Defining optimal tracer activities in pediatric oncologic whole-body ¹⁸F-FDG-PET/MRI

To explore the feasibility of reducing administered tracer activities and to assess optimal activities for combined ¹⁸F-FDG-PET/MRI in pediatric oncology.

30 ¹⁸F-FDG-PET/MRI examinations were performed on 24 patients with known or suspected solid tumors (10 girls, 14 boys, age 12 ± 5.6 [1-18] years; PET scan duration: 4 min per bed position). Low-activity PET images were retrospectively simulated from the originally acquired data sets using randomized undersampling of list mode data. PET data of different simulated administered activities (0.25-2.5 MBq/kg body weight) were reconstructed with or without point spread function (PSF) modeling. Mean and maximum standardized uptake values (SUV_mean and SUV_max) as well as SUV variation (SUV_var) were measured in physiologic organs and focal FDG-avid lesions. Detectability of organ structures and of focal ¹⁸F-FDG-avid lesions as well as the occurrence of false-positive PET lesions were assessed at different simulated tracer activities.

Subjective image quality steadily declined with decreasing tracer activities. Compared to the originally acquired data sets, mean relative deviations of SUV_mean and SUV_max were below 5 % at ¹⁸F-FDG activities of 1.5 MBq/kg or higher. Over 95 % of anatomic structures and all pathologic focal lesions were detectable at 1.5 MBq/kg ¹⁸F-FDG. Detectability of anatomic structures and focal lesions was significantly improved using PSF. No false-positive focal lesions were observed at tracer activities of 1 MBq/kg ¹⁸F-FDG or higher. Administration of ¹⁸F-FDG activities of 1.5 MBq/kg is, thus, feasible without obvious diagnostic shortcomings, which is equivalent to a dose reduction of more than 50 % compared to current recommendations.

Significant reduction in administered ¹⁸F-FDG tracer activities is feasible in pediatric oncologic PET/MRI. Appropriate activities of ¹⁸F-FDG or other tracers for specific clinical questions have to be further established in selected patient populations.