Arterial blood concentration curves by cardiac PET without arterial sampling or image reconstruction

Summary form only given. In cardiac positron-emission tomography (PET), arterial sampling may be replaced by left-ventricle cavity measurement during imaging. This in turn requires a very large number of images to be reconstructed, if a high-temporal-resolution blood curve is desired. The authors have investigated replacing arterial sampling with LV cavity measurements during bolus N-13 ammonia and H20-15 injections. To make the technique practical, they investigated the algorithm of Huesman (1984), which allows time-activity curves to be created from list-mode or projection data without the necessity of image reconstruction (saving as much as a factor of 30 in processing time). This rapid LV cavity method was tested on three dogs and on a dynamic cardiac phantom. In the dogs, agreement between LV-cavity and arterial samples was good (r>0.92) when gated and smaller ungated volumetric regions of interest were used. Significant size effects were observed with NH3. Spillover between adjacent slices was significant. In the phantom, correlation between LV cavity PET activity and `blood´ samples was high (r>0.98) after appropriate corrections