[The predictive value of cardiac MRI for the first episode of malignant ventricular arrhythmias in arrhythmogenic right ventricular cardiomyopathy].

<b>Objective:</b> To explore the value of cardiac magnetic resonance imaging (CMR) derived left ventricular late gadolinium enhancement (LV LGE) for the primary prevention of malignant ventricular arrhythmias in arrhythmogenic right ventricular cardiomyopathy (ARVC) patients. <b>Methods:</b> This was a single-center retrospective study. Consecutive ARVC patients who underwent CMR at Fuwai Hospital between January 2016 and September 2020, with no history of malignant ventricular arrhythmias at diagnosis, were enrolled. Clinical data and CMR characteristics were collected. The primary endpoint was defined as new-onset malignant ventricular arrhythmias related events, including sustained ventricular tachycardia, ventricular fibrillation/flutter, sudden cardiac death, cardiac arrest, and appropriate implantable cardioverter-defibrillator discharge. Follow-up via telephone interviews and medical records was conducted to confirm endpoint occurrences, and patients were categorized into event-free and event groups based on endpoint status. Univariable and multivariable Cox regression analysis were performed to identify independent risk factors for malignant ventricular arrhythmias in ARVC patients. Subgroup analyses were conducted based on the ARVC 5-year risk score (cutoff: 25%) and the median value of LV LGE percentage (cutoff: 13%). Kaplan-Meier curves were plotted, and log-rank tests were used to compare the difference in the incidence of primary endpoint events between subgroups. Receiver operating characteristic curves and likelihood ratio test were used to evaluate the incremental prognostic value of LV LGE percentage beyond the ARVC 5-year risk score. <b>Results:</b> A total of 172 ARVC patients were enrolled, aged (39.0±16.6) years, including 73 females (42.4%). During a follow-up of 53.1 (25.4, 76.9) months, 51 patients (29.7%) experienced malignant ventricular arrhythmias related events, including 3 cases of sudden cardiac death, 1 cardiac arrest, 33 sustained ventricular tachycardia and 14 appropriate implantable cardioverter-defibrillator discharges. Multivariable Cox regression analysis indicated that the ARVC 5-year risk score (<i>HR</i>=1.028, 95%<i>CI</i> 1.015-1.041, <i>P</i><0.001) and LV LGE percentage (<i>HR</i>=1.059, 95%<i>CI</i> 1.032-1.087, <i>P</i><0.001) were independent risk factors of the primary endpoint events. Kaplan-Meier analysis using composite stratification (ARVC 5-year risk score cutoff: 25%; LV LGE percentage cutoff: 13%) demonstrated that patients with both high risk scores (≥25%) and extensive LV LGE (≥13%) had the highest risk of primary endpoint events. Notably, among patients with ARVC 5-year risk scores <25%, those with LV LGE≥13% had a higher incidence of primary endpoint events than those without (log-rank <i>P</i>=0.037). The composite prediction model combining the 5-year risk score and left ventricular LGE percentage demonstrated significantly improved predictive performance (area under the curve (<i>AUC</i>)=0.82, 95%<i>CI</i> 0.75-0.90; likelihood ratio test all <i>P</i><0.001) compared to single-variable models (left ventricular LGE percentage alone: <i>AUC</i>=0.71, 95%<i>CI</i> 0.63-0.82, <i>P</i>=0.01; 5-year risk score alone: <i>AUC</i>=0.71, 95%<i>CI</i> 0.62-0.81, <i>P</i>=0.02). <b>Conclusion:</b> LV LGE percentage independently predict new-onset malignant ventricular arrhythmias in ARVC patients and provided incremental prognostic value based on the existing ARVC 5-year risk score.