Intravenously administered O-15 water is widely used to measure regional cerebral blood flow (rCBF) with PET, especially to perform repeat rCBF measurements in the same scan session, e.g., for brain mapping studies. The time interval between measurements is typically 12 mins., to allow for decay of the radioactivity background (half-life of O-15 is 2 mins.). We are developing a method to perform sequential measurements more rapidly, by including a correction for residual background in the tracer kinetic equation used to calculate rCBF. It is anticipated that this approach to increase the frequency of repeat rCBF measurements will provide greater flexibility in PET experiments such as brain mapping and studies of acute drug effects. It will also decrease the time required to perform such studies. After a baseline rCBF scan, a new physiologic state is established at time T. Scan data are then collected to record the residual brain radioactivity. A second bolus of O-15 water is injected, and further scan data are collected. By applying the one-compartment model of Kety, one can develop an equation which relates the time course of radioactivity in brain after time T to the rCBF, the background brain radioactivity value at time T, and the arterial time-activity curve. A rapid, lease-squares method is used to fit this equation for rCBF from the measured scan and blood data collected after time T. Simulation studies demonstrated that there is only a 2-3% increase in the noise in estimates of rCBF due to use of this method. The feasibility of the approach has been demonstrated in preliminary studies obtained in 4 subjects scanned at rest and during visual stimulation. For the rapidly repeated scans, the interscan interval was 6 mins. Mean values for % rCBF correction were 17% in white matter, 8% in insular cortex, and 6% in visual cortex. The absolute magnitude of the correction was greater with higher rCBF values, although the % correction was less. Further plans for this project include (1) comparing blood flow measurements obtained with the standard 12 min. interscan interval to those obtained with the new method; (2) performing further simulation studies of the experimental errors of the method; (3) extending the method so that it can be used to estimate relative rCBF without the requirement for an arterial time-activity curve.
