A method is presented for the acquisition and analysis of dynamic contrast-enhanced (DCE) MRI data, focused on the characterization of tumors in humans. Gadolinium (Gd) contrast was administered by bolus injection, and its effect was monitored in time by fast T1-weighted MRI. A simple algorithm was developed for automatic extraction of the arterial input function (AIF) from the DCE-MRI data. This AIF was used in the pixelwise pharmacokinetic determination of physiological vascular parameters in normal and tumor tissue. Maps were reconstructed to show the spatial distribution of parameter values. To test the reproducibility of the method 11 patients with different types of tumors were measured twice, and the rate of contrast agent uptake in the tumor was calculated. The results show that normalizing the DCE-MRI data using individual coregistered AIFs, instead of one common AIF for all patients, substantially reduces the variation between successive measurements. It is concluded that the proposed method enables the reproducible assessment of contrast agent uptake rates.
Copyright 2001 Wiley-Liss, Inc.