David R. Okada

Combining mechanistic modeling and machine learning, a multivariable sudden cardiac death predictor outperforms clinical metrics.

Cardiac sarcoidosis is a component of an often multiorgan granulomatous disease of still uncertain cause. It is being recognized with increasing frequency, mainly as the result of heightened awareness and new diagnostic tests, specifically cardiac magnetic resonance imaging and ¹⁸ F-fluorodeoxyglucose positron emission tomography scans. The purpose of this case-based review is to highlight the potentially life-saving importance of making the early diagnosis of cardiac sarcoidosis using these new tools and to provide a framework for the optimal care of patients with this disease. We will review disease mechanisms as currently understood, associated arrhythmias including conduction abnormalities, and atrial and ventricular tachyarrhythmias, guideline-directed diagnostic criteria, screening of patients with extracardiac sarcoidosis, and the use of pacemakers and defibrillators in this setting. Treatment options, including those related to heart failure, and those which may help clarify disease mechanisms are included.

Background: Transition zones between healthy myocardium and scar form a spatially complex substrate that may give rise to reentrant ventricular arrhythmias (VAs). We sought to assess the utility of a novel machine learning approach for quantifying 3-dimensional spatial complexity of grayscale patterns on late gadolinium enhanced cardiac magnetic resonance images to predict VAs in patients with ischemic cardiomyopathy. Methods: One hundred twenty-two consecutive ischemic cardiomyopathy patients with left ventricular ejection fraction ≤35% without prior history of VAs underwent late gadolinium enhanced cardiac magnetic resonance images. From raw grayscale data, we generated graphs encoding the 3-dimensional geometry of the left ventricle. A novel technique, adapted to these graphs, assessed global regularity of signal intensity patterns using Fourier-like analysis and generated a substrate spatial complexity profile for each patient. A machine learning statistical algorithm was employed to discern which substrate spatial complexity profiles correlated with VA events (appropriate implantable cardioverter-defibrillator firings and arrhythmic sudden cardiac death) at 5 years of follow-up. From the statistical machine learning results, a complexity score ranging from 0 to 1 was calculated for each patient and tested using multivariable Cox regression models. Results: At 5 years of follow-up, 40 patients had VA events. The machine learning algorithm classified with 81% overall accuracy and correctly classified 86% of those without VAs. Overall negative predictive value was 91%. Average complexity score was significantly higher in patients with VA events versus those without (0.5±0.5 versus 0.1±0.2; P P =0.002). Conclusions: Substrate spatial complexity analysis of late gadolinium enhanced cardiac magnetic resonance images may be helpful in refining VA risk in patients with ischemic cardiomyopathy, particularly to identify low-risk patients who may not benefit from prophylactic implantable cardioverter-defibrillator therapy. Visual Overview: A visual overview is available for this article.

Background: Patients with cardiac sarcoidosis (CS) may present with arrhythmic events (AE): atrioventricular block (AVB) and/ or ventricular arrhythmias (VA). We sought to: (a) use regional analysis of cardiac magnetic resonance imaging (CMR) to describe anatomic and functional phenotypes of patients with CS and AE; (b) Assess the association of regional CMR abnormalities with the combined endpoint of death, heart transplantation (HT) and AE; and (c) use machine learning (ML) to predict the combined endpoint based on CMR features.Methods: we included 76 patients with CS and CMR. We analyzed cine images to determine regional longitudinal (LS) and radial strain (RS); and late gadolinium enhancement imaging to determine regional scar burden (%scar).Results: Patients with AVB (n = 7), compared with those without, had higher %scar in the anterior (21.8 ± 27.4 vs 5.1 ± 8.9; P = 0.0005) and anteroseptal (19.3 ± 24.5 vs 3.5 ± 5.5; P < .0001) walls. Patients with VA (n = 12), compared with those without, had higher %scar in the basal inferoseptum (20.4 ± 30.8 vs 8.3 ± 13.4; P = .03). During mean follow‐up of 4.4 ± 3.3 years, four patients died or underwent HT; eight had VA; and zero developed AVB. Multiple regional abnormalities were associated with the combined endpoint, including scar in the anteroseptal wall (HR 1.06 [1.02‐1.09] per 1%scar increase, P = .002). The ML algorithm predicted the combined endpoint with a C‐statistic of 0.91.Conclusion: Regional CMR abnormalities are associated with AE in patients with CS.