A method to correct for scatter, spillover, and partial volume effects in region of interest analysis in PET

Scatter and spatial resolution effects degrade the accuracy of radioactivity concentration estimates obtained from positron emission tomography (PET) data. We present and evaluate a methodology for region quantification which accounts for these degradations. The method is based on analysis of sinogram data and does not require dynamic data sequences to be reconstructed. Moreover, estimates of region variance are also produced which may be used to define weights for model analyses that use weighted least squares minimization in order to obtain unbiased parameter estimates. We evaluate the method using both simulation and measured data and find that, with an appropriate model of scatter and spatial resolution effects, it is unbiased and capable of quantifying myocardial concentration with no more than a 5% error in accuracy for myocardium as thin as 10 mm.