Patlak plots of Gd-DTPA MRI data yield blood-brain transfer constants concordant with those of 14C-sucrose in areas of blood-brain opening.

The blood-to-brain transfer rate constant (K(i)) of Gd-DTPA was determined in MRI studies of a rat model of transient cerebral ischemia. The longitudinal relaxation rate, R(1), was estimated using repeated Look-Locker measurements. A model-independent analysis of deltaR(1), the Patlak plot, produced maps of K(i) for Gd-DTPA and the distribution volume of the mobile protons (V(p)) with intravascular-Gd changed R(1)'s. The K(i)'s of Gd-DTPA were estimated in regions of interest with blood-brain barrier (BBB) opening (regions of interest, ROIs) and compared to those of (14)C-sucrose determined shortly thereafter by quantitative autoradiography. The K(i)'s for both Gd-DTPA and sucrose were much higher than normal within the ROIs (n = 7); linear regression of K(i) for Gd-DTPA vs. K(i) for sucrose yielded a slope of 0.43 +/- 0.11 and r(2) = 0.72 (P = 0.01). Thus, K(i) for Gd-DTPA varied in parallel with, but was less than, K(i) for sucrose. In the ROIs, mean V(p) was 0.071 ml g(-1) and much higher than mean vascular volume estimated by dynamic-contrast-enhancement (0.013 ml g(-1)) or mean V(p) in contralateral brain (0.015 ml g(-1)). This elevated V(p) may reflect increased capillary permeability to water. In conclusion, K(i) can be reliably calculated from Gd-DTPA-MRI data by Patlak plots.