Reexamining the quantification of perfusion MRI data in the presence of bolus dispersion

PURPOSE: To determine the true impact of dispersion upon cerebral blood flow (CBF) quantification by removing an algorithm implementation-induced systematic error. MATERIALS AND METHODS: The impact of dispersion on the arterial input function (AIF) between measurement and entry into the tissue of interest on CBF estimates was simulated assuming: 1) contralateral circulation flow that introduces a true arterial tissue delay (ATD)-related dispersive component; and 2) the presence of an arterial stenosis that disperses and shifts the AIF peak entering the tissue; increasing the apparent ATD relative to the original AIF. RESULTS: Previously reported CBF estimates for the stenosis dispersion model were found to be a mixture of true dispersive effects and an algorithm implementation-induced systematic error. The true CBF(MEASURED)/CBF(NO-DISPERSION) ratios for short mean transit times (MTT) (normal) and long MTT (infarcted) tissue were similar for both dispersion models evaluated; this was an unanticipated result. The CBF quantification inaccuracies induced through the dispersion model truly related to ATD were lower than for the local stenosis-based dispersion for small ATD values. CONCLUSION: Correcting the systematic error present in a previous deconvolution study removes the reported ATD-related impact on CBF quantification. The impact of dispersion was smaller than half that reported in previous simulation studies.