[(18)F]CPFPX was proposed as a novel ligand for in vivo quantification of cerebral A(1) adenosine receptors (A(1)AR) using positron emission tomography (PET). The present study investigates the applicability of non-invasive non-compartment analyses to quantify the [(18)F]CPFPX total distribution volume (DV(') (t)) without arterial blood sampling. Five healthy volunteers underwent dynamic PET (90 min) with arterial and venous blood sampling after [(18)F]CPFPX bolus injection. Area under curve (AUC) analysis and Logan's graphical analysis (GA) were performed employing venous blood samples as non-invasive analyses. AUC analysis without data extrapolation (AUC(r)) and venous GA were also applied on voxel-level to generate parametric images. A conventional two-tissue compartment model (2TCM) using arterial blood samples served as reference method. Regional DV(') (t) estimates provided by venous AUC and GA analyses and 2TCM demonstrated very high agreement (venous AUC vs. 2TCM: r(2) = 0.968, linear regression slope = 0.943; venous GA vs. 2TCM: r(2) = 0.972, slope = 0.906). Although on voxel-level the venous AUC(r) and GA analyses were affected by a slight negative bias, they were still very highly correlated with the 2TCM (voxel-wise venous AUC(r) vs. 2TCM: r(2) = 0.969, slope = 0.858; voxel-wise venous GA vs. 2TCM: r(2) = 0.959, slope = 0.867). The proposed non-invasive analysis methods (particularly venous GA) allow reliable quantification of human [(18)F]CPFPX PET studies. In populations with altered metabolic/kinetic properties, the applicability of venous sampling has to be separately verified. The applicability of [(18)F]CPFPX PET in basic and clinical neurosciences will be considerably enhanced by the avoidance of arterial blood sampling.
(c) 2005 Wiley-Liss, Inc.