Development of FDG-PET quantitative procedures without blood sampling

Abstract

This study was to compare our input function estimation methods with the standard blood sampling and other competing methods. We have developed two input function estimation methods, namely simplified simultaneous estimation method and regression estimated input function method for FDG-PET model. The derivation of input functions from both methods requires no blood sample. Our results show that both methods have high correlation with the standard blood sampling method using nonlinear least squares method for parameter estimation. There were no significant differences in error percentage among the proposed methods and the existing methods such as image derived input function method, standard input function method, and simultaneous estimation method. Accordingly, our methods can reliably measure cerebral metabolic rate of glucose and provide patients and physicians with simple and noninvasive ways in FDG-PET quantitative studies.