Transfer of normal Tc-99m-ECD brain SPET databases between different gammacameras

A stereotactic, normal perfusion database is imperative for optimal clinica l brain single-photon emission tomography (SPET). However, interdepartmenta l use of normal data necessitates accurate transferability of these data se ts. The aim of this study was to investigate transfer of three normal perfu sion databases obtained in the same large population of healthy volunteers who underwent sequential scanning using multihead gamma cameras with differ ent resolution. Eighty-nine healthy adults (46 females, 43 males; aged 20-8 1 years) were thoroughly screened by history, biochemistry, physical and fu ll neurological examination, neuropsychological testing and magnetic resona nce imaging. After injection of 925 MBq technetium-99m labelled ethyl cyste inate dimer (ECD) under standard conditions, 101 scans were acquired from a ll subjects (12 repeat studies) on a triple-head Toshiba GCA-9300A (measure d average FWHM 8.1 mm). Ninety-one sequential scans were performed on a dua l-head Elscint Helix camera (FWHM 9.6 mm) and 22 subjects also underwent im aging on a triple-head Prism 3000 (FWHM 9.6 mm). Images were transferred to the same processing platform and reconstructed by filtered back-projection with the same Butterworth filter (order 8, cut-off 0.9 cycles/cm) and unif orm Sorensen attenuation correction (mu =0.09). After automated rigid intra subject registration, all subjects were automatically reoriented to a stere otactic template by a nine-parameter affine transformation. The databases w ere analysed using 35 predefined volumes of interest (VOIs) with normalisat ion on total VOI counts. For comparison, the high-resolution data were smoo thed with a 3D Gaussian kernel to achieve moro similar spatial resolution. Hoffman phantom measurements were conducted on all cameras. Partial volume effects after smoothing varied between -6.5% and 10%, depending on VOI size . Between-camera reproducibility was 2.5% and 2.7% for the Toshiba camera v ersus the Helix and the Prism database, respectively. The highest reduction in between-camera variability was achieved by resolution adjustment in com bination with linear washout correction and a Hoffman phantom-based correct ion. In conclusion, transfer of normal perfusion data between multihead gam ma cameras can be accurately achieved, thereby enabling widespread interdep artmental use, which is likely to have a positive impact on the diagnostic capabilities of clinical brain perfusion SPET.