Abstract

A new technique for measuring tissue cellular volume fraction, based on an improved modeling of the dynamic distribution of Gd-DTPA and the effect of proton exchange, is described. This technique uses peak T1 enhancement and blood Gd-DTPA concentration to compute tissue cellular volume fraction. The feasibility of this technique is demonstrated with computer simulations that explore the limits of the simplifying assumptions (small vascular space, slow vascular-extravascular proton exchange), and by direct comparison of MR and radionuclide cell fraction measurements made in muscle, liver, and tumor tissue in a rat model. The computer simulations demonstrate that with slow to intermediate vascular proton exchange and vascular fractions less than 10% the error in our cell fraction measurements typically remains less than 10%. Consistent with this prediction, a direct comparison between MR and radionuclide measurements of cell fraction demonstrates mean percent differences of less than 10%:1.9% in muscle (n = 4); 9% in liver (n = 1) and 9.5% in tumor (n = 4). Similarly, for all rats studied, the MR-measured cell fractions (muscle (0.92 +/- 0.04, n = 20); liver (0.76 +/- 0.11, n = 9); whole tumor (0.69 +/- 0.15, n = 22)) agree with the cell fraction values reported in the literature. In general, the authors' results demonstrate the feasibility of a simple method for measuring tissue cell fraction that is robust across a broad range of vascular volume, flow, and exchange conditions. Consequently, this method may prove to be an important means for evaluating the response of tumors to therapy.