Work

New ECG biomarkers and sex-stratified models for the detection of Arrhythmogenic Cardiomyopathy with left ventricular involvement

Year: 2024

Type: article

Abstract: Arrhythmogenic Cardiomyopathy (ACM) is a rare cardiac disease that can lead to severe cardiac structural and electrical abnormalities. Given its genetic basis, other relatives might be affected. Thus, family screening is highly recommended. This clinical diagnosis remains challenging and relies upon a multiparametric approach, which includes several ECG biomarkers without sex distinction. Aiming to improve the yield of current criteria in patients with left ventricular and biventricular ACM involvement, our study introduces innovative biomarkers based on the ECG, the Vectocardiogram (VCG), and the Principal Components Analysis (PCA) that fed different Logistic Regression models. Furthermore, a sex-focused stratification was additionally performed to find differences in the performance of the models. In this work, 87 individuals (age 43±18 years, 53% females) underwent cardiological studies in the setting of family screening of ACM, where 41 were considered ACM patients and 46 healthy relatives. Values of the G-metric that combines sensitivity and specificity were obtained from our classification models.The best G value reached in the whole cohort to identify patients was 0.84 (sensitivity 0.85, specificity 0.83). When analyzing the dataset stratified by sex, the models showed improved performance for males, with a G value of 0.88 (sensitivity 0.87, specificity 0.89). The female-specific model achieved a maximum G value of 0.81 (sensitivity 0.83, specificity 0.79).These results suggest that sex-stratified models may enhance ACM detection, specially in males. After validation in larger cohorts, the proposed ECG biomarkers and machine learning approaches might improve diagnostic accuracy for ACM during family screening.

Source: Biomedical Signal Processing and Control

DOI: 10.1016/j.bspc.2024.107224

Authors: Santiago Jiménez-Serrano, Jorge Sanz-Sánchez, Yolanda Vives Gilabert, José Millet, Esther Zorio, Francisco Castells