Recognition of lead reversals in pediatric electrocardiograms

Misplacement of electrodes during the recording of an electrocardiogram occurs occasionally and this may lead to misdiagnosis and improper treatment.(1,2) 2 Lead reversals have been found in nearly 2% of the electrocardiograms recorded at an emergency department,(3) and even experienced readers of electrocardiograms often overlook the lead reversals. Therefore, commonly used computerized electrocardiographs could present a warning to the technician that "appearance suggests arm lead reversal." This type of electrocardiograph contains interpretation programs with rule-based criteria for the detection of right/left arm lead reversal.(4) Recently, another technique-artificial neural networks-was presented for automated recognition of lead reversals.(3,5) Artificial neural networks represents a computer-based method that lias been shown to be of value in pattern recognition tasks, e.g. diagnostic tasks in clinical medicine.(6-8) The purpose of this study was to assess the performance of existing methods for recognizing right/left arm lead reversal in a group of pediatric electrocardiograms and to develop a neural network for improved recognition of right/left arm lead reversal in pediatric electrocardiograms.