Harmonisation of PET/CT contrast recovery performance for brain studies

• DOI: 10.1007/s00259-021-05201-w
• Publication Date: 2021-01-31T08:02:25Z
• Authors: E. E. Verwer; S. S. V. Golla; A. Kaalep; M. Lubberink; F. H. P. van Velden; V. Bettinardi; M. Yaqub; T. Sera; S. Rijnsdorp; A. A. Lammertsma; R. Boellaard
• Keywords: Grey matter; Image contrast; Comparability

Abstract Purpose In order to achieve comparability of image quality, harmonisation PET system performance is imperative. this study, prototype criteria for brain studies were developed. Methods Twelve clinical PET/CT systems (4 GE, 4 Philips, Siemens, including SiPM-based “digital” systems) used acquire 30-min scans a Hoffman 3D Brain phantom filled with ~ 33 kBq·mL −1 [ 18 F]FDG. Scan data reconstructed using various reconstruction settings. The images rigidly coregistered template (voxel size 1.17 × 2.00 mm 3 ) onto which several volumes interest (VOIs) defined. Recovery coefficients (RC) and grey matter white ratios (GMWMr) derived eroded (denoted in the text by subscript e) non-eroded (GM) (WM) VOIs as well mid-phantom cold spot (VOI from Hammers atlas. addition, left-right hemisphere differences voxel-by-voxel compared reference assessed. Results Systematic observed reconstructions without point-spread-function modelling (PSF ON PSF OFF , respectively). Normalising image-derived activity, upper lower limits ensuring follows: RC GMe = [0.97–1.01] GMWMr e [3.51–3.91] VOI GM [0.78–0.83] [1.77–2.06] VOI, [0.92–0.99] [3.14–3.68] [0.75–0.81] [1.72–1.95] VOI. Conclusions To inter-scanner comparability, we propose selecting settings based on specified “Results”. These proposed standards should be tested prospectively validate and/or refine criteria.