Analysis of kinetic parameter estimates for dynamic PET

Abstract

In this work, a mathematical framework is developed for analytical investigation of bias and variance in kinetic parameter estimation of compartment models for dynamic PET imaging. As implicit functions are used for kinetic parameter estimation, an open form equation cannot be derived for bias and variance. The results of the developed framework is compared against Monte Carlo simulations. Bias and variance in the kinetic parameter estimations can be computed within 10 % accuracy, if the noise in time activity curves is not very high.