AN INTEGRATED APPROACH TO LOCALIZE EPILEPTIC FOCI USING RELATIVE SPECT SUBTRACTION

摘要

Accurate epileptic focus localization using SPECT images has proven itself to be a challenging endeavor. Firstly, this is partly due to the fact that radiopharmaceuticals such as hexamethylpropylene amine oxime (HMPAO) quantitatively underestimate large blood flows; this may lead to SPECT images that do not reflect the true physiological conditions in the brain. This often leads to subtracted SPECT images that contain non-distinct epileptic foci. Therefore, better methods of SPECT image analysis need to be devised that can overcome this quantitative defect. The relative change subtraction method proposed alleviates the burden imposed by this problem. Secondly, the analysis of SPECT images has traditionally been performed by physicians and is often tune-consuming and prone to error. Though software packages exist to aid this process in distinct phases, an integrated and fully automated approach is a more desirable outcome. Interactivity will be provided through a visual interface for medical expert feedback in the decision-making process. To meet such desired outcomes, an integrated algorithm has been designed to automate the process of SPECT image analysis so that minimal human intervention is required.