Maximum likelihood estimator image reconstruction for emission tomography in nuclear medicine

Abstract

Emission tomography is a technique which allows the three-dimensional representation of radioisotope labelled chemical compounds that have been inroduced into the human body for the purpose of studying metabolic processes, for the detection of tumors and/or the diagnosis of a variety of diseases. There are two principal methodologies: single photon emission tomography (SPET) and positron emission tomography (PET). Both methods of emission tomography can benefit from a new method of image reconstruction. This method, the maximum likelihood estimator (MLE) method of reconstruction, has the practical advantage that the image noise is smaller in regions of low radioisotope concentration, instead of the rather uniformly distributed noise obtained with conventional filtered backprojection (FBP) reconstruction methods. The author describes the most important mathematical ideas underlying the MLE method and show some results obtained with it, in comparison with FBP results. >