Assessment of postreperfusion myocardial hemorrhage using proton NMR imaging at 1.5 T

Intramyocardial hemorrhage occurs frequently after reperfusion of acute myocardial infarction. However, its significance has not yet been established, mainly because of the lack of methods for detecting such hemorrhage. The following ex vivo study was carried out to assess the potential of nuclear magnetic resonance (NMR) imaging to detect and quantitate postreperfusion intramyocardial hemorrhage.

Sixteen adult mongrel dogs underwent 3 hours of coronary occlusion followed by 1 hour of reperfusion, and three dogs underwent 4 hours of occlusion without reperfusion. Radiolabeled microspheres and 51Cr-labeled red blood cells were used to assess flow and evaluate the extent of hemorrhage. These results were later compared with both NMR and histology. Spin-echo NMR imaging was performed on the excised hearts using a 1.5-T system. Macroscopic assessment of the sliced myocardium revealed the existence of hemorrhage in 14 of the 16 dogs that underwent reperfusion but in none of those with occlusion only. In all 16 dogs with reperfusion, zones of increased signal intensity (SI) ratio (1.68 +/- 0.41 compared with control, p less than 0.05) were seen in regions relating to the distribution of the occluded coronary artery, whereas in 13 of the 16 dogs, areas of decreased SI within the zone of increased SI ratio (0.81 +/- 0.16 compared with control, p less than 0.05) were also seen, corresponding to regions with macroscopic hemorrhage. In contrast, in the three dogs without reperfusion, no macroscopic hemorrhage was observed, and likewise, no NMR zones of reduced SI were detected. Hemorrhage size by NMR (decreased SI zones), correlated well with hemorrhage size calculated from tissue slices (r = 0.96, SEE = 0.92%, p less than 0.01), or by 51Cr labeling (r = 0.78, SEE = 1.5, p = 0.1). In the reperfusion group, T2 relaxation times in the infarcted hemorrhagic zone (58 +/- 9 msec) were significantly lower than the infarcted zones without hemorrhage (98 +/- 13 msec, p less than 0.001). In contrast, when compared with control (964 +/- 72 msec), T1 relaxation times were significantly increased in both infarct zones, either with (1,284 +/- 176 msec) or without (1,266 +/- 103 msec) hemorrhage. The selective shortening of T2 relaxation times in the hemorrhagic regions is consistent with the paramagnetic effects of deoxyhemoglobin.

NMR imaging may provide a noninvasive approach for the detection and quantitation of intramyocardial hemorrhage. This observation may provide a means to further characterize pathological processes associated with acute myocardial infarction and assess the role of myocardial hemorrhage after reperfusion therapy.