OBJECTIVE: The purpose of the study was to study factors affecting SUV of PET imaging with 18F-FDG. MATERIAL AND METHOD: PET/CT Biograph 64 was used to acquire the data. A NEMA PET phantom with 6 spheres varying in diameter from 10 to 37 mm was used to mimic the human body and tumors. Background activity of 18F in the phantom was 0.14 microCi/ml and tumor-to-background ratios (TBR) of 2:1, 5:1 and 10:1 were studied. For each TBR, thirty sinograms were acquired with 3-min scan durations. Different scan durations varying from 3 to 20 min using a TBR of 5:1 were studied and three datasets of each scan time were collected Sinograms were reconstructed using the Ordered Subset Expectation Maximization (OSEM) algorithm with 5 mm Full-Width-at-Half-Maximum (FWHM) Gaussian filtering. Sinograms at TBR of 5:1 were reconstructed by varying the number of iterative updates of OSEM (N) from 8 to 168 and SUVavg and SUVmax were measured. The percentage of underestimation of SUVs was used to study the effect of tumor size and TBR. Intraclass correlation coefficient (ICC) was used to test the reliability of SUVmax with different scan durations. RESULTS: The results showed that both the SUVavg and SUVmax rapidly increased when N was <48 and slightly increased afterwards. At TBRs ranging from 2:l to 10:1, the percentages of underestimation of SUVmax ranged from 8.17 to 22.46 and that of SUVavg were ranged from 41.44 to 52.33 for 37-mm sphere and from 40.38 to 54.52 and from 48.97 to 67.73 for 10-mm sphere respectively. Different scan durations gave reliable SUVs(max) with ICC of 0.996. CONCLUSION: SUVs increased as N increased The percentage of underestimation of the SUV depended on tumor size and TBR. Scan duration did not affect SUVs.
OBJECTIVE: The purpose of the study was to study factors affecting SUV of PET imaging with 18F-FDG. MATERIAL AND METHOD: PET/CT Biograph 64 was used to acquire the data. A NEMA PET phantom with 6 spheres varying in diameter from 10 to 37 mm was used to mimic the human body and tumors. Background activity of 18F in the phantom was 0.14 microCi/ml and tumor-to-background ratios (TBR) of 2:1, 5:1 and 10:1 were studied. For each TBR, thirty sinograms were acquired with 3-min scan durations. Different scan durations varying from 3 to 20 min using a TBR of 5:1 were studied and three datasets of each scan time were collected Sinograms were reconstructed using the Ordered Subset Expectation Maximization (OSEM) algorithm with 5 mm Full-Width-at-Half-Maximum (FWHM) Gaussian filtering. Sinograms at TBR of 5:1 were reconstructed by varying the number of iterative updates of OSEM (N) from 8 to 168 and SUVavg and SUVmax were measured. The percentage of underestimation of SUVs was used to study the effect of tumor size and TBR. Intraclass correlation coefficient (ICC) was used to test the reliability of SUVmax with different scan durations. RESULTS: The results showed that both the SUVavg and SUVmax rapidly increased when N was <48 and slightly increased afterwards. At TBRs ranging from 2:l to 10:1, the percentages of underestimation of SUVmax ranged from 8.17 to 22.46 and that of SUVavg were ranged from 41.44 to 52.33 for 37-mm sphere and from 40.38 to 54.52 and from 48.97 to 67.73 for 10-mm sphere respectively. Different scan durations gave reliable SUVs(max) with ICC of 0.996. CONCLUSION: SUVs increased as N increased The percentage of underestimation of the SUV depended on tumor size and TBR. Scan duration did not affect SUVs.
Authors: Sied Kebir; Laurèl Rauschenbach; Martin Glas; Manuel Weber; Lazaros Lazaridis; Teresa Schmidt; Kathy Keyvani; Niklas Schäfer; Asma Milia; Lale Umutlu; Daniela Pierscianek; Martin Stuschke; Michael Forsting; Ulrich Sure; Christoph Kleinschnitz; Gerald Antoch; Patrick M Colletti; Domenico Rubello; Ken Herrmann; Ulrich Herrlinger; Björn Scheffler; Ralph A Bundschuh Journal: J Neurooncol Date: 2021-01-27 Impact factor: 4.130