ABSTRACT
The ultimate goal of medical imaging is to answer a clinically relevant questions about the disease status of patients. As an example, one of the frequent tasks is to determine if a patient who undergoes medical imaging procedures has cancer or not. For this decision problem there are two possible outcomes: cancer present or not. Another common task is to determine the concentration of a tracer in the organ of interest. For this case there is usually an infinite number of outcomes as the concentration is a continuous variable. Another task may be to determine if the concentration of tracer changes in longitudinal studies. The change can be classified as substantial, marginal, or no change in which case three types of outcomes are identified. Since the data acquired by nuclear imaging medical scanner is inherently random, a clear choice of the optimal decision will seldom be available. The goal of decision theory is to provide tools that can be used for making “the best” decision based on available relevant information.
