Direct parametric reconstruction from dynamic projection data in emission tomography including prior estimation of the blood volume component.

Dynamic emission tomography measures the amount of the injected radiotracer through multiple consecutive time frames. The temporal information is used to estimate the pharmacokinetic properties of the radio-labeled molecules. One complicating factor is that the scanner measures the signal that is generated not only from the molecules in the tissues of interest, but also from those in the blood. This is usually corrected for by subtraction of the blood term from the radioactivity image before tracer kinetic modeling. However, some methods estimate the pharmacokinetic parameters directly from the measured data (e.g. sinograms) and in such cases the subtraction procedure for correcting the data cannot be applied. In other advanced direct reconstruction approaches, the blood radioactivity curve is included in the model. However, there may be models for tracers, which are not feasible as the blood volume component can be undistinguishable. In this study, we address these issues and show that these can be solved by viewing the blood radioactivity as a 'physiological' background term similar to the physical background (e.g. scatter and random counts). The proposed technique can be applied to any radiotracers in emission tomography, independent of the kinetic model.