Detection of acute reperfusion myocardial hemorrhage with cardiac MR imaging: T2 versus T2

Abstract

Purpose: To evaluate T2 and T2* changes in acute reperfused hemorrhagic and nonhemorrhagic myocardial infarctions and to determine which technique is more suitable in the detection of intramyocardial hemorrhage at 1.5 T.

Materials and methods: Patient studies were approved by the institutional review board and were HIPAA compliant. Patients (n = 14, three women) with first ST-elevation myocardial infarction underwent cardiac magnetic resonance (MR) imaging 3 days after angioplasty. T2* maps, T2 short inversion time inversion-recovery (STIR) images, and late gadolinium enhancement (LGE) images were acquired. Animal studies were approved by the institutional animal care and use committee. Canines (n = 20) were subjected to ischemia-reperfusion injury, and cardiac MR imaging was performed 5 days after reperfusion. T2* and T2 maps and T2 STIR and LGE images were acquired. Repeated-measures analysis of variance or the Friedman test was used to compare T2 and T2* changes in patients with hemorrhagic infarctions and those with nonhemorrhagic infarctions.

Results: Relative to remote myocardium, mean T2* of hemorrhagic infarctions was 54% ± 13 (standard deviation) lower in patients (15.9 msec ± 4.5 vs 35.2 msec ± 2.1, P < .001) and 40% ± 10 lower in canines (23.0 msec ± 4.0 vs 39.3 msec ± 2.5, P < .001). Mean T2* of nonhemorrhagic infarctions was marginally elevated by 6% ± 2.5 (37.8 msec ± 2.5, P = .021) in patients and by 8% ± 5 (44.6 msec ± 4.8, P = .012) in canines. In contrast, mean T2 STIR signal intensity (SI) of both hemorrhagic infarctions and nonhemorrhagic infarctions was higher than that in remote myocardium both in patients (hemorrhagic: 37% ± 19, P < .001; nonhemorrhagic: 78% ± 27, P < .001) and in canines (hemorrhagic: 42% ± 22, P < .001; nonhemorrhagic: 65% ± 22, P < .001). Consistent with STIR SI findings, mean T2 of both hemorrhagic (62.0 msec ± 4.9) and nonhemorrhagic (71.7 msec ± 7.3) infarctions in canines was elevated relative to mean T2 of remote myocardium (52.1 msec ± 4.8) by 18% ± 9 and 38% ± 13, respectively (P < .001 for both).

Conclusion: T2* cardiac MR imaging is more suitable than T2 cardiac MR imaging in the detection and characterization of acute reperfusion myocardial hemorrhage.

Supplemental material: http://radiology.rsna.org/lookup/suppl/doi:10.1148/radiol.13122397/-/DC1.

RSNA, 2013

Figures

Figure 1:

Cardiac MR-based detection of acute hemorrhagic (Hemo+) and nonhemorrhagic (Hemo-) myocardial infarctions in patients. Representative LGE images (inversion recovery–prepared fast low-angle shot sequence; 0.1 mmol of gadopentetate dimeglumine per kilogram of body weight; repetition time/echo time msec/inversion time msec, two R-R intervals/3.32/250; flip angle, 25°), T2 STIR images (two or three R-R intervals/61/170), and T2* maps (multigradient echo; repetition time msec/echo time msec, 15.8/2.6, 4.8, 7.0, 9.3, 11.5, and 13.7; flip angle, 10°) acquired on day 3 after angioplasty in a 49-year-old man with hemorrhagic infarction and a 57-year-old man with nonhemorrhagic infarction. LGE images enabled us to confirm the presence of infarction (arrows) in both patients. Marked T2* decreases were observed within hemorrhagic territories, while hypointense cores were observed within the hyperintense edematous territories on T2 STIR images. Marked hyperintensities were observed within the nonhemorrhagic territories on both T2* maps and T2 STIR images.

Figure 2:

Cardiac MR-based detection of acute hemorrhagic (Hemo+) and nonhemorrhagic (Hemo-) myocardial infarctions in canines. Representative LGE images (inversion recovery–prepared steady-state free precession sequence; 0.1 mmol/kg gadopentetate dimeglumine; 3.75/1.75/230; flip angle, 40°), T2 maps (T2-prepared steady-state free precession sequence; 2.2/1.1; T2 preparation time, 0, 24, and 55 msec; flip angle, 70°), T2 STIR images (2–3 R-R intervals/64/170), and T2* maps (multigradient echo sequence; 21/3.4, 6.4, 9.4, 12.4, 15.4, and 18.4; flip angle, 12°) acquired 5 days after I-R injury are shown in a control animal and animals with hemorrhagic and nonhemorrhagic infarctions. Control animals did not sustain any myocardial infarction, as evidenced by LGE images. Corresponding T2 map, T2 STIR image, and T2* map did not show any distinct signal features generally observed in the presence of an acute myocardial infarction. LGE images enabled us to confirm the presence of myocardial infarctions (arrows) in animals subjected to I-R injury. Marked T2* decreases were observed in hemorrhagic territories, while hypointense cores were observed in hyperintense edematous territories on T2 maps and T2 STIR images. Nonhemorrhagic territories showed marked hyperintensity on T2 maps and T2 STIR images, while marginal T2* increases were observed on T2* maps.

Figure 3:

Histopathologic validation of acute hemorrhagic and nonhemorrhagic myocardial infarctions in canines. Representative histopathologic images obtained from an animal in the control group and two animals in the infarct group (one with hemorrhagic [Hemo+] and one with nonhemorrhagic [Hemo-] infarction) are shown. Triphenyltetrazolium chloride staining enabled us to confirm the presence of infarctions within the infarct group. Corresponding hematoxylin-eosin (H&E) staining of hemorrhagic and nonhemorrhagic territories showed diffuse necrosis with massive infiltration of inflammatory cells (arrows and insets). Hematoxylin-eosin staining of hemorrhagic territories also showed interstitial extravasation of red blood cells into the infarcted territories. Perl staining enabled us to confirm the presence of localized iron in hemorrhagic territories (arrows and inset) but not in nonhemorrhagic territories. Scale bar = 200 µm.