efficient prediction of injection radiotracer amount that optimizes the current clinical amount up to 28%. In addition, we found that the total injected radiotracer dosage for adult patients are mostly affected when considering patient weight rather than patient age.