Comparison of the variability of SUVs normalized by skeletal volume with that of SUVs normalized by body weight in normal 18 F-fluoride PET/CT studies.

Objective: to test the hypothesis that the variability of the standardized uptake values (SUVs) normalized by skeletal volume (SV) in 18 F-fluoride PET/CT (18 F-NaF) studies is lower than that of the SUVs normalized by body weight (BW). Methods: the SUVs were obtained in whole skeletal volumes of interest (wsVOIs) in 163 selected 18 F-NaF studies. These studies were performed for the investigation of bone metastases and considered normal by evaluation of PET and CT images. The SUVs were calculated using normalization by BW, by SV and without normalization (WN). The total skeletal volume for each patient was also estimated based on the volume of the wsVOIs defined on the CT component of the PET/CT study. The SUV variability estimate for each patient was calculated in the following manner: the absolute value of the difference between the SUV for each patient and the mean of the SUVs in the whole group of patients divided by the mean of the SUVs in the whole group of patients. The variabilities of the SUVs calculated by the three methods were compared using paired one-tailed Wilcoxon. Results: the means of the variabilities for the SUVs BW, SV and WN were 0.16, 0.13 and 0.16, respectively. There were statistically significant differences between the variabilities for SUV SV and BW ( P = 0.03), and between the variabilities for SUV SV and WN (p<0.01). There was no statistically significant difference between the variabilities for SUV BW and WN ( P = 0.4). Conclusion: In patients with normal 18 F-NaF studies, the SUVs normalized by skeletal volume present lower variability than the SUVs normalized by body weight.