Scar architecture determinants of ventricular arrhythmias in patients with hypertrophic cardiomyopathy: a cardiac magnetic resonance study

Abstract

Abstract Funding Acknowledgements Type of funding sources: None. Background Hypertrophic cardiomyopathy (HCM) patients with extensive cardiac magnetic resonance (CMR) late gadolinium enhancement (LGE) are at greater risk of ventricular arrhythmias and sudden cardiac death (SCD). LGE-CMR enables the identification within the scar of dense (core) and diffuse (border zone; BZ) fibrosis, as well as of corridors of BZ tissue connecting areas of normal myocardium between core zones (BZ channels). These BZ channels are essential components of reentry circuits and may be critical to entail VT/VF and SCD in HCM patients. Purpose We conducted a retrospective analysis of a cohort of consecutive unselected HCM patients from two referral centers. We sought to describe scar architecture and assess scar composition as a predictor of VT/VF. Methods We analyzed CMR images, generated color-coded pixel signal intensity (PSI) maps based on LGE using the ADAS 3D LV software (Galgo Medical, Barcelona, Spain), and characterized hyper-enhanced areas as core zone, BZ or healthy tissue. BZ channels were identified as corridors of BZ tissue surrounded by scar core or anatomical barriers. The study endpoint (VT/VF) was composed by the occurrence of SCD, resuscitated cardiac arrest (RCA), sustained VT, or appropriate ICD therapy for VT/VF. Cox multivariable analyses were adjusted for the ESC-5 years risk of SCD, end-stage disease, apical aneurysm, and LV scar mass. Results 130 consecutive HCM patients (age 43±17 y; 73% males) with available good quality cardiac MRI were included. ESC 5-years risk was 4.0±2.9, 15% had end-stage disease, 4.6% had apical aneurism and 58% had an ICD implanted. Mean LV mass was 147 g (IQR:108-201). Eighty-six patients (66%) had LGE. Median scar mass was 14.8 g (IQR:6.3-32.8), representing 9.7% (IQR:4.3-18.4) of LV mass. The scar was mainly composed of BZ tissue (median: 12.7 g, IQR:5.5-28.0), while median dense scar mass was 1.5g (IQR:0.2-3.8). BZ channels were found in 46 (35%) patients. During a median follow-up of 87 months (IQR:54-108), 18 (13.8%) patients reached the primary composite endpoint. Scar mass (49.9 g, IQR:24.4-66.5 vs 12.3 g, IQR:3.3-26.7), BZ mass (37.7 g, IQR:21.2-54.8 vs 10.7 g, IQR:2.4-22.7) and core mass (5.7 g, IQR:2.4-11.5 vs 1.3 g, IQR:0.1-2.7) were significantly higher in VT/VF patients (p<0.01 for all comparisons) (Figure 1). BZ channels (Figure 2, A) were observed in 15 out of 18 (83%) patients with VT/VF as compared with 31 out of 112 (28%) of those without (p<0.01) (Figure 2, B). Median BZ channels mass was higher in VT/VF patients versus controls (1.9, IQR:0.8-3.6 vs 0.0, IQR:0.0-0.47, p<0.001). In multivariate Cox regression, the presence of BZ channels was associated with highly increased risk of VT/VF (HR:7.0; 95% CI:1.9-25.4; p=0.03). Conclusions In an unselected cohort of HCM patients, scar architecture affected the risk of VT/VF. Borderzone channels emerged as a powerful risk predictor beyond the magnitude of scar mass.