Renal kinetic parameter estimation from fast activity curves sampling with a three-headed SPECT system

Recently, we developed a technique for the accurate determination of time-activity curves (TACs) based on the direct reconstruction of region-of-interest (ROI) contents from a series of three-projections data sets acquired with a three-headed SPECT system. This technique allows to reach a very fast sampling of the TACs and has been applied in renal studies for the determination of the blood pool input function and of the renal TACs (sampling interval: 0.67 sec). A simple data analysis method has been developed for the estimation of the renal clearance values directly from the comparison of these curves. The approach is in principle analogous to the conventional one of fitting the tracer kinetic model to the TACs generated from the ROIs segmented on a temporal series of images, but, the new fast sampling procedure is not affected by the problems of inconsistency of data and unbiased estimates can be retrieved. The method has been compared with the Direct Least Squares Estimation (DLSE) one and it has shown many advantages both under the computational and the theoretical aspects. The number of equations is strongly reduced and iterative estimation can be avoided; the number of unknowns is reduced too, since the method does not require the description, often not realistic at all, of the tracer behavior in regions different from kidneys and blood. The method has been compared to the DLSE one by means of simulations and showed, in general, better performances.