Phenotypic stratification and prognostic value of cardiac magnetic resonance in non-dilated left ventricular cardiomyopathy

Background Non-dilated left ventricular cardiomyopathy (NDLVC), characterised by non-ischaemic scar/fatty replacement or isolated systolic dysfunction without dilatation, lacks validated risk stratification tools. We aimed to define cardiac magnetic resonance (CMR)-based phenotypes and evaluate their association with clinical outcomes. Methods In 515 patients with NDLVC (mean age 45 (16) years), three phenotypes were classified by CMR: late gadolinium enhancement (LGE+)/H− (LGE with preserved left ventricular ejection fraction (LVEF), n=130), LGE−/H+ (hypokinesia without LGE, n=226) and LGE+/H+ (LGE with reduced LVEF, n=159). The primary endpoint was all-cause death/heart transplantation; secondary endpoints included heart failure (HF) events and malignant ventricular arrhythmia (MVA). Results Over a mean follow-up of 6.5 (1.9) years, 29 patients (5.6%) reached the primary endpoint, while 81 (15.7%) and 19 (3.7%) experienced HF and MVA, respectively. The LGE+/H+ subgroup demonstrated the highest risk for composite clinical endpoints compared with other phenotypic groups (p<0.001). Multivariable analysis identified New York Heart Association class >II (HR 3.42, 95% CI 1.58 to 7.39, p=0.002), LVEF (HR 0.91 per 1% increase, 95% CI 0.88 to 0.95, p<0.001) and LGE extent (HR 1.14 per 3% increase, 95% CI 1.07 to 1.21, p<0.001) as independent predictors of the primary endpoint, with excellent discriminative power (C-statistic 0.862). In the adjusted model, LGE extent also independently predicted HF (HR 1.11 per 3%, 95% CI 1.06 to 1.17, p<0.001). The univariable Cox regression analysis indicated LGE extent was significantly associated with MVA (HR 1.12 per 3%, 95% CI 1.02 to 1.23, p=0.021). Conclusion CMR phenotyping enables risk stratification in NDLVC. LGE extent provides an objective marker to identify high-risk patients—even with preserved ejection fraction—supporting its integration into routine evaluation.