Long-term ST database: A reference for the development and evaluation of automated ischaemia detectors and for the study of the dynamics of myocardial ischaemia

Abstract

The long-term ST database is the result of a multinational research effort. The goal was to develop a challenging and realistic research resource for development and evaluation of automated systems to detect transient ST segment changes in electrocardiograms and for supporting basic research into the mechanisms and dynamics of transient myocardial ischaemia. Twenty-four hour ambulatory ECG records were selected from routine clinical practice settings in the USA and Europe, between 1994 and 2000, on the basis of occurrence of ischaemic and non-ischaemic ST segment changes. Human expert annotators used newly developed annotation protocols and a specially developed interactive graphic editor tool (SEMIA) that supported paperless editing of annotations and facilitated international co-operation via the Internet. The database contains 86 two- and three-channel 24 h annotated ambulatory records from 80 patients and is stored on DVD-ROMs. The database annotation files contain ST segment annotations of transient ischaemic (1155) and heart-rate related ST episodes and annotations of non-ischaemic ST segment events related to postural changes and conduction abnormalities. The database is intended to complement the European Society of Cardiology ST-T database and the MIT-BIH and AHA arrhythmia databases. It provides a comprehensive representation of 'real-world' data, with numerous examples of transient ischaemic and non-ischaemic ST segment changes, arrhythmias, conduction abnormalities, axis shifts, noise and artifacts.

Towards the prediction of transient ST changes

“…The set associated with a given class corresponds to models associated with the KLTs found in the 30s of ECG signal before the event. The details of the KLT formation are found in [9,11].…”

“…The free portion of the dataset consists of 43 records from 42 patients (1772 events using the B protocol) and is used as the training set for this study. The fee portion of the dataset consists of 43 recordings from 38 patients (1974 events using the B protocol) and is used as the test set for this study [8,9]. A second partitioning of the dataset is formed for the purposes of ten-fold cross validation.…”

Automatic distinguishing between ischemic and heart-rate related transient ST segment episodes in ambulatory ECG records

“…The Long-Term ST Database (LTST DB) [1] contains 86 2-or 3-lead 24-hour ambulatory ECG records, sampled at 250 samples s −1 per channel, and is intended for development and testing of automatic ischemia detectors. The records were collected during routine clinical practice to model significant number of real-world clinical conditions.…”

Evaluation of a Neyman-Pearson heart-rate turbulence detector

“…The Long Term ST database (LTST) [6] contains 68 twolead and eighteen three-lead 24-hour ambulatory records with significant ST events annotated by medical experts. These records were sampled at 250 Hz with a 12-bit resolution.…”

“…These records were sampled at 250 Hz with a 12-bit resolution. The database [6]also provides automatic QRS annotations including beat classification supervised by an expert Holter technician.…”