Classification of ischaemic episodes with ST/HR diagrams.

Coronary artery disease is the developed world's premier cause of mortality and the most probable cause of myocardial ischaemia. More advanced diagnostic tests aside, in electrocardiogram (ECG) analysis it manifests itself as a ST segment deviation, targeted by both exercise ECG and ambulatory ECG. In ambulatory ECG, besides ischaemic ST segment deviation episodes there are also non-ischaemic heart rate related episodes which aggravate real ischaemia detection. We present methods to transform the features developed for the heart rate adjustment of ST segment depression in exercise ECG for use in ambulatory ECG. We use annotations provided by the Long-Term ST Database to plot the ST/HR diagrams and then estimate the overall and maximal slopes of the diagrams in the exercise and recovery phase for each ST segment deviation episode. We also estimate the angle at the extrema of the ST/HR diagrams. Statistical analysis shows that ischaemic ST segment deviation episodes have significantly steeper overall and maximal slopes than heart rate related episodes, which indicates the explored features' utility for distinguishing between the two types of episodes. This makes the proposed features very useful in automated ECG analysis.