A modeling-based factor extraction method for determining spatial heterogeneity of Ga-68 EDTA kinetics in brain tumor

The ROI method used in a Ga-68 EDTA PET dynamic study for quantitative determination of brain tumor (blood brain barrier) BBB permeability assumes that the tumor is homogeneous in terms of Ga-68 EDTA kinetics, even though it is known to be highly heterogeneous. It is desirable to examine regions of different kinetics separately. In this study, we have developed an efficient and effective method to separate tissue regions of different Ga-68 EDTA kinetics. The method uses a two-compartment model to extract three principal component factors (vascular component, fast and slow components) from whole-tumor kinetics by model fitting, then each pixel kinetics in the tumor was expressed in terms of these factors by least-square regression to provide factor images. The whole tumor was separated into two regions-one with mainly fast kinetics and one with slow kinetics. The two regions have markedly different uptake and clearance rates. This method has combined the advantage of statistical factor analysis and a modeling approach. The PET-to-MRI image registration program was employed in this study for registering Ga-68 EDTA PET images to MRI T1 weighted images.