Effect of PSF and TOF on the PET/CT image quality at various noise levels

Objectives The aim of this study was to investigate the effect of the point-spread-function (PSF) and time-of-flight (TOF) on improving the PET/CT image quality at various noise levels using a large body phantom measuring 40 cm in diameter.

Methods We used a Biograph mCT and studied two phantom models; including a NEMA IEC body phantom and a large 40 cm diameter phantom with a built-in NEMA IEC body phantom (background activity was 5.31 or 2.65 kBq/mL, respectively). The PET data were acquired in the 3-dimensional list-mode and then were reconstructed with the baseline OSEM algorithm, with the OSEM+PSF algorithm and with the OSEM+PSF+TOF algorithm. The NECphantom, noise (CV) and contrast were all assessed in relation to the phantom size, radioactivity, acquisition times and iterations.

Results The NECphantom and the noise were inferior in the large body phantom and low activity models. In comparison to the CV of OSEM algorithm with 7.17% (high activity, body phantom), 8.64% (low activity, body phantom), 13.1% (high activity, large phantom) and 16.6% (low activity, large phantom), these of OSEM+PSF+TOF algorithm were 6.33%, 7.39%, 11.7% and 14.3%, respectively. PSF and TOF diminished the noise, especially for the large phantom, low activity models and short acquisition time. TOF provided a higher contrast recovery on a 10 mm sphere at an acceptable noise level.

Conclusions PSF and TOF provided a significant improvement in the quality of the both the body phantom and the large body phantom images, especially in the large body phantom, the low activity models and short acquisition time. PSF and TOF may therefore be potentially useful for obtaining good image quality, and thus are expected to reduce both patient inconvenience and the radiation burden