ATRIAL FIBRILLATION AND SHANNON ENTROPY

Abstract

Early identification of Atrial Fibrillation (AFIB) is essential to prevent the severe complications associated with this arrhythmia. The diagnosis derived from ECG-Holter Monitoring is ineffective unless the arrhythmia manifests during these examinations. The paper introduces a novel, robust AI (Artificial Intelligence)-driven methodology for the prediction and diagnosis of AFIB, utilizing Heart Rate Variability (HRV) analysis of a patient, and suitable for implementation in clinical practice. This procedure is based on the findings of an observational study involving 7,315 individuals who underwent experimental HRV monitoring. Among the numerous markers assessed in the HRV analysis, four appear capable of diagnosing AFIB with high sensitivity but limited specificity, and only if AFIB occurs during the recording period. Notably, one marker, Shannon Entropy, demonstrates remarkable performance, being able to detect AFIB with both high sensitivity and specificity, even if episodes occurred in the recent or distant past, exhibiting a surprising "memory effect." This fact could enable clinicians to make early predictions of AFIB episodes, facilitating preventive measures and precautionary treatments.