Quantitative precision and least significant change of ¹⁸F-FDG PET/CT study in a rabbit VX2 tumor model

Objectives: To investigate the precision and least significant change (LSC) of metabolic parameters for assessment of tumor response in a rabbit VX2 tumor model.

Methods: Repeated ¹⁸F-FDG PET/CT scans were performed for 3 consecutive days in 15 VX2 tumor-bearing rabbits. All metabolic parameters used for assessment of solid tumor response were measured and their precision and LSC were calculated. DXA scans of rabbits and pathological specimens of tumor tissue were also collected. Two-way ANOVA was used for statistical intra-group comparison.

Results: The tumor growth rate of rabbit VX2 tumor model was 100%. The precision of SUV parameters, including SUVmax, SUVmean and SUVpeak ranged from 18.3% to 18.8%, which was similar to that (18.0%-18.4%) of SUL parameter. The precision of TLG ranged from 26.4% to 26.8%, higher than other quantitative metabolic parameters. Using 80% confidence interval, the LSC (RMS-CV and RMS-SD) of SUVmax and SULpeak were 33.1%-34.1% and 1.34-2.10, 32.6%-33.3% and 1.25-1.96, respectively. Using 95% confidence interval, the LSC (RMS-CV and RMS-SD) of SUVmax and SULpeak were 50.1%-52.1% and 2.04-3.22, 50.0%-51.0%and 1.91-3.00, respectively.

Conclusion: A novel technique to evaluate the quantitative precision of solid tumor response to therapy in animal PET imaging was established. TLG was proved to be less precise than other metabolic parameters in our study. Further investigations were required to determine the proper confidence interval to calculate LSC in clinical practice. Research Support: No