The interpretation of quantitative imaging data obtained from Positron Emission Tomography (PET) studies requires an understanding of the measurement variability due to instrumental effects, especially in clinical trials that involve large numbers of patients being scanned in several sites with different types of scanners using a variety of acquisition and analysis techniques.
Although the data from most PET studies are often used on a relative basis, the quality of the analysis will be greatly improved if the data can be normalized using a common reference standard that allows instrumental variability to be minimized. This requires long-lived PET phantoms to be available to all sites throughout the entire study and the activity content in the phantoms to be linked to a single standard.
Working in collaboration with a university medical center and a commercial source manufacturer, we have designed and calibrated two prototype large volume (30 cm length x 20 cm diameter, approximately 9 L) cylindrical phantom sources containing 68Ge in epoxy as a surrogate for 18F that can be used to accomplish these tasks. The design is similar to the phantoms used clinically with 18F to calibrate the PET scanners and are the first ones having a calibrated value for the amount of 68Ge in the phantom that is directly traceable to a national standard for that radionuclide.
Because of their large physical volume, direct calibration measurements on the phantoms were deemed unlikely to satisfy the requirement of having a combined uncertainty on the 68Ge activity concentration of about 1 %. Therefore, a sampling procedure was adopted in which samples of the epoxy containing 68Ge were taken during the phantom preparation process and dispensed into a standardized geometry for calibration. The epoxy sources were measured on several HPGe systems that been calibrated for this specific geometry using a previously standardized 68Ge solution. Corrections were made for possible differences in attenuation and scattering arising from the differences in composition and density between the epoxy and solution sources. From the known mass of the epoxy in each of the samples and these activity measurements, the activity concentration of the epoxy in the phantoms was determined with a combined standard uncertainty of less than 1 %. The relative difference in activity concentration between the two prototypes, as well as the measured variability in the activity concentration in each phantom (uniformity), was confirmed by scanning the phantoms in the NIST PET-CT scanner.
Using this new procedure, we have begun to calibrate several types of 68Ge epoxy phantoms for different manufacturers. These are now becoming commercially available and will provide the nuclear medicine community with the necessary tools to calibrate and monitor the performance of their scanners during clinical trials in a way that is traceable to national standards.