An investigation of the effect of finite system resolution and photon noise on the bias and precision of dynamic cardiac SPECT parameters.

The exchange of a radionuclide (and similarly of oxygen) between blood and the myocardium is a dynamic process. Dynamic SPECT offers the possibility of directly quantifying the kinetic parameters which describe this process. In this paper we investigate and quantify the effect of typical SPECT system resolution and photon counting statistics on the bias and precision of dynamic cardiac SPECT parameters. System resolution and photon noise are only two of the image degrading processes which occur in SPECT. Therefore, the bias and precision quoted should be viewed as a lower limit on those which could be expected with an experimental study. Dynamic SPECT projection data are simulated using a realistic human torso phantom. Data are simulated assuming both perfect system resolution and a system resolution typical of a clinical SPECT system. A triple detector SPECT system which acquires a full set of projection data in 10 s using continuous detector motion is modelled. Kinetic parameters are estimated using a number of myocardial regions of interest. The results show that the rate constant characterizing the washing of activity into the myocardium is more sensitive to region of interest position than is the washout rate constant. The bias and precision of the dynamic parameters are estimated from multiple realizations of projection data exhibiting various noise levels. The main effect of increased photon noise in the projection data is to decrease the precision of the estimated parameters. However, there is also some evidence that for high noise levels the bias of the parameters may also be affected.