Variability of lesion detectability and standardized uptake value according to the acquisition procedure and reconstruction among five PET scanners

The objective of this study was to assess differences in the semiquantitative values of 18F-fluorodeoxyglucose (18F-FDG) uptake among different positron emission tomographic (PET) systems. A phantom study was performed to compare standardized uptake value (SUV) in five PET scanners including a dedicated PET scanner and four PET/computed tomography (CT) scanners. Radioactivity simulating the SUV of 2.5 was filled in the hot spheres (8 mm, 11 mm, 14 mm, 18 mm, 22 mm, and 27 mm) that were set in the cylindrical phantom with the background SUV of 1.0. Data acquisition and reconstruction were performed according to routine and standardized conditions. The standardized condition was as follows: CT acquisition (120 kVp, 50 mA) and PET acquisition (2-min acquisition with a slice thickness of 2 mm); reconstruction was performed by ordered subsets expectation maximization+Fourier rebinding. Detectability of hot spheres and SUV was compared between routine condition and standardized condition with five PET scanners. On routine condition, two cameras could detect a 14-mm sphere clearly. On the other hand, the visualization of hot spheres by the standardized condition was remarkably variable. Semiquantitative evaluation revealed that a maximum of 45.7% error was recognized with the 27-mm sphere by the routine condition, although the standardized condition could reduce the error to 22.6%. Detectability depends not only on the PET machine but also on the imaging protocol. The results indicate that SUV is variable with PET machines under routine conditions of data acquisition and reconstruction. Standardization of the reconditions can reduce variability and maximum difference in the SUV by half. (author)