BackgroundFor two decades, bright-blood late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) has been considered the reference standard for the non-invasive assessment of myocardial viability. While bright-blood LGE can clearly distinguish areas of myocardial infarction from viable myocardium, it often suffers from poor scar-to-blood contrast, making subendocardial scar difficult to detect. Recently, we proposed a novel dark-blood LGE approach that increases scar-to-blood contrast and thereby improves subendocardial scar conspicuity. In the present study we sought to assess the clinical value of this novel approach in a large patient cohort with various non-congenital ischemic and non-ischemic cardiomyopathies on both 1.5T and 3T CMR scanners of different vendors.MethodsThree hundred consecutive patients referred for clinical CMR were randomly assigned to a 1.5T or 3T scanner. An entire short-axis stack and multiple long-axis views were acquired using conventional phase sensitive inversion recovery (PSIR) LGE with TI set to null myocardium (bright-blood) and proposed PSIR LGE with TI set to null blood (dark-blood), in a randomized order. The bright-blood LGE and dark-blood LGE images were separated, anonymized, and interpreted in a random order at different time points by one of five independent observers. Each case was analyzed for the type of scar, per-segment transmurality, papillary muscle enhancement, overall image quality, observer confidence, and presence of right ventricular scar and intraventricular thrombus.ResultsDark-blood LGE detected significantly more cases with ischemic scar compared to conventional bright-blood LGE (97 vs 89, p=0.008), on both 1.5T and 3T, and led to a significantly increased total scar burden (3.32.4 vs 3.0 +/- 2.3 standard AHA segments, p=0.015). Overall image quality significantly improved using dark-blood LGE compared to bright-blood LGE (81.3% vs 74.0% of all segments were of highest diagnostic quality, p=0.006). Furthermore, dark-blood LGE led to significantly higher observer confidence (confident in 84.2% vs 78.4%, p=0.033).Conclusions The improved detection of ischemic scar makes the proposed dark-blood LGE method a valuable diagnostic tool in the non-invasive assessment of myocardial scar. The applicability in routine clinical practice is further strengthened, as the present approach, in contrast to other recently proposed dark- and black-blood LGE techniques, is readily available without the need for scanner adjustments, extensive optimizations, or additional training.