Reliable Model and Cluster Aided Formation of Parametric Images in Functional Imaging

Parameter estimation in functional imaging provides unique quantitative measures in clinical diagnosis, and in the evaluation of treatment response and new drugs. Voxel-by-voxel parameter estimations can construct parametric images which visualize the spatial distribution of functional parameters. The low signal-to-noise ratio in single photon emission computed tomography (SPECT) may cause physiologically meaningless estimates using the general linear least square method (GLLS). A proof-of-principle framework is proposed in this study for constructing simultaneously multiple parametric images using a model-aided GLLS method and fuzzy clustering for dynamic SPECT. Computer simulations were performed to evaluate the accuracy and reliability of estimates for the studied methods. The results show that the model-aided GLLS with fuzzy clustering did enhance reliability for voxel-byvoxel parameter estimation with a slight overestimation of the volume of distribution. The method employing normalization of TTAC was superior to the method without normalization.