Abstract
^1^H NMR imaging has been used to define zones of myocardial infarction (MI), which appear as areas of relatively increased signal intensity (SI). However, zones of decreased SI have been observed within or around the areas of infarction in NMR images acquired at high magnetic fields. To determine the cause of these areas of reduced SI, ex vivo spin‐echo ^1^H NMR imaging at 1.5 T was performed in eight dogs following 72 h of coronary artery occlusion. In all dogs, a zone of increased SI (122 ± 7% compared to control myocardium; P < 0.01) was observed in the territory of the occluded coronary artery. In seven of the dogs, additional zones were also seen, within or around the central zone of increased SI, which displayed SI that was reduced in comparison with the local enhanced intensity, but was similar to the intensity of normal myocardium (97± 7% compared to control; P = NS). Gross inspection and histological assessment of sliced myocardium disclosed hemorrhage in these regions characterized by locally decreased NMR SI. Image‐derived calculation of T~2~ in the various infarct regions revealed a significant shortening of T~2~, in the hemorrhagic infarct zones characterized by decreased SI, in comparison with the nonhe‐morrhagic infarct zones characterized by increased SI (59 ± 7 ms vs 73 ± 10 ms, P < 0.05). No difference was found, however, between the observed T~2~ of hemorrhagic infarct and of control tissue (57 ± 4 ms). Using a biexponential analysis of T~2~ from the hemorrhagic infarct zones, the intrinsic T~2~ of water protons affected by hemorrhage was determined to be 43 ± 9 ms, significantly reduced in comparison with the values obtained with the standard monoexponential fit. The reduction in T~2~ in the hemorrhagic zone is consistent with the paramagnetic effects of deoxyhemoglobin associated with intramyocardial hemorrhage. Thus the apparent T~2~, measured in hemorrhagic infarct tissue, represents the result of an averaging effect of infarct and hemorrhage on T~2~ relaxation times. These observations improve our understanding of the changes in NMR SI within the infarcted regions, and may provide a noninvasive method for the detection and quantitative assessment of intramyocardial hemorrhage. © 1992 Academic Press, Inc.