Improved image quality with optimal energy window in a compact brain-dedicated TOF-PET

To make a high-sensitivity brain PET system with a limited number of detectors, we have proposed and developed a helmet-type TOF-PET prototype for which the coincidence timing resolution (CTR) was 245 ps. The energy resolution of 12.6% was obtained with the energy window of 420-590 keV. For such a compact system, the scattered coincidence counts increase due to the increased solid angle coverage. Our aim in this study was to investigate an optimal energy window of this prototype in order to obtain better count rate performance and better image quality. CTR, noise-equivalent count rate with TOF gain (TOF-NECR), and image quality were evaluated with 16 different patterns of the energy windows (range, 400-700 keV). A brain-sized image quality (IQ) phantom and a hemispherical Hoffman phantom were used for image quality evaluation. The highest TOF-NECR was obtained with the energy window of 440-590 keV, and the CTR was improved to 236.1 ps. Comparing the image quality between the previous (420-590 keV) and the optimized (440-590 keV) energy windows, the background variability of the IQ phantom image was reduced by 5.4%, and the contrast in the Hoffman phantom image was increased by 14.5%. In conclusion, we significantly enhanced imaging performance of the helmet-type TOF-PET prototype by using the optimized energy window. We found that the lower energy window setting was a key factor to get a higher TOF-NECR and better image quality in compact TOF-PET systems.2020 IEEE Nuclear Science Symposium & Medical Imaging Conferenc