Differences among [¹⁸F]FDG PET-derived parameters in lung cancer produced by three software packages

Investigation of differences in derived [¹⁸F]FDG PET metabolic and volumetric parameters among three different software programs in lung cancer. A retrospective analysis was performed on a group of 98 lung cancer patients who underwent a baseline [¹⁸F]FDG PET/CT study. To assess appropriate delineation methods, the NEMA phantom study was first performed using the following software: Philips EBW (Extended Brilliance Workstation), MIM Software and Rover. Based on this study, the best cut-off methods (dependent on tumour size) were selected, extracted and applied for lung cancer delineation. Several semiquantitative [¹⁸F]FDG parameters (SUV_max, SUV_mean, TLG and MTV) were assessed and compared among the three software programs. The parameters were assessed based on body weight (BW), lean body mass (LBM) and Bq/mL. Statistically significant differences were found in SUV_mean (LBM) between MIM Software and Rover (4.62 ± 2.15 vs 4.84 ± 1.20; p < 0.005), in SUV_mean (Bq/mL) between Rover and Philips EBW (21,852.30 ± 21,821.23 vs 19,274.81 ± 13,340.28; p < 0.005) and Rover and MIM Software (21,852.30 ± 21,821.23 vs 19,399.40 ± 10,051.30; p < 0.005), and in MTV between MIM Software and Philips EBW (19.87 ± 25.83 vs 78.82 ± 228.00; p = 0.0489). This study showed statistically significant differences in the estimation of semiquantitative parameters using three independent image analysis tools. These findings are important for performing further diagnostic and treatment procedures in lung cancer patients.