TY - JOUR T1 - Multi PET scanner statistical analysis comparison JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 2046 LP - 2046 VL - 52 IS - supplement 1 AU - Stephen Adler AU - Roberto Maass-Moreno AU - P. Choyke AU - Karen Kurdziel Y1 - 2011/05/01 UR - http://jnm.snmjournals.org/content/52/supplement_1/2046.abstract N2 - 2046 Objectives Obtaining PET images of equivalent quality among different scanners and reconstruction methods may require different injected dose and acquisition time combinations. To guide this choice in a systematic and reproducible way, a method is proposed based on a simple characterization of the scanner image signal-to-noise as a function of the count density exposure in the FOV. Methods The NEMA NU2 image quality phantom was used in the study. It was scanned with varying background activity levels and in whole body mode, using varying scan times/bed position (BP). The scans were done on a Philips Gemini TF, a GE Discovery ST and a Siemens Biograph mCT and images reconstructed using FBP and iterative algorithms (ITER). Plots of background signal-to-noise squared (SN2) vs estimated count density in the FOV were generated and fit to second-order polynomials. Interpreting the slope of linear segment of the fit as an effective sensitivity measurement, such was measured for each background activity and scanner/reconstruction combination. Using constant SN2 between scanners as an iso-quality metric, one can thus obtain either equivalent scan times, injected doses or a combination of each to achieve images of the same SN. Results Among the major vendors and reconstruction algorithms, differences were observed in SN with the range of effective sensitivity varying from .4e-3% to 7.1e-3% and from 1.2e-3% to 22.7e-3% for FBP and ITER algorithms respectively. Tables of equivalent scan times and/or doses which generate equivalent SN between scanners will be presented. Conclusions Estimating equivalent acquisition times or injected dose for different scanners and reconstruction methods is difficult due to non-linearities and varying noise propagation characteristics. The proposed method allows one to specifying the bed position times and injected activity required to obtain equivalent SN among different PET/CT scanners thus a way to circumvent these difficulties. Research Support Funded by NCI Contract No. HHSN261200800001E ER -