A prostate tumor model in rats was used to compare histometric parameters of prostate cancer physiology with those obtained by magnetic resonance imaging (MRI). The study was focused on vascular physiology as reflected by relative blood volume v(b). Histometry and MRI showed a significant increase in mean v(b) in tumor compared to normal prostate tissue (histometry: normal tissue v(b)=0.69+/-0.19%, tumor tissue v(b)=1.10+/-0.31%, P<.001; MRI: normal tissue v(b)=0.67+/-0.23%, tumor tissue v(b)=1.77+/-0.67%, P<.001). The experimental work showed that MRI yielded a 60.9+/-0.76% higher v(b) than histometry in tumors, while no significant difference in v(b) was found between both methods in normal prostate tissue. Water exchange is known to affect signal intensity on contrast-enhanced MRI. This article investigated the influence of water exchange between intravascular and extravascular space to account for the discrepancy in the values of v(b) obtained with a dynamic inversion-prepared gradient echo MRI sequence and histometry in tumor and normal prostate tissue. The expected influence of water exchange on v(b) was modeled by a computer simulation of the MRI signal and compared with experimental results measured with MRI and histometry. The simulation was based on a two-compartment model indicating that v(b) may be overestimated by MRI. The magnitude of overestimation leads from 10% for the slow water exchange regime to 70% for fast water exchange. Since slow water exchange is probably predominant and even if the observed histological differences in tumor tissue are considered, an overestimation of only 15% due to water exchange is predicted by the simulation. Therefore the overestimation of tumor blood volume by MRI of 60.9% compared to histometry seems to be attributable to additional causes besides water exchange.