ECG based prediction of Atrial Fibrillation using Support Vector Classifier

Abstract

Original scientific paperIn patients undergoing Coronary Artery Bypass Grafting (CABG) surgery postoperative atrial fibrillation (AF) occurs with prevalence of up to 40%. The highest incidence is between the second and third day after the operation. Following cardiac surgery AF causes various complications, hemodynamic instability, and can cause heart attack, cerebral and other thromboemolisms. AF increases morbidity, duration and expense of medical treatment. This study aims to identify patients at high risk of postoperative AF. An early prediction of AF would provide a timely prophylactic treatment and would reduce incidence of arrhythmia. Patients at low risk of postoperative AF could be excluded from the side effects of anti-arrhythmic drugs. The investigation included 50 patients in whom lead II electrocardiograms were continuously recorded for 48 hours following CABG. Univariate statistical analysis was used in the search of signal features that might predict AF. The most promising identified features were: P wave duration, RR interval duration and PQ segment level. On the basis of these a nonlinear multivariate prediction model was made deploying a Support Vector Machine (SVM) classifier. The prediction accuracy was found uprising over the time. At 48 hours following CABG; the measured best average sensitivity was 95.9% and specificity 93.4%. The positive and negative predictive accuracy were 88.9% and 98.8%, respectively and the overall accuracy was 94.6%. In regard to the prediction accuracy, the risk assessment and prediction of postoperative AF are optimal to be done in the period between 24 and 48 hours following CABG.Key words: Post-operative atrial fibrillation, Coronary artery bypass grafting (CABG), Cubic spline baseline correction, PQ segment levelPredvianje postoperacijske fibrilacije atrija korištenjem SVM klasifikatora. Postoperacijska fibrilacija atrija (AF) pojavljuje se u oko 40% pacijenata podvrgnutih operaciji aortokoronarnog premoštenja (CABG), s najvećom učestalosti pojavljivanja oko trećeg dana nakon operacije. Postoperacijska AF može stvoriti brojne komplikacije poput hemodinamske nestabilnosti, srčanog udara, cerebralnih i drugih tromboembolija; povećava morbiditet, trajanje i troškove liječenja. Studija ima za cilj rano otkrivanje pacijenta sa visokim rizikom razvoja postoperacijske AF, što bi osiguralo pravovremenu profilaktičku terapiju i smanjilo učestalost aritmije, dok bi pacijenti sa niskim rizikom razvoja postoperacijske AF bili pošteeni nuspojava antiaritmičkih lijekova. Podatkovni skup uključuje 50 pacijenata, snimanih II standardnim odvodom elektrokardiografa, kontinuirano u razdoblju od 48 sati nakon operacije. Univarijatna statistička analiza korištena je za odreivanje parametara signala koji bi mogli predvidjeti AF, te su kao najznačajniji odreeni: trajanje P vala, trajanje RR intervala i razina PQ spojnice; na temelju kojih je izveden nelinearni multivarijatni predikcijski model zasnovan na SVM klasifikatoru. Ukupna predikcijska točnost modela povećava se s vremenom. ...

Analyzing artificial intelligence systems for the prediction of atrial fibrillation from sinus-rhythm ECGs including demographics and feature visualization

“…Traditional techniques for AF prediction combine discrete features extracted from ECGs and machine learning classifiers 12 , 13 . Promising features for the prediction of AF, such as P wave duration, RR interval duration, and PQ segment level, have been considered in the literature 14 . However, this kind of approaches takes into account only little information from the ECG and require a time-consuming phase to extract and select handcrafted features 15 , also prone to errors.…”

Simplified 2D Bidomain Model of Whole Heart Electrical Activity and ECG Generation

“…In an earlier study [21,22], we claimed that changes in the PQ segment level (defined as elevation / depression of the segment level between P wave end and Q wave onset) and the falling P wave slope are related to premature atrial repolarization, APD shortening and loss of AP plateau, and this is often found in patients prone to atrial fibrillation [21,22]. These hypotheses were tested again with the 2D cardiac model of this study and the results (Fig.8.…”