Synthetic late gadolinium enhancement cardiac magnetic resonance for diagnosing myocardial scar

Abstract Objectives. Late gadolinium enhancement (LGE) is the in vivo reference standard for assessing focal myocardial fibrosis. Post-contrast T1-mapping by Modified Look-Locker Inversion recovery (MOLLI) can be used to generate synthetic late gadolinium enhancement (SynLGE) images with an image contrast similar to conventional LGE images. We hypothesized that SynLGE has an accuracy that approaches conventional LGE for diagnosing focal myocardial fibrosis. Methods. Consecutive patients (n = 109, mean ± SD age 50 ± 16 years, 63% male) referred for clinical cardiac magnetic resonance imaging underwent LGE and post-contrast MOLLI starting 10–15 and 20–25 minutes post contrast, respectively. A cardiac short-axis stack and three long-axis views were acquired for SynLGE and LGE. SynLGE were generated from post-contrast T1-maps. Only LGE and SynLGE images were analyzed by two blinded observers for agreement regarding localization and origin of focal myocardial fibrosis on a per-patient basis. Results. Consensus identified focal fibrosis by LGE in 44/109 (40%) patients. Compared to LGE, SynLGE yielded a diagnostic sensitivity of 34/44 (77%), specificity of 64/65 (98%), positive predictive value of 34/35 (97%), negative predictive value of 64/74 (86%), and an overall accuracy of 98/109 (90%). In cases where SynLGE missed focal fibrosis (n = 10), these were either small non-ischemic focal fibrosis (n = 8) or infarction in a thin myocardial wall (n = 2). In one case, SynLGE identified midmural non-ischemic focal fibrosis not identified by LGE. Discussion. Overall, SynLGE showed good agreement with LGE. SynLGE derived from post-contrast T1-maps may provide the complementary ability to increase confidence in assessment of LGE images for focal myocardial fibrosis.