Abstract
Logistic regression and Gaussian mixture model (GMM) classifiers have been trained to estimate the probability of acute myocardial infarction (AMI) in patients based upon the concentrations of a panel of cardiac markers. The panel consists of two new markers, fatty acid binding protein (FABP) and glycogen phosphorylase BB (GPBB), in addition to the traditional cardiac troponin I (cTnI), creatine kinase MB (CKMB) and myoglobin. The effect of using principal component analysis (PCA) and Fisher discriminant analysis (FDA) to preprocess the marker concentrations was also investigated. The need for classifiers to give an accurate estimate of the probability of AMI is argued and three categories of performance measure are described, namely discriminatory ability, sharpness, and reliability. Numerical performance measures for each category are given and applied. The optimum classifier, based solely upon the samples take on admission, was the logistic regression classifier using FDA preprocessing. This gave an accuracy of 0.85 (95% confidence interval: 0.78–0.91) and a normalised Brier score of 0.89. When samples at both admission and a further time, 1–6 h later, were included, the performance increased significantly, showing that logistic regression classifiers can indeed use the information from the five cardiac markers to accurately and reliably estimate the probability AMI.
Similar content being viewed by others
References
Peterson S, Rayner M (2002) Coronary heart disease statistics: 2002 edition, British Heart Foundation
Wu A (Ed) (1998) Cardiac markers, pathology and labratory medicine. Humana Press, Totowa, NJ
Goldman L, Cook E, Brand D, Lee T, Rouan G, Weisberg M, Acampora D, Stasiulewicz C, Walshon J, Terranova G, Gottlieb L, Kobernick L, Goldstein-Wayne B, Copen D, Daley K, Brandt A, Jones D, Mellors J, Jakubowski R (1998) A computer protocol to predict myocardial infarction in emergency department patients with chest pain. New Engl J Med 318(13):797–803
Goldman L, Weinberg M, Weisberg M, Olshen R, Cook E, Sargent R, Lams G, Dennis C, Wilson C, Deckelbaum L, Fineberg H, Stiratelli R (1982) A computer-derived protocol to aid in the diagnosis of emergency room patients with acute chest pain. New Engl J Med 307(10):588–596
Selker H, Beshansky J, Griffith J, Aufderheide T, Ballin D, Bernard S, Crespo S, Feldman J, Fish S, Gibler W, Kiez D, McNutt R, Moulton A, Ornato J, Podrid P, Pope J, Salem D, Sayre M, Woolard R (1998) Use of the Acute Cardiac Ischemia Time-Insensitive Predictive Instrument (ACI-TIPI) to assist with triage of patients with chest pain or other symptoms suggestive of acute cardiac ischemia: a multicenter, controlled clinical trial. Ann Intern Med 129(1):845–855
Fesmire F, Hughes A, Fody E, Jackson A, Fesmire C, Gilbert M, Stout P, Wojcik J, Wharton D, Creel J (2002) The Eulanger chest pain evaluation protocol: a one-year experience with serial 12-lead ECG monitoring, two-hour delta serum marker measurements, and selective nuclear stress testing to identify and exclude acute coronary syndromes. Ann Emerg Med 40(6):584–594
Pozen M, D’Agostino R, Selker H, Sytkowski P, Hood W (1984) A predictive instrument to improve coronary-care-unit admission practices in acute ischemic heart disease. New Engl J Med 310(20):1273–1278
Tierney W, Roth B, Psaty B, McHenry R, Fitzgerald J, Stump D, Anderson K, Ryder K, McDonald C, Smith D (1985) Predictors or myocardial infarction in emergency room patients. Crit Care Med 13(7):526–531
Harrison R, Marshall S, Kennedy R (1992) Minimum-risk decisions in the management of suspected heart attack: an application of the Boltzmann perceptron network. In: Rogers S (Ed) Proceedings of SPIE. Applications of Artificial Neural Networks III, vol 1709, pp 1701–1082
Fraser H, Pugh R, Kennedy R, Ross P, Harrison R (1994) A comparison of back propagation and radial basis functions in the diagnosis of myocardial infarction. In: Proceedings of the International Conference on Neural Networks and Expert Systems in Medicine and Health Care, pp 76–84
Ellenuis J, Groth T, Lindahl B, Wallentin L (1997) Early assessment of patients with suspected acute myocardial infarction by biochemical monitoring and neural network analysis. Clin Chem 43(10):1919–1925
Groth T, Ellenius J (2002) Provision of decision support for acute myocardial infarction. United States Patent 6:443–889
Bishop C (1995) Neural networks for pattern recognition. Clarendon Press, Oxford
de Dombal F (1996) Medical informatics: the essentials, Butterworth-Heinemann, Linacre House. Jordan Hill, Oxford, UK
Hanley J, McNeil B (1982) The meaning and use of the area under a receiver operator characteristic (ROC) curve. Radiology 143:29–36
Hilden J, Habbema J, Bjerregaard B (1978) The measurement of performance in probabilistic diagnosis: II. Trustworthiness of exact values of the diagnostic probabilities. Methods Inform Med 17(4):227–237
Hilden J, Habbema J, Bjerregaard B (1978) The measurement of performance in probabilistic diagnosis: III. Methods based on continuous functions of diagnostic probabilities. Methods Inform Med 17(4):238–246
Habbema J, Hilden J, Bjerregaard B (1978) The measurement of performance in probabilistic diagnosis: I. The problem, descriptive tools, and measures based on classification matricies. Methods Inform Med 17(4):217–226
Habbema J, Hilden J (1981) The measurement of performance in probabilistic diagnosis: IV. Utility considerations in therapeutics and prognostics. Methods Inform Med 20(2):80–96
Habbema J, Hilden J, Bjerregaard B (1981) The measurement of performance in probabilistic diagnosis: V. General recommendations. Methods Inform Med 20(2):97–100
Huang P, Harris C, Nixon M (1998) Comparing different template features for recognising people by their gait. In: Proceedings of the Ninth British Machine Vision Conference, vol 2, pp 639–648
Nabney I (2002) Netlab: algorithms for pattern recognition, advances in pattern recognition. Springer, Berlin
Inc., TM, Matlab version 6.1 release 12.1, Software (2001)
Armstrong G, Petry C, Wagner G, Wu A (1999) Reflex algorithm for early and cost effective diagnosis of myocardial infarctions suitable for automated diagnostic platforms. European Patent Application number: 99110216.1
Magee D, Boyle R (1998) Improving class seperation in principal component analysis using delta analysis, research Report: School of Comnputer Studies. University of Leeds
Fesmire F, Percy R, Bardoner J, Wharton D, Calhoun F (1998) Serial creatinine kinase (CK) MB testing during the emergency department evaluation of chest pain: utility of a 2-hour delta CK-MB of +1.6 ng/ml. Am Heart J 136(2):237–244
Acknowledgments
The first author would like to the Department for Employment and Learning, Northern Ireland, for providing a Ph.D. studentship, and Randox Laboratories for financial support and provision of data.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Wilson, P.C., Irwin, G.W., Lamont, J.V. et al. Probabilistic classification of acute myocardial infarction from multiple cardiac markers. Pattern Anal Applic 12, 321–333 (2009). https://doi.org/10.1007/s10044-008-0126-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10044-008-0126-x