Cardiovascular magnetic resonance late gadolinium enhancement risk score for mortality in hypertrophic cardiomyopathy: the HCM-LGE risk score

Abstract Background Prognostic stratification is the cornerstone of the management of patients with hypertrophic cardiomyopathy (HCM).(1) Our working group has developed the concept of "late gadolinium enhancement (LGE) granularity" using cardiovascular magnetic resonance (CMR) with an incremental prognostic value in HCM patients.(2) Purpose To develop a readily interpretable score based on the CMR-LGE granularity concept to predict all-cause mortality in HCM patients. Methods Between 2008 and 2021, all patients referred for HCM assessment using CMR, without history of cardiovascular disease or clinical history of myocarditis were prospectively recruited in two French centers. We selected only patients with CMR-LGE for the purpose of this analysis. The outcome was all-cause death using the French National Registry of Death. Using the variables of the "LGE granularity" model (LGE extent, septal location and subepicardial associated midwall pattern), the HCM-LGE score, was derived from coefficients of the Cox regression (Figure 1). The first center (N=723) was split into training set (N=586) designated for score development, and testing set (N=137) to test the performance of the score, as well as the second center (N=139) was used as the external validation cohort. Categories for the prognostic score were identified using a survival conditional inference tree analysis on the training set, aiming to maximize the log-rank score. Performance evaluation of the score was conducted using Kaplan-Meier curves analysis and Cox regression on the overall population. Results Overall, 862 patients (52±7 years, 54% males) with HCM and CMR-LGE were included. After a median (IQR) follow-up of 9 (7–11) years, 283 (33%) patients died. The proportion of mortality rate for each score points is presented in Figure 2A. In the overall population (N=862), intermediate and high-risk categories were strongly associated with all-cause mortality (hazard ratio (HR) 8.61, 95% CI: 5.96-12.45, p