Parameter estimation in clinical nuclear medicine studies

The advent of sophisticated computerized data acquisition systems in nuclear medicine has provided a voluminous amount of dynamic scintigraphic data. The problem at hand is the interpretation of these data into meaningful diagnostic information. It is pointed out that there are a number of difficulties one normally encounters in this endeavor and how we and others have chosen to cope with these problems is indicated. We have suggested: the use of multiple views of a radionuclide distribution for improved accuracy of quantitation, the use of simulation modelling analysis for complicated systems, the arbitrary selection of symmetric regions of interest when no a priori reason to select a particular region is available, and the use of potentially powerful statistical tests such as the analysis of variance or factor analysis to detect anomalous regions when on adequate model is available