Methods for the correction of vascular artifacts in PET O-15 water brain mapping studies

Abstract

While positron emission tomographic (PET) measurements of regional cerebral blood flow (CBF) can be used to map regions of the brain that are involved in normal and pathological human behaviors, measurements in the anteromedial temporal lobe can be confounded by the combined effects of radiotracer activity in neighboring arteries and partial volume averaging. The authors now describe two alternative methods for addressing this potential confound. One method utilizes the early frames of a dynamic PET study, while the other method utilizes a coregistered magnetic resonance image (MRI) to characterize the vascular region-of-interest (ROI). Both methods subsequently assign a common value to each pixel in the vascular ROI. To simulate the vascular artifact, four dynamically acquired PET, O-15 water scans in the same subject during the same behavioral state were used to compute 4 vascular images (0-60 s after radiotracer administration) which included vascular activity and 4 control images (20-80 s after radiotracer administration) which did not. t-score maps were used to characterize regional blood flow differences related to vascular activity before and after the application of each vascular artifact-correction method. Both methods eliminated the observed differences in vascular activity, as well as the vascular artifact observed in the anteromedial temporal lobes.