Interobserver agreement in the assessment of septal LGE in patients with suspected cardiomyopathies and conduction defects: the impact of genetic data on cardiac MRI interpretation

Abstract Introduction Dilated cardiomyopathy (DCM) and non-dilated left ventricular cardiomyopathy (NDLVC) are heterogeneous conditions associated with significant arrhythmic risk (1). Cardiac magnetic resonance (CMR) imaging, particularly late gadolinium enhancement (LGE), is a key tool for arrhythmic risk stratification, with septal LGE being a strong predictor of arrhythmic events, independent of left ventricular ejection fraction (2). However, in patients with left bundle branch block (LBBB), altered ventricular activation and mechanical dyssynchrony can cause imaging artifacts, potentially leading to overestimation of myocardial fibrosis (3). Purpose This study aimed to investigate whether the presence of septal LGE is assessed consistently by observers with varying levels of experience in patients with LBBB and suspected DCM or NDLVC. Additionally, the study explored whether the inclusion of genetic data influences the interpretation of septal LGE on CMR, potentially modifying the certainty of imaging findings. Methods A retrospective review of CMR scans from February 2022 to January 2025 was conducted in patients who underwent next-generation sequencing (NGS). CMR was performed on a 1.5 T Siemens Magnetom Sola scanner, including Cine, T1 and T2 mapping, and LGE (Mag-IR and PSIR). Three observers (two experienced, one non-experienced) assessed septal LGE and rated their certainty using a Likert scale. Images were first assessed without genetic data and re-evaluated one month later with genetic information. Interobserver agreement was assessed using Fleiss' Kappa, and Cohen's weighted Kappa was used to compare blinded and genetic conditions. Non-parametric bootstrap was applied to assess the significance of differences between conditions. Results The study included 23 patients (18 male, 5 female). Nineteen patients (83%) carried at least one genetic variant, of which the majority were classified as variants of uncertain significance (C3). Six patients (26%) carried two or more variants. The interobserver agreement for the presence of septal LGE was moderate, with a Fleiss' Kappa of 0.718 (95% CI 0.373, 0.930). Pairwise Cohen's weighted Kappa showed an increase in agreement with the inclusion of genetic data (from 0.743 to 0.851 among the two expert readers), but these differences were not statistically significant, indicating minimal impact of genetic information on the certainty of LGE interpretation. Conclusions Septal LGE is a key prognostic factor in arrhythmic risk assessment in patients with suspected cardiomyopathies. However, its interpretation requires caution, particularly in individuals with conduction defects, as genetic data did not significantly influence imaging interpretation, particularly in a cohort with a predominance of variants of uncertain significance. Standardizing LGE assessment, including the integration of certainty ratings, may enhance diagnostic precision and facilitate more informed clinical decision-making.Septal LGE certainty rating: 2 Septal LGE certainty rating: 5