A new graphic plot analysis for cerebral blood flow and partition coefficient with iodine-123-iodoamphetamine and dynamic SPECT validation studies using oxygen-15-water and PET

Abstract

To estimate regional cerebral blood flow (rCBF) and brain-blood partition coefficient (lambda) using a dynamic measurement, a new graphic plot analysis is proposed. By assuming a two-compartment model for tracer kinetics, we derived the linear relationship as Y(t) = K1 - k2 X(t), where Y(t) is the ratio of brain tissue activity-to-time-integrated arterial blood activity and X(t) is the ratio of time-integrated brain tissue activity-to-time-integrated arterial blood activity. A plot of Y(t) against X(t) yields a straight line and the y- and x-intercept of the regression line represent rCBF (K1) and lambda, respectively. The slope is a washout constant (-k2). This method was applied to 14 subjects with N-isopropyl-p-iodine-123 iodoamphetamine ([123I]IMP). The mean values of K1 and lambda for normal subjects were 41.3 +/- 6.7 ml/100 g/min and 29.6 +/- 6.5 ml/g, respectively, in the gray matter. A comparative study with positron emission tomography (PET) using an H2(15)O autoradiographic method revealed good correlation between IMP K1 and PET rCBF [r = 0.822; K1 = 0.842 rCBF + 0.030 (ml/g/min)]. The values of K1 using the graphical method were in excellent agreement with those using a nonlinear least-squares fitting technique (r = 0.992 for K1; r = 0.941 for lambda). The estimated K1 values in the graphical method were not changed when scanning times were varied. We conclude that a two-compartment model is acceptable for IMP kinetics within a scan time of 60 min. The graphical method gives a reliable and rapid estimation of rCBF when applied to dynamic data.