The addition of a cerebral blood volume (CBV) compartment in the [18F]2-fluoro-2-deoxy-D-glucose (FDG) model produces estimates of local CBV simultaneously with glucose metabolic rates when kinetic FDG studies are performed. We investigated the influence of this term upon CMRglc values in a series of brain tumor patients and found that significant overestimations of CMRglc are possible if the effect of CBV upon the model is ignored. The magnitude of this potential overestimation is directly related to the absolute value of CBV locally and inversely related to the CMRglc value. The kinetic estimates also permitted an evaluation of the FDG model in an environment with a variable disruption of the blood-brain barrier. Incorporating the vascular compartment in the FDG model also frequently improved the statistical accuracy of model fits to tissue kinetic data. The sampling requirements for this model configuration were also investigated in a series of computer simulations.