Objective: Many studies have documented the clinical usefulness of standardized uptake values (SUV) for diagnosis. However, in the event of injection error, accurate measurements cannot be obtained if the radioactivity of fluorodeoxyglucose (FDG) leakage is not subtracted from the administered dosage. Here, a correction formula for radioactivity estimation that takes into account the radioactivity of FDG leakage was derived on the basis of a phantom experiment. Furthermore, to determine whether SUV could be accurately calculated by the correction formula, we performed a volunteer study.
Methods: Images were displayed by altering the conversion constant from 1.0, 0.1 to 0.01, and the range of correctable counts was verified on the basis of image inversion. To estimate the radioactivity of FDG leakage by imaging, the count of the leakage was measured, converted into a radioactivity concentration using a cross-calibration factor (CCF), and multiplied by volume, as measured by imaging. Three factors that markedly affect count, i.e., count rate performance, partial volume effect and crosstalk, were assessed in phantom studies in order to derive a correction formula. In addition, to clarify the accuracy of the correction formula, we attached to the right elbow.
Results: With a conversion constant of 0.1, there was no image inversion at <or=1.565 MBq/ml. At concentrations below this, the average detection rate was 90%. This suggests that count rate performance can be corrected at <or=1.0 MBq/ml. Crosstalk investigations clarified that the effects of adjacent radioactivity concentrations on FDG leakage were not marked. On the basis of investigations on partial volume effect and count rate performance, the following formulas were derived: For leakages of >or=28 mm Leakage radioactivity (MBq)=positron emission tomography (PET) radioactivity (MBq)x0.9. For leakages of >or=15 mm but <28 mm Leakage radioactivity (MBq)=PET radioactivity (MBq)x0.9x(0.0517xleakage size (mm)-0.4029). In a volunteer study with 10 MBq leakage, SUV recalculated using the formula achieved 99.97% correction, whereas with 100 MBq leakage, SUV achieved 67.5% resulting in poor correction.
Conclusions: The present correction technique can accurately calculate SUV and could be useful for the clinical diagnosis of malignant tumors.