Tri-dimensional automatic segmentation of PET volumes based on measured source-to-background ratios: influence of reconstruction algorithms

Abstract

A method for automatic volume segmentation of functional imaging based on a relationship between source-to-background ratio and the iso-activity level to be used is described. Such method, which has been established with radioactive spheres in a phantom, is independent of the a priori knowledge of the lesion of interest and is valid for small (>2 ml) and/or poorly contrasted (S/B>1.5) lesions. Accuracy varies slightly with the image reconstruction algorithm used.