Issues of discontinuity in the impulse residue function for deconvolution analysis of dynamic contrast-enhanced MRI data

Abstract

Recent dynamic contrast-enhanced MRI studies using the adiabatic tissue homogeneity model have highlighted potential issues of difficulty in convergence during data-fitting and reduced parameter precision, due to discontinuities in the adiabatic tissue homogeneity model. This study presents two solutions (an analytic approach and a discrete correction method) to such convergence problems and show that these problems can be attributed to an inaccurate approximation of the convolution integral based on the standard trapezoidal quadrature. It is further explained that such issues of discontinuity in the impulse residue function do not pertain only to the adiabatic tissue homogeneity model, but are generic to all tracer kinetic models, if the difference in bolus arrival time between the arterial input and tissue voxels were to be accounted for simultaneously during model-fitting.