Prediction of ventricular arrhythmic outcomes in suspected cardiac sarcoidosis: a comparison of cardiovascular magnetic resonance phenotyping vs. societal recommendations for implantable cardioverter-defibrillator placement

Abstract Background and Aims Implementing societal recommendations for primary prevention implantable cardioverter-defibrillators (ICDs) in cardiac sarcoidosis requires an accurate diagnosis. However, cardiac sarcoidosis diagnostic schemes are inconsistent and often produce conflicting results. This study aimed to compare the discriminative accuracy of cardiovascular magnetic resonance imaging (CMR) phenotyping with the societal recommendations for predicting long-term ventricular arrhythmic outcomes in patients with suspected cardiac sarcoidosis, regardless of their diagnostic status. Methods This multicentre study included patients with histology-supported sarcoidosis who underwent CMR for suspected cardiac involvement and were ineligible for secondary prevention ICDs. The study outcome was a composite of fatal or life-threatening ventricular arrhythmias. Outcomes were compared based on eligibility for ICDs by societal recommendations and CMR phenotyping. Results Among 1514 patients, 84 experienced the study outcome during a median follow-up of 4.5 years and a maximum follow-up of 10 years. The high-risk CMR phenotype was associated with higher 5- and 10-year incidences of the outcome (24.0% and 35.0%, respectively) compared with those who met societal recommendations. Patients with low-risk phenotypes had lower incidences (0.7% and 2.6%). Cardiovascular magnetic resonance imaging phenotyping outperformed societal recommendations in discriminative accuracy, with areas under the curve of 0.861 and 0.776 for 5- and 10-year outcomes, respectively. Additionally, CMR phenotyping had the highest adjusted subdistribution hazard ratio (19.8) for the study outcome. Conclusions In patients with suspected cardiac sarcoidosis, CMR phenotyping showed greater discriminative accuracy than societal recommendations for predicting fatal or life-threatening ventricular arrhythmias, suggesting that it may be more effective at identifying candidates for primary prevention ICDs.