Impact of Time-of-Flight and Point-Spread-Function for Respiratory Artifact Reduction in PET/CT Imaging: Focus on Standardized Uptake Value

BACKGROUND: The most important advantage of positron emission tomography/computed tomography (PET/CT) imaging is its capability of quantitative analysis. The aim of the current study was to choose the proper standardized uptake value (SUV) threshold, when the time-of-flight (TOF) and point spread function (PSF) were used for respiratory artifact reduction in the liver dome in a new-generation PET/CT scanner. MATERIALS AND METHODS: The current study was conducted using a National Electrical Manufacturers Association International Electrotechnical Commission body phantom, with activity ratios of 2:1 and 4:1. A total of 27 patients, with respiratory artifacts in the thorax region, were analyzed. PET images were retrospectively reconstructed using either a high definition (HD) + PSF (i.e., a routine protocol) algorithm or HD+PSF+TOF (PSF+TOF; i.e., to reduce the respiratory artifact) algorithms, with various reconstruction parameters. The SUV(max) and SUV(mean), at different thresholds (i.e., at 45%, 50%, and 75%), were also assessed. RESULTS: Although in comparison to the routine protocol a higher SUV was observed when using the PSF+TOF method, this approach was used to reduce the respiratory artifact. The appropriate threshold for SUV was strongly related to the lesion size, reconstruction parameters, and activity ratio. The mean of the relative difference between PSF+TOF algorithm and routine protocol for SUV(max) varied from 10.58±14.99% up to 35.49±32.60% (which was dependent on reconstruction parameters). CONCLUSION: In comparison with other types of SUVs, the SUV(max) value illustrated its significant overestimation, especially at the 4:1 activity ratio. The poor agreement between SUV(max) and SUV(50%) was also observed. When the TOF and PSF are utilized to reduce respiratory artifacts, the SUV(50%) can be an accurate semi-quantitative parameter for PET/CT images, for all lesion sizes. For smaller lesions, however, a smaller filter size was required to observe an accurate SUV.