Myocardial fibrosis, represented by the accumulation of collagen within the extracellular myocardial space, results in elevated ventricular stiffness and is closely related to clinically evident systolic and diastolic dysfunction^1,2. Consequently, extracellular myocardial space expansion may represent a key intermediate phenotype that precedes morbidity and mortality. Therefore, a non-invasive method that can reliably quantify fibrosis would be desirable. The conventional late gadolinium enhancement (LGE) method relies on the difference in signal intensity between fibrotic and unaffected myocardium as a “nulled” reference, limits its usefulness in diagnosis of the diffuse interstitial fibrosis frequently observed in histological specimens of dilated cardiomyopathy (DCM).³ Recently, myocardial T1 mapping has been applied to quantify the extracellular volume fraction (ECV) and has shown potential for a better characterization of myocardial tissue composition.⁴In this study, we determined whether a T1 map-derived ECV could be associated with disease severity and major adverse cardiovascular events (MACEs).Fifty-two patients with DCM underwent gadolinium-enhanced cardiac MRI with a 3T clinical machine (MAGNETOM Verio, Siemens AG Healthcare Sector, Erlangen, Germany). T1mapping using the Fast MOLLI method was performed for short-axial slices at the basal, mid-ventricular, and apical levels at pre-contrast and post-contrast 21 minutes. Fast-MOLLI was implemented as a two inversion recovery (IR) sequence with the first of 3 and the second of 5 consecutive image acquisitions, which decreased the acquisition time by about one-third when compared with the original MOLLI method.⁵ The SSFP (True-FISP) sequence was used for readout (single-slice, single-shot, TE/TR=1.1/2.5msec, flip angle 35°, FOV 320×223mm, matrix 192×256, and slice thickness 8mm). A T1 map was reconstructed using 8 source images with different inversion times. The ECV was quantified according to the following formula : ECV=λ × (1 – hematocrit), where λ=⊿R1myocardium /⊿R1bloodpool (⊿R1 : the change of R1(=1/T1) between pre-contrast and post-contrast). The pre-contrast and post-contrast T1 map was used for automatic reconstruction of the ECV map with dedicated software (MapMaker prototype®, Medis, Leiden, Netherlands) (Figure 1). The ECV map was segmented into 16 segments (American Heart Association classification). Segment-based ECV values were measured and the mean of 16 segmental values was calculated. The LGE was also performed with IR True-FISP (fixed inversion time=350msec) after gadolinium administration. The presence or absence of LGE was decided visually on a 10-minute-delayed image. The association with parameters for disease severity [left ventricular endo-diastolic volume index (LVEDVI) and ejection fraction (LVEF) measured by cine MRI, and serum BNP] and MACE was assessed.The mean ECV was negatively correlated with LVEF (r=-0.54, p=0.01) and positively correlated with LVEDVI (r=0.44, p=0.004) and serum BNP (r=0.65, p=0.001). During the follow up period (median 290days), MACEs were observed in 7 patients (1 cardiac death, 1 left ventricular assist device implantation, 4 heart failures, and 1 appropriate implantable cardioverter-defibrillator discharge). Patients with MACEs showed significantly higher ECV compared with patients without MACE (p=0.006) (Figure 2), whereas no significant difference was noted in the presence of LGE (p=0.50) (Figure 3). ROC curve analysis identified ECV>38.8% as the optimal cutoff value for predicting a MACE (area under the curve 0.84, p=0.008). Multivariate Cox regression analysis identified ECV>38.8% as a significant independent predictor when compared with the presence of LGE and other risk factors (LVEF, LVEDVI, BNP) (Table 1). Figure 4 shows a Kaplan-Meier Curve stratified by ECV, indicating that ECV was significantly associated with event-free rates.In this study, we used myocardial T1 mapping to measure the ECV and to quantify interstitial fibrosis that may not be evident using conventional LGE methods. The principal finding of our study is the significant relationship between the T1-map-derived ECV and prognosis. We speculate that our data may support an expanded paradigm of cardiac dysfunction and subsequent risk to include myocyte dysfunction (represented by EF) and myocardial loss (represented by LGE), as well as expansion of ECV, indicating diffuse interstitial fibrosis. Our study population was relatively small, so further investigation is required in larger populations.This study supports an association between T1-map-derived ECV measurements and parameters of disease severity and MACEs. ECV may therefore represent a novel prognostic indicator in patients with DCM.