Performance evaluation of the Philips ´Gemini´ PET/CT System

Methods to test the practical performance of the Philips Gemini PET/CT system during clinical imaging are described and results presented. Measurements were performed using the EEC Emission phantom, alongside phantoms provided by Philips. The test methods used were based on NEMA (NU2-2001) and those suggested by a task group of the commission of European Communities (EEC). The clinical protocols in use were based on the single-pass emission-transmission whole body scanning protocols provided by Philips. Images are reconstructed using a 3D row action maximum likelihood algorithm (RAMLA). Attenuation correction was applied using both CT data (CTAC) and Cs-137 transmission scans (CsAC) to allow a comparison of the two methods. Images can be acquired with the PET and CT gantries adjacent (closed) or separated by 60 cm (open); geometric calibrations show that gantry alignment is most consistent in the closed position. Results show the non-uniformity of the images varied across the slices, for the non-attenuation corrected (NOAC) images, but that uniformity was improved and more consistent in the both the CTAC and CsAC images. The reconstructed scatter fractions show that the RSF was better for NOAC (0.15) than for CsAC (0.18) and CTAC (0.22), the sinogram derived scatter fraction was 0.36. At the centre the transverse resolution was 5.3 mm in air (5.4 mm in water) and the axial resolution was 5.8 mm (6.5 mm). Nearer the FOV periphery the transverse radial resolution was 5.3 mm (5.3 mm), the transverse tangential resolution was 5.3 mm (5.5 mm) and the axial resolution was 7.5 mm (7.4 mm). The recovery coefficients show that partial volume only affects objects smaller than 22 mm.