Artificial neural networks and robust Bayesian classifiers for risk stratification following uncomplicated myocardial infarction

• Published in: International journal of cardiology Vol. 101; no. 3; pp. 481 - 487
• Main Authors: Bigi, Riccardo, Gregori, Dario, Cortigiani, Lauro, Desideri, Alessandro, Chiarotto, Francesco A., Toffolo, Gianna M.
• Format: Journal Article
• Language: English
• Published: Shannon Elsevier Ireland Ltd 08.06.2005; Elsevier Science
• Subjects: Algorithms; Angina, Unstable - epidemiology; Angina, Unstable - prevention & control; Artificial intelligence; Bayes Theorem; Biological and medical sciences; Cardiology. Vascular system; Confidence Intervals; Coronary heart disease; Death, Sudden, Cardiac - epidemiology; Death, Sudden, Cardiac - prevention & control; Echocardiography, Stress; Electrocardiography; Exercise ECG; Female; Follow-Up Studies; Heart; Humans; Incidence; Male; Medical sciences; Middle Aged; Myocardial infarction; Myocardial Infarction - diagnosis; Myocardial Infarction - epidemiology; Myocarditis. Cardiomyopathies; Neural network; Neural Networks (Computer); Predictive Value of Tests; Prognosis; Reproducibility of Results; Retrospective Studies; Risk Assessment - statistics & numerical data; Risk Factors; Secondary Prevention; Severity of Illness Index; Stress echocardiography; Survival Rate; Time Factors; Myocardial infarction; Stress echocardiography; Prognosis; Neural network; Artificial intelligence; Exercise ECG; Sonography; Physical exercise; Echocardiography; Cardiovascular disease; Risk; Myocardial disease; Phlebology; Stress; Electrodiagnosis; Heart disease; Risk factor; Electrocardiography; Exercise tolerance test; Circulatory system; Cardiology; Stratification
• ISSN: 0167-5273; 1874-1754
• DOI: 10.1016/j.ijcard.2004.07.008

To compare artificial neural networks (ANN) and robust Bayesian classifiers (RBC) in predicting outcome following acute myocardial infarction (AMI). Clinical, exercise ECG and stress echo variables by 496 patients with AMI were used to predict the cumulative end-point of cardiac death, nonfatal reinfarction and unstable angina. Revascularized patients were censored. Short (200 days)-, medium (400 days)- and long (1000 days)-term observation intervals, including 50%, 75% and 90% of the events, respectively, were considered. At each interval, any patient was binary assigned to the “event” or “no event” class. A multilayer feedforward ANN, trained by a back propagation algorithm, was used. RBC, using the leave-one-out technique, were derived. The accuracy of both techniques was compared to the default accuracy (DA) obtained by assigning all subjects to the largest class. 14 death, 27 reinfarction and 29 unstable angina were observed during a mean follow-up of 24 [95% confidence interval (CI) 19 to 22] months. The accuracy of ANN and RBC and DA were 70%, 81% and 74% at short, 67%, 73% and 56% at medium and 64%, 68% and 62% at long-term follow-up. (1) ANN do not improve the prognostic classification of patients with uncomplicated AMI as compared to RBC. (2) In particular, short-term prognostic accuracy seems insufficient.