Optimizing PET scan protocol for small animal study: Correlation between scan duration, injected activity, target dimension, target-to-background ratio and imaging performance

Objectives Increasing the injected activity or the scanning duration is usually considered to achieve target imaging performance in an animal PET study, but may thus affect the processes under study due to the increased radiation and anesthetic dose. The aim of this study was to investigate how PET image quality behave when varying scan duration and injected activity.

Methods PET data sets of a cylindric water phantom (4 cm in diameter and 8.2 cm in height) including spheres with different object properties were simulated using the GATE Monte Carlo simulation software. A multivariate approach was used to investigate the simultaneous effects of target dimension (D_targ = 4, 5, 6, 8, 12 mm in diameter), target-to-background ratio (T/B ratio = 2, 4, 8, 12, 20), scan duration (T_acq = 30, 60, 90 min) and injected activity (A_acq = 4.7×10⁴, 9.3×10⁴, 1.4×10⁵ Bq/ml) on contrast-to-noise ratio (CNR) and background variability (BV).

Results The T/B ratio resulted as the most significant predictor of CNR, followed by D_targ, T_acq and A_acq of the given sphere. D_targ, T_acq and A_acq were the most important, the second most important and the least important predictor of BV, respectively. The fittings of both models were good with R² larger than 0.80 in both cases.

Conclusions Our preliminary results should provide practical guidance on how to define appropriate scan duration and radioactivity to achieve the target imaging performance when imaging object with different intrinsic characteristics