NEMA NU 2-2001 performance testing of a Philips Gemini GXL PET/CT scanner

Post installation acceptance testing is vital to demonstrate that the equipment meets the vendor's specification and is suitable for clinical studies. The test procedures described in the NEMA NU 2-2001 document form the basis of vendor performance specifications of PET scanners and hence are also appropriate for acceptance testing. Initial installation performance tests of the Philips Gemini GXL PET/CT scanner installed at Liverpool Hospital revealed that the peak noise equivalent count rate (NECR) measurement of 57.5 kcps was substantially lower than the specification of 70 kcps and the scatter fraction of 38.5% was 10% higher than the specification of <= 35%. These performance parameters potentially impact on clinical image quality and hence the deviations were considered unacceptable. The performance issues were addressed by the vendor through a hardware upgrade and optimisation of the energy window by increasing the lower discriminator value from 408.8 to 434.4 keV. Several repeat performance measurements taken from post-optimisation demonstrated improvement in peak NECR to 67 kcps with delayed window randoms subtraction on and 72 kcps with delayed window randoms subtraction off. Scatter fraction taken from post upgrade and optimisation improved to an average of 33%, which is well within specification of <35%. All other performance tests (resolution, sensitivity, accuracy of corrections) were within specifications both prior and post-optimisation changes. This experience demonstrates the importance of equipment acceptance testing prior to clinical use so that deficiencies in performance can be addressed before the equipment is placed in clinical service.