Prospective validation of artificial neural network trained to identify acute myocardial infarction

BACKGROUND Artificial neural networks apply non-linear statistics to pattern recognition problems. One such problem is acute myocardial infarction (AMI), a diagnosis which, in a patient presenting as an emergency, can be difficult to confirm. We report here a prospective comparison of the diagnostic accuracy of a network and that of physicians, on the same patients with suspected AMI. METHODS Emergency department physicians who evaluated 1070 patients 18 years or older presenting to the emergency department of a teaching hospital in California, USA with anterior chest pain indicated whether they thought these patients had sustained a myocardial infarction. The network analysed the patient data collected by the physicians during their evaluations and also generated a diagnosis. FINDINGS The physicians had a diagnostic sensitivity and specificity for myocardial infarction of 73.3% (95% confidence interval 63.3-83.3%) and 81.1% (78.7-83.5%), respectively, while the network had a diagnostic sensitivity and specificity of 96.0% (91.2-100%) and 96.0% (94.8-97.2%), respectively. Only 7% of patients had had an AMI, a low frequency but typical for anterior chest pain. INTERPRETATION The application of non-linear neural computational analysis via an artificial neural network to the clinical diagnosis of myocardial infarction appears to have significant potential.

[1]  Russell C. Eberhart,et al.  Neural network paradigm comparisons for appendicitis diagnoses , 1991, [1991] Computer-Based Medical Systems@m_Proceedings of the Fourth Annual IEEE Symposium.

[2]  Halbert White,et al.  Bootstrapping Confidence Intervals for Clinical Input Variable Effects in a Network Trained to Identify the Presence of Acute Myocardial Infarction , 1995, Neural Computation.

[3]  Richard S. Johannes,et al.  Using the ADAP Learning Algorithm to Forecast the Onset of Diabetes Mellitus , 1988 .

[4]  A Hart,et al.  Evaluating black-boxes as medical decision aids: issues arising from a study of neural networks. , 1990, Medical informatics = Medecine et informatique.

[5]  David G. Bounds,et al.  A comparison of neural network and other pattern recognition approaches to the diagnosis of low back disorders , 1990, Neural Networks.

[6]  W. Baxt.  Use of an artificial neural network for the diagnosis of myocardial infarction. , 1991, Annals of internal medicine.

[7]  R. L. Kennedy,et al.  The early diagnosis of heart attacks: a neurocomputational approach , 1991, IJCNN-91-Seattle International Joint Conference on Neural Networks.

[8]  E F Cook,et al.  Sensitivity of routine clinical criteria for diagnosing myocardial infarction within 24 hours of hospitalization. , 1987, Annals of internal medicine.

[9]  H. Sox,et al.  Clinical prediction rules. Applications and methodological standards. , 1985, The New England journal of medicine.

[10]  R D Cebul,et al.  The importance of disease prevalence in transporting clinical prediction rules. The case of streptococcal pharyngitis. , 1986, Annals of internal medicine.

[11]  A. Ellrodt,et al.  Selecting the best triage rule for patients hospitalized with chest pain. , 1989, The American journal of medicine.

[12]  Prediction rules for patients with suspected myocardial infarction. Applying guidelines in community hospitals. , 1987, Archives of internal medicine.

[13]  Halbert White,et al.  Neural-network learning and statistics , 1989 .

[14]  Use of a rapid assay of subforms of creatine kinase MB to diagnose or rule out acute myocardial infarction. , 1994, The New England journal of medicine.

[15]  Emile Servan-Schreiber,et al.  A Connectionist Approach to the Diagnosis of Dementia , 1988 .

[16]  L. Goldman,et al.  Candidates for thrombolysis among emergency room patients with acute chest pain. Potential true- and false-positive rates. , 1989, Annals of internal medicine.

[17]  William G. Baxt,et al.  Use of an Artificial Neural Network for Data Analysis in Clinical Decision-Making: The Diagnosis of Acute Coronary Occlusion , 1990, Neural Computation.

[18]  R S Wigton,et al.  Transportability of a decision rule for the diagnosis of streptococcal pharyngitis. , 1986, Archives of internal medicine.

[19]  E F Cook,et al.  A computer protocol to predict myocardial infarction in emergency department patients with chest pain. , 1988, The New England journal of medicine.

[20]  A L Brannen,et al.  Prediction of outcome from critical illness. A comparison of clinical judgment with a prediction rule. , 1989, Archives of internal medicine.

[21]  E. S. Pearson,et al.  THE USE OF CONFIDENCE OR FIDUCIAL LIMITS ILLUSTRATED IN THE CASE OF THE BINOMIAL , 1934 .

[22]  William G. Baxt,et al.  Improving the Accuracy of an Artificial Neural Network Using Multiple Differently Trained Networks , 1992, Neural Computation.

[23]  James L. McClelland.  Explorations In Parallel Distributed Processing , 1988 .

[24]  A. V. Sebald,et al.  Use of neural networks for detection of artifacts in arterial pressure waveforms , 1989, Images of the Twenty-First Century. Proceedings of the Annual International Engineering in Medicine and Biology Society,.

[25]  W. Baxt.  A Neural Network Trained to Identify the Presence of Myocardial Infarction Bases Some Decisions on Clinical Associations That Differ from Accepted Clinical Teaching , 1994, Medical decision making : an international journal of the Society for Medical Decision Making.

[26]  D B Matchar,et al.  Noninvasive Carotid Artery Testing: A Meta-analytic Review , 1995, Annals of Internal Medicine.