TY - JOUR
T1 - Best leads in the standard electrocardiogram for the emergency detection of acute coronary syndrome.
AU - Green, Michael
AU - Ohlsson, Mattias
AU - Lundager Forberg, Jakob
AU - Björk, Jonas
AU - Edenbrandt, Lars
AU - Ekelund, Ulf
PY - 2007
Y1 - 2007
N2 - Background and Purpose: The purpose of this study was to determine which leads in the standard 12-lead electrocardiogram (ECG) are the best for detecting acute coronary syndrome (ACS) among chest pain patients in the emergency department. Methods: Neural network classifiers were used to determine the predictive capability of individual leads and combinations of leads from 862 ECCs from chest pain patients in the emergency department at Lund University Hospital. Results: The best individual lead was aVL, with an area under the receiver operating characteristic curve of 75.5%. The best 3-lead combination was III, aVL, and V-2, with a receiver operating characteristic area of 82.0%, compared with the 12-lead ECG performance of 80.5%. Conclusions: Our results indicate that leads III, aVL, and V2 are sufficient for computerized prediction of ACS. The present results are likely important in situations where the 12-lead ECG is impractical and for the creation of clinical decision support systems for ECG prediction of ACS.
AB - Background and Purpose: The purpose of this study was to determine which leads in the standard 12-lead electrocardiogram (ECG) are the best for detecting acute coronary syndrome (ACS) among chest pain patients in the emergency department. Methods: Neural network classifiers were used to determine the predictive capability of individual leads and combinations of leads from 862 ECCs from chest pain patients in the emergency department at Lund University Hospital. Results: The best individual lead was aVL, with an area under the receiver operating characteristic curve of 75.5%. The best 3-lead combination was III, aVL, and V-2, with a receiver operating characteristic area of 82.0%, compared with the 12-lead ECG performance of 80.5%. Conclusions: Our results indicate that leads III, aVL, and V2 are sufficient for computerized prediction of ACS. The present results are likely important in situations where the 12-lead ECG is impractical and for the creation of clinical decision support systems for ECG prediction of ACS.
KW - acute coronary syndrome
KW - artificial neural networks
KW - myocardial infarction
KW - electrocardiography
U2 - 10.1016/j.jelectrocard.2006.12.011
DO - 10.1016/j.jelectrocard.2006.12.011
M3 - Article
SN - 1532-8430
VL - 40
SP - 251
EP - 256
JO - Journal of Electrocardiology
JF - Journal of Electrocardiology
IS - 3
ER -