Electrocardiographic and Vectorcardiographic Markers of Cardiac Involvement in Systemic Lupus Erythematosus: A Machine Learning Approach

Alejandro Perez1, Elisa Ramirez2, Francisco Castells3, Muhammad Soyfoo4, Ruben Casado5, Jose Millet6
1ITACA Institute, 2Institute ITACA, Universitat Politecnica de Valencia, 3Universitat Politècnica de Valencia, 4Rhumatologie, Hopital Erasme, 5Université Libre de Bruxelles, 6BioITACA-UPV


Abstract

Systemic lupus erythematosus (SLE) is a chronic autoimmune disease frequently associated with cardiovascular involvement, often subclinical and difficult to detect. This study investigates differences in cardiac electrical activity between SLE patients and healthy controls by analysing standard ECG signals and reconstructed vectorcardiographic (VCG). ECG recordings from 64 SLE patients and 64 matched controls were processed. A statistical comparison using the Mann-Whitney U test was conducted to identify features with significant differences between groups, with T loop speed and shape among the most discriminative. The set of derived features was used for dimensionality reduction via principal component analysis (PCA) and four machine learning models were applied. The models achieved high performance, with random forest reaching up to 96% accuracy. These findings suggest that ECG and VCG analysis, combined with machine learning, could aid in the early detection of cardiac involvement in SLE. However, larger datasets are required to confirm these findings and support clinical translation.