Quantification of myocardial blood flow and blood flow reserve in the presence of arterial dispersion: A simulation study

Myocardial blood flow (MBF) can be quantified using dynamic T-1-weighted MRI of diffusible tracers and a mathematical model of underlying vasculature. Quantification of MBF by means of T-1-weighted MRI requires knowledge of the arterial input function (AIF). The AIF can be estimated from the left ventricular (LV) cavity. However, dispersion may occur between the LV and the tissue of interest because of the laminar blood flow profiles, branching of venules, and because of stenosis. To evaluate the influence of dispersion on the results of MBF quantification, a simulation study was performed. The dispersion was described as a convolution of the AIF with an exponential residue function. Synthetic tissue and AIF curves were analyzed and the derived parameters fit to the simulated parameters. The results show that an unaccounted dispersion may result in a systematic underestimation of MBF up to approximately 50%. Underestimation increases with increasing dispersion and with increasing MBF. Assuming equal dispersion at rest and during hyperemia, myocardial perfuslon reserve (MPR) estimates are also susceptible to underestimation of approximately 20%. An unaccounted dispersion therefore can lead to systematic underestimation of both blood flow and perfusion reserve. (C) 2002 Wiley-Liss, Inc.