Experimental estimates of the constants relating signal change to contrast concentration for cerebral blood volume by T2*MRI

Estimates of cerebral blood volume (CBV) obtained from dynamic contrast T2*-weighted magnetic resonance imaging (MRI) tend to be significantly higher than values obtained by other methods. This may relate to the common assumption that the proportionality constants relating signal change to contrast concentration are equal in tissue and artery. To test this hypothesis and provide estimates for the ratio of those proportionality constants, the authors compared measurements of CBV by both MRI and computed tomography (CT) scans in nine healthy volunteers obtained using identical kinetic paradigms for the two imaging modalities. Both boluses and infusions of contrast were studied. Measurements were made in nine anatomic regions of interest of the cerebral hemispheres bilaterally. Cerebral blood volume values obtained by CT were generally lower than those obtained by MRI, especially in the cerebral cortex. As a result, the calculated values of the ratios of proportionality constants relating signal change to concentration in tissue and artery after bolus injections were significantly less than 1 in cortex (0.69) and white matter (0.76), although not in deep gray matter structures (0.87). Values of the ratios based on infusion measurements were closer to 1. In addition, CBV measurement errors with bolus MRI were significantly larger than those observed with infusion MRI or by CT. The reasons that the constants differ from 1 and for the larger variance of bolus MRI are discussed in terms of the T2* signal mechanisms. These studies help define the magnitude by which CBV is overestimated with typical T2*-weighted perfusion imaging. Infusion measurements of CBV can reduce the variance intrinsic to T2* MRI and lessen the likelihood of type II error.