Improving prognostic accuracy in ischemic cardiomyopathy: the “CMR-LGE score”

Abstract

Introduction

Several studies have shown the impact of cardiovascular magnetic resonance (CMR) imaging findings notably late gadolinium enhancement (LGE) to stratify the risk in ischemic cardiomyopathy (ICM) but an easy interpretable score to guide clinical practice is lacking.

Objective

To determine the CMR parameters most predictive of mortality in a cohort of ICM patients with reduced left ventricular ejection fraction (LVEF) less than 50%. Then, we will develop a readily interpretable score based on these CMR parameters.

Method

Between 2008 and 2022, consecutive patients with ICM, LVEF < 50% and presence of ischemic-LGE on CMR, were recruited by two independent centers. The primary outcome was all-cause death. The first center (i.e. the derivation cohort, ICPS Hospital, n = 2900) was used for variable selection and score development, and the second center (i.e. the validation cohort, Lariboisiere Hospital, n = 691) was used to evaluate the performance of the score. Clinical variables (17) and CMR variables (15) were initially assessed. Feature selection was performed using Random Survival Forest (RSF). Using the selected variables, our CMR-LGE score was derived from beta coefficients of the Cox regression. Performance evaluation was conducted using Harrel's C index compared with known prognostic factors associated with poor outcomes in ICM. Prognostic score categories were defined using survival tree analysis in the derivation cohort.

Results

Among the 3591 patients included (mean age 65 years; 75% male; mean LVEF 44%), 549 (15%) died over a median follow-up of 9 years. Feature selection with RSF highlighted that the most important variables to predict death were LGE variables: the LGE extent, its location (septal and anterior) and the transmurality of ischemic-LGE as well as the extent of additional midwall-LGE. Following these findings, we developed the CMR-LGE score using Cox regression coefficients (Fig. 1A). Harrel's C index of CMR-LGE score outperformed known prognostic factors, including LVEF (0.88 vs 0.69). Finally, based on our CMR-LGE score, we identified a low-risk population (score ≤ 5) and a high-risk population (score ≥ 9), validated using survival curves in the validation cohort (Fig. 1B).

Conclusion

Using RSF, we identified that the 5 most important CMR variables to predict mortality in ICM were all LGE features. Our CMR-LGE score showed excellent performance to stratify patient risk compared to traditional prognostic factors including LVEF.