Assessment of performance for the dynamic SPECT (dSPECT) method

The four-dimensional images reconstructed using the dSPECT method may be used to quantify temporal changes in regional radiotracer concentrations within a patient which are believed to reflect the functional ability of the investigated organ. The accuracy of the diagnosis will certainly depend on the accuracy of the dynamic reconstruction. Therefore, before the method can be used clinically it is important to evaluate its performance and limitations. The accuracy of the dSPECT reconstructions may be assessed using several different approaches and here the authors present a summary of their research in this area and discuss different validation techniques. Analysis of simulation data created using 6 different acquisition protocols with three camera configurations demonstrated that the errors in reconstructed dynamic activity distributions range from about 50% for single head cameras to just a few percent for triple head systems. The accuracy of kinetic parameter evaluation depends strongly on the modelled situation, but in general the shapes of the reconstructed time activity curves match closely the shapes of the true curves. Quantitative analysis of the data showed good agreement for washout half-lives in the range 3-8 minutes but larger errors were present for longer T/sub 1/2/ and for multi exponential time-activity functions. These findings were confirmed in phantom experiments, although the results were more difficult to qualify in this case. Numerical values of the renal GFR measurements performed using blood sampling method, dynamic planar and dSPECT scans for normal volunteers are reported.