Abstract
1341
Objectives: To improve the reliability of an eight-unit PET/CT multi-center clinical trial, and to setup the reasonable control in the trial design, we evaluated the consistency of the image quality and the quantification of these eight PET/CT units.
Methods: One person did the phantom preparation, PET/CT scanning and data analyses using a PET NEMA image quality phantom (Biodex) on eight Discovery 690 PET/CTs from GE Healthcare. The 18F-NaF concentration in the background was 5.4kBq/ml, while the one in the balls with diameter 37mm, 22mm, 17mm and 10mm was 8 times higher and the one in the balls with diameter 28 mm and 13mm was 0kBq/ml. The PET/CT scanning parameters were the same with the clinical ones used in the multi-center trial. For PET scans, 2 min/bed was used with ordered-subset expectation maximization (OSEM) construction using 24 subsets, 2 iterations, with a cutoff frequency 6.4 Gaussian based post filter and with time-of-flight and point spread function technologies. For CT scans, 140 KVp bulb voltage and automatic bulb current (15-180mA) were used with a pitch of 0.984. The consistency of hot sphere recovery coefficient (HRC), cold sphere recovery coefficient (CRC), hot sphere contrast (QH) and cold sphere contrast (Qc) among these 8 PET/CTs were analyzed.
Results: The HRC of mean uptake (HRCmean) and that of CRC (CRCmean) of the balls were: HRCmean37mm=0.91±0.04, HRCmean22mm=0.74±0.04 , HRCmean17mm=0.67±0.03 , HRCmean10mm=0.46±0.04, CRCmean28mm=0.32±0.03 and CRCmean13mm=0.52±0.05. The variation of HRCmean and CRCmean among the 8 PET/CTs was less than 10%. The HRC of maximum uptake (HRCmax) of the balls were: HRCmax37mm=1.17±0.09, HRCmax22mm=1.12±0.11 , HRCmax17mm=1.12±0.09 , HRCmax10mm=0.64±0.09. For the balls with diameter 蠅17mm, the variations of HRCmax among the 8 PET/CTs were less than 10%. While for the ball with diameter 10mm, the variation of HRCmax was greater than 10%. The QH and Qc of the balls were: QH37mm =79.08±4.62%, QH22mm =69.08±4.83%, QH17mm =55.02±4.15%, QH10mm=32.31±8.51%, Qc28mm =69.08±3.70% and Qc13mm=48.28±5.84%. For the hot balls with diameter 蠅17mm, the variations of QH among the 8 PET/CTs were less than 10%. While for the cold balls with diameter 13mm and 10mm, the variations of Qc were greater than 10%. In addition, all HRCmaxs of the hot balls with diameter larger than 17mm were larger than 1. All performance parameters went down along with the decrease of sphere size.
Conclusion: The variation of the main image quality and quantification parameters of the eight PET/CT systems in this multi-center trial was within 10%. So, the multi-center trial results based upon these parameters will be reliable. It is suggested that SUVmean or SUVpeak should be used in this trial, instead of SUVmax. For the hot lesions with diameter larger than 17mm, SUVmax will over-estimate the uptake. While for the hot lesions with diameter less than 13mm, both the SUVmax and SUVmean will under-estimate the uptake. For the lesions with diameter less than 13mm, the variation among PET/CT systems is larger than 10% for some quantification parameters. The result from these small lesions should be careful. Research Support: