How is SUV affected by counting statistics in current clinical practice

1226 Objectives Three-dimensional (3D) PET-CT studies using dose suitable for 2D acquisition has resulted in enhanced sensitivity (1). However, it is not clear how quantitation is affected by a decrease in dose or the imaging time. We have evaluated the precision of standardized uptake value (SUV) with the objective of reducing the dose using 3D acquisition. Methods Twenty five consecutive PET-CT studies using 15 mCi F18-FDG were evaluated retrospectively. All studies were acquired in list mode (LM) and images were reconstructed with 0.5, 1, 1.5, 2, 2.5, and 3-min-per bed position. They were evaluated by a board certified nuclear medicine physician, who was aware of the clinical diagnoses, in terms of the image quality with selected lesions measured in size, maximum SUV (SUVm), and average SUV (SUVavg) at each time interval. A large region of interest was drawn in the normal liver and SUVavg as well as SUVm were quantified to serve as a reference to biodistribution. Results Statistically significant differences in SUVm and SUVavg were found only at 0.5 min compared to 3 min acquisition with variation at other time intervals decreasing as they approach to 3 min. Alternatively, the Bland Altman plots demonstrated decreasing mean differences, and hence, smaller standard deviation from these differences, to be in very good agreement between 2, 2.5 and 3 min acquisitions. Conclusions We have demonstrated a threshold beyond which the quantitation can be confidently reported with minimal variation. Furthermore, we can confidently decrease either injected dose or imaging time without compromising the validity of SUVs from previous studies