Automatic classification of transient ischaemic and transient non-ischaemic heart-rate related ST segment deviation
1Laboratory of Biomedical Computer Systems and Imaging, Faculty of Computer and Information Science, University of Ljubljana, Ljubljana, Slovenia. jana.faganeli@fri.uni-lj.si
Insights
This study developed an automated method to distinguish between ischemic and non-ischemic ST segment deviations in ECGs. Combining heart rate, Mahalanobis distance, and Legendre polynomial features achieved high accuracy in classifying these cardiac events.
Area of Science:
- Cardiology
- Biomedical Engineering
- Signal Processing
Background:
- Ambulatory ECG monitoring reveals transient ST segment deviations, including both ischemic and non-ischemic heart-rate related types.
- Non-ischemic ST segment deviations, driven by heart rate changes, can hinder accurate detection of true ischemic episodes.
- Automated classification systems require robust methods to differentiate these ST segment deviation types.
Purpose of the Study:
- To develop and evaluate an automated method for classifying transient ischemic and non-ischemic heart-rate related ST segment deviation episodes in ambulatory ECG records.
- To identify the most effective features for distinguishing between these two types of ST segment deviations.
Main Methods:
- Utilized features including heart rate changes, Mahalanobis distance of Karhunen-Loève transform (KLT) coefficients of the QRS complex, time-domain ST segment morphology parameters, and Legendre orthonormal polynomial coefficients.
- Employed decision trees for classification, leveraging Legendre polynomials for their suitability in representing ST segment morphology.
- Validated the method on the Long-Term ST Database, assessing performance using aggregate statistics and bootstrap methods.
Main Results:
- The combination of heart rate features, Mahalanobis distance, and Legendre orthonormal polynomial coefficients yielded the best classification performance.
- Achieved an average sensitivity of 98.1% and an average specificity of 85.2% in distinguishing between ischemic and non-ischemic ST segment deviations.
- The chosen features, particularly Legendre polynomials, provided effective insight into ST segment morphology changes.
Conclusions:
- An automated decision tree-based classification method effectively differentiates between ischemic and non-ischemic heart-rate related ST segment deviations.
- The integration of specific ECG signal features significantly improves the accuracy of automated ischemic event detection.
- This approach holds promise for enhancing the reliability of automated analysis in ambulatory ECG monitoring.
Abstract:
In ambulatory ECG records, besides transient ischaemic ST segment deviation episodes, there are also transient non-ischaemic heart-rate related ST segment deviation episodes present, which appear only due to a change in heart rate and thus complicate automatic detection of true ischaemic episodes. The goal of this work was to automatically classify these two types of episodes. The tested features to classify the ST segment deviation episodes were changes of heart rate, changes of the Mahalanobis distance of the first five Karhunen-Loève transform (KLT) coefficients of the QRS complex, changes of time-domain morphologic parameters of the ST segment and changes of the Legendre orthonormal polynomial coefficients of the ST segment. We chose Legendre basis functions because they best fit typical shapes of the ST segment morphology, thus allowing direct insight into the ST segment morphology changes through the feature space. The classification was performed with the help of decision trees. We tested the classification method using all records of the Long-Term ST Database on all ischaemic and all non-ischaemic heart-rate related deviation episodes according to annotation protocol B. In order to predict the real-world performance of the classification we used second-order aggregate statistics, gross and average statistics, and the bootstrap method. We obtained the best performance when we combined the heart-rate features, the Mahalanobis distance and the Legendre orthonormal polynomial coefficient features, with average sensitivity of 98.1% and average specificity of 85.2%.
Related Concept Videos
01:29Electrocardiogram
Three major waveforms are present in a typical ECG recording: the P wave, the QRS complex, and the T...
01:24ECG Interpretation of Rhythms
Components of the Electrocardiogram
The primary components of a normal ECG waveform in Normal sinus rhythm(NSR) include the P wave, PR interval, QRS complex, ST segment, T wave, and occasionally a U wave.
ECG waveforms are divided by vertical and horizontal lines at standard intervals.
The horizontal axis measures time and rate, and the vertical axis measures amplitude or voltage. When...
01:28Electrocardiogram Fundamentals
An electrocardiogram (ECG) is a diagnostic tool for identifying cardiac conditions such as arrhythmias, conduction abnormalities, and myocardial ischemia.
Definition
An electrocardiogram (ECG) visualizes the heart's electrical activity by tracing the electrical movement associated with each heartbeat on a graph or monitor. As the heart beats, an electrical wave passes through it, correlating with the cardiac cycle events.
Parts of an ECG
An ECG utilizes electrodes on the skin to...
01:25ECG Interpretation of Arrhythmias II: Atrial, Junctional and Ventricular Arrhythmias
01:23Holter Monitor: 24-Hour Monitoring
01:30Pulse rhythm
Conversely, an irregular pulse pattern is termed dysrhythmia, stemming from disruptions in cardiac muscle...