The purpose of this project is to determine an optimal method for performing functional brain mapping studies with three-dimensional positron emission tomography (PET) and O-15 water, a tracer that maps regional cerebral blood flow (rCBF). Current tomographs permit three- dimensional imaging, substantially increasing sensitivity. However, image noise increases at the high count rates at which rCBF studies with bolus injections of O-15 water have typically been performed. Although the image signal-to-noise ratio (S/N) is improved at low count rates with small amounts of injected radioactivity, the improvement falls off at high count rates. Thus, several factors are important in designing brain mapping experiments: the amount of radioactivity per injection, the number of injections, and type of injection (fast bolus vs. slow infusion). Simulations and experiments in humans are being performed to find an optimal combination of these parameters for the GE Advance tomograph. In human studies, scan data were obtained with bolus intravenous injections of 2.5 to 30 mCi of 0-15 water in 4 subjects performing a word generation task. The noise equivalent count (NEC) rate for whole brain peaked at an injected dose of 12-15 mCi; 10 mCi gave an NEC of 92% of the peak value. Although consistent focal rCBF changes were detected with scan durations of 60, 90, and 120 sec, their statistical significance decreased as scan time increased. Therefore, we selected a 1 min scan using 10 mCi for bolus injections. We then performed simulations to evaluate the effect on S/N of longer scan durations with slow tracer infusions. Data on image noise as a function of local radioactivity were obtained by scanning a brain phantom. Then, simulations were used to predict the tissue radioactivity that would be observed with a slow tracer infusion. Combining the measured phantom noise data with the predicted tissue data allowed calculation of the S/N. This showed an equivalent S/N for a 60 sec scan with a bolus injection and a single longer scan with a slow infusion, given the same amount of injected radioactivity. The time over which tracer uptake contributes to the total integrated counts is greater with a slow infusion than a bolus injection. Thus, investigators may wish to customize scanning parameters for different types of brain mapping experiments while maintaining the S/N of bolus injections.
