Large PET/CT dataset statistics for the correlation between torso organ standardized uptake value and personal physiological features

ABSTRACT Background: The knowledge of [18F] fluoro-deoxy-glucose (18F-FDG) uptake range in normal torso organ is useful for metabolic disease diagnosis and treatment planning. This study aims to reveal the relationship between torso organ uptakes and personal physiological features, so as to quickly calculate the normal uptake value for patients according to his/her own physiological parameter. Materials and Methods: This study collected 767 18F-FDG positron emission tomography/computed tomography (PET/CT) images of normal people. The regions of torso organs were segmented and the mean values of classical body weight-normalized standard uptake value (W-SUV) and lean-body-mass-normalized standard uptake value (L-SUV) were calculated for each organ. The Pearson correlation coefficients and multiple regression formulas were calculated to reveal the influence of physiological features on W-SUV and L-SUV. Comparison between W-SUV and L-SUV were also conducted in terms of the correlation strength, regression confidence and inter-hospital consistency. Results: This study obtains the regression formulas for normal organ uptake calculation according to personal physiological parameters. The analysis reveals that lean body mass is the major factor affecting the uptakes of most organs (R > 0.5). It is also worthy to find that the L-SUV formulas have better prediction confidence and inter-hospital consistency than the W-SUV, indicating the advantage of L-SUV for organ uptake evaluation. Conclusion: Compared to the previously similar studies, this study clarifies the influential factors of torso organ 18F-FDG uptake. We derived detailed formulas for computing individually adapted normal organ SUVs. The comparison between L-SUV and W-SUV also highlights the stability and reliability of L-SUV for cross-center organ uptake analysis.