BackgroundThe risk stratification of left ventricular noncompaction (LVNC) remains ambiguous. LV entropy derived from late gadolinium enhancement (LGE) in cardiac magnetic resonance (CMR) as a novel measurement of myocardial heterogeneity may serve as the substrate of major adverse cardiovascular events (MACEs). This retrospective study aimed to investigate the value of LV entropy for predicting MACEs in LVNC patients. MethodsConsecutive patients who underwent CMR and met the diagnosis criteria of LVNC were included. All patients were follow-up for MACEs (cardiac death, ventricular arrhythmia requiring therapy or heart failure hospitalization), and their LV entropy derived from the distribution of pixel signal intensities in the LGE of the LV myocardium was analyzed. ResultsOne hundred and forty-three patients (mean age 40 years, 64.3% male) were followed for a median of 3.2 years, and forty-two (29.4%) experienced MACEs. Presenting of symptoms, left ventricular end-diastolic diameter (LVEDD), LV end-diastolic volume (LVEDV) index, LV end-systolic volume (LVESV) index, LV ejection fraction (LVEF), LGE extent, and LV entropy showed association with MACEs. LV entropy maintained independent association with MACEs (HR: 4.76, 95%CI 3.68–5.15, p < 0.001) in multivariable analysis. Entropy was also strong independent predictor of MACEs in patients with and without LGE (HR: 5.89, 95% CI4.18–7.73, p < 0.001; HR: 3.06, 95% CI:1.53–4.80, p = 0.013, respectively). ConclusionsLV entropy can predict MACEs in LVNC patients and provide incremental prognostic value on top of LVEF and LGE. Also, LV entropy may help risk stratification in LGE-negative LVNC patients.