This study assessed several proposed imaging strategies and analytic methods based on gadopentetate-enhanced MRI to differentiate benign from malignant breast tumors in a blinded experimental animal study. Steady-state dynamic MRI and first-pass imaging, performed with either T(1)- or T*(2)- weighted sequences, were compared. Semiquantitative and quantitative analysis methods, based on empirical measures of the data or physiological models, were subsequently applied to the imaging datasets. Comparative measures provided pathologic distinction of benign from malignant tumors, tumor grading, and histologic determination of microvascular density. Of the eight tested methods, only one, an estimate of first-pass perfusion using T *(2)-weighted imaging, showed an almost significant (P = 0.05) difference between benign and malignant tumors and correlated almost significantly (r =.3, P = 0.06) with the tumor grade. All other tests, performed either with steady-state imaging or with T(1)-weighted first-pass imaging, failed to differentiate benign from malignant tumors. In addition, they yielded poor correlations with tumor grade and microvascular density.