Temperature dependence of the LabPET small-animal PET scanner

INTRODUCTION In quantitative PET imaging it is important to correct for all image-degrading effects, for example detector efficiency variation. Detector efficiency variation depends on the stability of detector efficiency when operating conditions vary within normal limits. As the efficiency of APD-based light detection strongly depends on ambient temperature, temperature-dependent detector efficiency normalization may be needed in APD-based PET scanners. We have investigated the temperature dependence of the LabPET APD-based small-animal PET scanner. MATERIALS AND METHODS First a simulation study was performed to evaluate the effect of different APD temperature coefficients on the temperature dependence of scanner sensitivity. Five experiments were also performed. First the immediate effect of temperature changes on scanner sensitivity was evaluated. Second, the effect of temperature changes that have stabilized for a few hours was investigated. In a third experiment the axial sensitivity profile was acquired at 21°C and 24°C. Next, two acquisitions of the NEMA image quality phantom (at 21°C and 23°C) were performed and absolute quantification was done based on normalization scans acquired at the correct and incorrect temperature. Finally, the feasibility of maintaining a constant room temperature and the stability of the scanner sensitivity under constant room temperature was evaluated. RESULTS Simulations showed that the relation between temperature-dependent APD gain changes and scanner sensitivity is quite complex. A temperature deviation leading to a 1 % change in APD gain corresponds to a much larger change in scanner sensitivity due to the shape of the energy histogram. In the first and second experiment a strong correlation between temperature and scanner sensitivity was observed. Changes of 2.24 kcps/MBq and 1.64 kcps/MBq per °C were seen for immediate and stabilized temperature changes respectively. The NEMA axial - ensitivity profile also showed a decrease in sensitivity at higher temperature. The quantification experiment showed that a larger quantification error (up to 13%) results when a normalization scan acquired at the incorrect temperature is used. In the last experiment, temperature variability was 0.19°C and counts varied by 10.2 Mcts (1.33%). CONCLUSION The sensitivity of the LabPET small-animal PET scanner strongly depends on room temperature. Therefore, room temperature should be kept as stable as possible and temperature-dependent detector efficiency normalization should be used. However, with constant room temperature excellent scanner stability is observed. Temperature should be kept constant within 0.5°C and weekly normalization scans are recommended.