| Measurement of cerebral blood flow (CBF) using positron emission tomography (PET) with 15O-labelled water is commonly achieved using the one-compartment model of Kety. Unlike this model, a recently developed two-compartment CBF model accounts for nonextracted intravascular radioactivity and therefore provides with perfusion (CBF) as well as vascular, i.e., blood volume (CBV) related, information. The two parameters obtained from this two-compartment analysis, namely cerebral water clearance, KH2O1 , a measure of perfusion, and the apparent tracer distribution volume, Vo, provide more complete haemodynamic information than CBF alone. Still, accurate K1 values can only be obtained if appropriate corrections for tracer delay (Deltat) and dispersion (tau) have been applied. Building on this two-compartment model, we have derived new nonlinear and linear solutions where all parameters involved appear explicitly. Simulations were performed to evaluate the accuracy and precision with which K1, k2, Vo, Deltat and tau can be determined from multiparameter regression of the model equations on dynamic PET data. It was found that K1 and k2 can be estimated with less than 2% error using 5-parameter fitting.;We have applied our solutions to dynamic 15O-water bolus PET studies where a CBF change was produced either through neuronal activation triggered by visual stimulation (in normal subjects), or by external pharmacological control through vasodilatory insults induced by Diamox injection (in patients with cerebrovascular disease) and antioxidant LY231617 administration (in monkeys with unilateral middle cerebral artery occlusion).;We found that in addition to accurate pixel-by-pixel estimation of K 1 and k2, accurate changes in K1, Vo and the ratio Vo/K1 between baseline and activation states can also be obtained using our solutions of the two-compartment CBF model with implicit corrections for delay, dispersion and residual vascular radioactivity. These parameter changes provide potential indices of cerebrovascular reserve capacity, and of variations in cerebral vascular volume and mean vascular transit time. Our method may prove to be an elegant tool in the selection of candidates for revascularisation surgery. |
