Direct least-squares estimation of spatiotemporal distributions from dynamic SPECT projections using a spatial segmentation and temporal B-splines

Artifacts can result when reconstructing a dynamic image sequence from inco nsistent, as well as insufficient and truncated, cone beam single photon em ission computed tomography (SPECT) projection data acquired by a slowly rot ating gantry, The artifacts can lead to biases in kinetic model parameters estimated from time-activity curves generated by overlaying volumes of inte rest on the images. However, the biases in time-activity curve estimates an d subsequent kinetic parameter estimates can be reduced significantly by fi rst modeling the spatial and temporal distribution of the radiopharmaceutic al throughout the projected field of view, and then estimating the time-act ivity curves directly from the projections. This approach is potentially us eful for clinical SPECT studies involving slowly rotating gantries, particu larly those using a single-detector system or body contouring orbits with a multidetector system. We have implemented computationally efficient methods for fully four-dimens ional (4-D) direct estimation of spatiotemporal distributions from dynamic SPECT projection data, Temporal B-splines providing various orders of tempo ral continuity, as well as various time samplings, were used to model the t ime-activity curves for segmented blood pool and tissue volumes in simulate d cone beam and parallel beam cardiac data acquisitions. Least-squares esti mates of time-activity curves were obtained quickly using a workstation. Gi ven faithful spatial modeling, accurate curve estimates were obtained using cubic, quadratic, or linear B-splines and a relatively rapid time sampling during initial tracer uptake. From these curves, kinetic parameters were e stimated accurately for noiseless data and with some bias for noisy data, A preliminary study of spatial segmentation errors showed that spatial model mismatch adversely affected quantitative accuracy, but also resulted in st ructured errors (projected model versus raw data) that were easily detected in our simulations. This suggests iterative refinement of the spatial mode l to reduce structured errors as an area of future research.