Estimation of physiological parameters in the subspace of arterial input function in dynamic contrast-enhanced magnetic resonance imaging

Estimation of physiological parameters by fitting data to a proper pharmacokinetic model plays an important role in application of DCE-MRI for characterization of tissue microvasculature in diagnosis and prognosis of various diseases. In this study, by allowing asymmetric permeability a generalized two-compartment exchange model (G2CXM) is presented that uniformly includes the extensively applied Patlak model, Tofts model, and extended Tofts model as well as the two-compartment exchange model as special instances. The identifiable physiological parameters in the G2CXM for various tissue types determined by the special values or boundary values of physiological parameters are obtained by analyzing its impulse response function. To obviate failures occurring at times in the conventional fitting method, an approach of estimating physiological parameters in the subspace of arterial input function (SAIF) is proposed. Simulation result shows that the SAIF can obviate failure, significantly increase accuracy and decrease bias of estimated physiological parameters in a wide signal to noise ratio range.