Clinical evaluation of two dark blood methods of late gadolinium quantification of ischemic scar

Background Late gadolinium enhancement (LGE) imaging was validated for diagnosis and quantification of myocardial infarction (MI). Despite good contrast between scar and normal myocardium, contrast between blood pool and myocardial scar can be limited. Dark blood LGE sequences attempt to overcome this issue. Purpose To evaluate T-1 rho (T-1 rho)-prepared dark blood sequence and compare to blood nulled (BN) phase sensitive inversion recovery (PSIR) and standard myocardium nulled (MN) PSIR for detection and quantification of scar. Study Type Prospective. Population Thirty patients with prior MI. Field Strength/Sequence Patients underwent identical 1.5 T MRI protocols. Following routine LGE imaging, a slice with scar, remote myocardium, and blood pool was selected. PSIR LGE was repeated with inversion time set to MN, to BN, and T-1 rho FIDDLE (flow-independent dark-blood delayed enhancement) in random order. Assessment Three observers. Qualitative assessment of confidence scores in scar detection and degree of transmurality. Quantitative assessment of myocardial scar mass (grams), and contrast-to-noise ratio (CNR) measurements between scar, blood pool, and myocardium. Statistical Tests Repeated-measures analysis of variance (ANOVA) with Bonferroni correction, coefficient of variation, and the Cohen kappa statistic. Results CNRscar-blood was significantly increased for both BN (27.1 +/- 10.4) and T-1 rho (30.2 +/- 15.1) compared with MN (15.3 +/- 8.4 P < 0.001 for both sequences). There was no significant difference in CNRscar-myo between BN (55.9 +/- 17.3) and MN (51.1 +/- 17.8 P = 0.512); both had significantly higher CNRscar-myo compared with the T-1 rho (42.6 +/- 16.9 P = 0.007 and P = 0.014, respectively). No significant difference in scar size between LGE methods: MN (2.28 +/- 1.58 g) BN (2.16 +/- 1.57 g) and T-1 rho (2.29 +/- 2.5 g). Confidence scores were significantly higher for BN (3.87 +/- 0.346) compared with MN (3.1 +/- 0.76 P < 0.001) and T-1 rho (3.20 +/- 0.71 P < 0.001). Data Conclusion PSIR with inversion time (TI) set for blood nulling and the T-1 rho LGE sequence demonstrated significantly higher scar to blood CNR compared with routine MN. PSIR with TI set for blood nulling demonstrated significantly higher reader confidence scores compared with routine MN and T-1 rho LGE, suggesting routine adoption of a BN PSIR approach might be appropriate for LGE imaging. Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019;50:146-152.