Measurement of Myocardium at Risk with Cardiovascular MR: Comparison of Techniques for Edema Imaging

Purpose: To determine variability and agreement for detecting myocardial edema with T2-weighted short-tau inversion recovery (STIR), acquisition for cardiac unified T2 edema (ACUT2E), T2 mapping, and early gadolinium enhancement (EGE) after successfully reperfused ST-segment-elevation myocardial infarction (STEMI) and diagnostic accuracy of each sequence to predict infarct-related artery (IRA). Materials and Methods: Local ethics committee approved the study, with patient informed written consent. On day 2 after successful primary angioplasty for STEMI, 53 patients were prospectively enrolled; 40 patients (mean age, 60 years) completed study. Two sets of cardiac magnetic resonance (MR) images were obtained on same day 6 hours apart. Basal, midcavity, and apical sections were obtained with each sequence. Interobserver, intraobserver, and interimage variability (1 minus intraclass correlation coefficient) and agreement (Bland-Altman method) were assessed. Results: Size of myocardial edema significantly differed. Mean size of myocardium at risk was similar between T2-weighted STIR (18.2 g) and T2 mapping (17.3 g) (P =.54). Mean size differed between T2-weighted STIR (18.2 g) and ACUT2E (14.0 g) (P =.01) and between T2-weighted STIR (18.2 g) and EGE (14.2 g) (P =.003). T2 mapping and EGE had best agreement (interobserver bias: T2-weighted STIR, -0.9 [mean difference] +/- 9.6 [standard deviation]; ACUT2E, -2.5 +/- 6.9; T2 mapping, -3.8 +/- 4.7; EGE, -5.3 +/- 5.9; interimage bias: T2-weighted STIR, 1.5 +/- 5.8; ACUT2E, -0.8 +/- 4.9; T2 mapping, 3.1 +/- 4.0; EGE, 1.1 +/- 4.9; intraobserver bias: T2-weighted STIR, 1.4 +/- 5.8; ACUT2E, 0.6 +/- 4.7; T2 mapping, 2.2 +/- 3.1; EGE, 1.7 +/- 2.9). Variability was lowest for T2 mapping (intraobserver, 0.05; interobserver, 0.09; interimage, 0.1) followed by EGE (intraobserver, 0.03; interobserver, 0.14; interimage, 0.14), with improved detection of territory of IRA versus ACUT2E (intraobserver, 0.11; interobserver, 0.22; interimage, 0.12) and T2-weighted STIR (intraobserver, 0.1; interobserver, 0.32; interimage, 0.1). Conclusion: Cardiac MR methods to detect and quantify infarct myocardial edema are not interchangeable; T2 mapping is the most reproducible method, followed by EGE, ACUT2E, and T2-weighted STIR. (C) RSNA, 2014