The Use of Artificial Intelligence for the Prioritization of Causative Genetic Background in a Brugada Syndrome Cohort: an Observational Retrospective Study

Brugada Syndrome (BS) is an inherited cardiac electrical disorder that can cause syncope and sudden cardiac arrest in young asymptomatic individuals. It is suspected to contribute to 4-12% of cases of sudden cardiac death in the general population. Diagnosis relies on identifying a type I ECG pattern characterized by ST-segment elevation with a coved morphology in the right precordial leads. The prevalence in Western countries is estimated at 1:5000. Currently, implantation of a cardioverter defibrillator (ICD) is the only treatment option, but risk stratification guidelines remain incomplete, particularly for asymptomatic individuals.

BS is inherited as an autosomal dominant trait with incomplete penetrance. While 23 genes have been associated with BS susceptibility, 70% of patients remain genetically uncharacterized, suggesting a more complex inheritance pattern. Genetics have not been incorporated into risk stratification guidelines, despite evidence linking certain genetic variants to higher arrhythmic risk. This knowledge gap underscores the importance of expanding our understanding of BS genetics to enhance diagnostic sensitivity and patient management.

This protocol builds upon preliminary data from a study granted by the Italian Ministry of Health (GR-2016-02362316), in which next-generation sequencing (NGS) was used to investigate the entire coding regions (Whole Exome Sequencing_WES) of 200 BS patients. The study aimed to identify new BS candidate genes and characterize the genetic basis of the condition.

The cohort was selected based on the presence of a type I ECG, confirmed either spontaneously or induced by flecainide or ajmaline. Patients underwent thorough cardiac evaluations to rule out other conditions. Follow-up included yearly assessments and more frequent evaluations for patients with a higher risk of ventricular tachycardia.

A large number of genetic variants were identified by exploiting WES, prompting the use of Artificial Intelligence (AI) to prioritize the sequencing data. AI techniques, including advanced algorithms and machine learning, can streamline the identification of potentially disease-causing genetic variations by filtering out common variants, predicting pathogenicity, and integrating clinical data.