Comparison of the Kinetic Parameters Estimated with Different Numerical Methods in Dce-mri

Introduction: Dynamic contrast enhanced MRI (DCE-MRI) has been shown its ability to noninvasively assess characterization of tumor by the kinetic parameters, such as volume transfer constant (K) and rate constant (kep). The Tofts and Kermode (TK) model [1] and modification of Tofts and Kermode (mTK) model [2] are both popular pharmacokinetic models. Nonlinear least-squares (NLSQ) method is often used to calculate kinetic parameters in TK and mTK models. However, a disadvantage of NLSQ method is the long calculation time. To overcome this disadvantage, an alternative method, linear least-squares method, have been developed in previous study [3]. The study had demonstrated that the LLSQ method not only can reduce calculation time but also can increase accuracy of estimation of kinetic parameter in lower SNR. However, the previous study only demonstrated that LLSQ is better method than NLSQ in solving mTK model equation. In this study, we further propose to apply the LLSQ method to solve TK model and use computer simulations to compare the accuracy of the kinetic parameters.