The purpose of this study was to characterize the contrast caused by a susceptibility MRI contrast agents, on spin echo T-2-weighted imaging of reperfused myocardial infarction. Our interest in this model focused on the expected requirement that such agents be compartmentalized in the tissue to cause signal loss on spin echo images, a condition which may not be present in reperfused infarcted myocardium. Accordingly, nine rats were subjected to 2 h of left coronary artery occlusion followed by 3 +/- 0.5 h of reperfusion prior to administration of contrast media. Three sets of MR images were acquired: (a) baseline axial images at the midventricle, both T-1-weighted (TR/TE = 300/20) and T-2-weighted (TR/TE = 1500/60); (b) T-1-weighted images after administering a T-1-enhancing agent, Gd-DTPA-BMA (0.2 mmol/kg), to document that contrast media is delivered to the reperfused infarction; and (c) T-2-weighted images after administering the susceptibility agent, Dy-DTPA-BMA (1.0 mmol/kg). Gadolinium-enhanced T-1 images depicted reperfused infarction as regions with greatly enhanced signal intensity compared with uninfarcted myocardium, indicating that contrast agent was delivered to the infarcted zone. Dysprosium-enhanced T-2 images depicted the injury as a region of persistent signal intensity relative to depletion of signal in normal myocardium, consistent with failure of the contrast agent to cause signal loss. Similar infarction sizes were observed for unenhanced T-2- weighted images (33 +/- 5%), gadolinium-enhanced T-1-weighted images (36 +/- 5%) and postmortem staining (30 +/- 6%); strong correlations (r > 0.9) were noted in comparisons of these data. Dysprosium-enhanced images exhibited a smaller region of differential signal presumed to be infarction (20 +/- 5%, P < 0.05) and weak correlations (r < 0.75) with the other measurements. We conclude that the smaller infarction depicted on dysprosium-enhanced images is a subregion of the true infarction in which myocardial necrosis is sufficiently advanced that the agent is homogeneously distributed throughout all tissue compartments, preventing T-2*-dependent phase loss on spin echo images.