Clinical evaluation of automated parametric imaging for FDG brain PET imaging

1412 Objectives: A study was conducted to acquire parametric FDG brain images using automated software on a clinical PET/CT system. Methods: Nine subjects (4 male; 5 female) were scanned on a Vision 600 PET/CT system (Siemens Medical Solutions, Knoxville, USA). A six minute single bed dynamic scan over the heart was acquired as subjects were injected with 296 ± 2 MBq of FDG using a MEDRAD Intego PET infusion system (Bayer HealthCare LLC, Whippany, USA), followed immediately by 15 or 16 continuous bed motion passes from vertex to heart. The total PET scan duration was 90 min. Parametric images of the net uptake rate constant of FDG (Ki) and distribution volume were reconstructed using integrated direct Patlak reconstruction of passes from 40 to 70 min (early) and 60 to 90 min (late). A subject-specific input function (SSIF) was measured from a volume of interest automatically located on the descending aorta. Further a population-based input function (PBIF) derived from the subjects in this study was used. Static uptake images were also generated from 60 to 70 min using the same reconstruction parameters as were used in the Patlak reconstruction. Images were spatially normalized to an FDG brain template with associated atlases of grey and white matter regions. Median values of Ki in each region for each subject were measured and the mean regional Ki across all subjects was calculated.\n Results: The mean Ki for global grey matter with SSIF was 0.0311 ± 0.0025 ml/min/ml and 0.0290 ± 0.0025 ml/min/ml for early and late imaging respectively. These values are similar to those published by Huisman (EJNMMI Research 2012; 2:63.). Use of PBIF increased Ki for global grey matter by 0.1 ± 4.5% and 0.6 ± 6.5% for early and late imaging respectively. Global grey to white matter Ki ratios were 2.31 ± 0.10 and 2.23 ± 0.16 for early and late imaging respectively. Images of Ki were of good quality; images of distribution volume appeared to be more susceptible to artifacts from patient motion than images of Ki.\n Conclusions: This study demonstrated that parametric imaging using integrated acquisition and direct reconstruction software was feasible and provided expected values of Ki. The use of a PBIF simplifies and shortens the study duration as the patient does not need to be imaged during injection.