Noise-Resilient Automatic Interpretation of Holter ECG Recordings
This work provides an incremental improvement in the accuracy of automatic Holter ECG interpretation, which can reduce the workload for cardiologists and improve diagnostic efficiency.
The paper addresses the challenge of automatically interpreting long-term Holter ECG recordings, which are often noisy and time-consuming for doctors to analyze. The proposed three-stage method, combining neural networks for segmentation and classification with gradient boosting decision trees, outperforms commercial software and existing literature methods on a dataset of 5095 Holter recordings.
Holter monitoring, a long-term ECG recording (24-hours and more), contains a large amount of valuable diagnostic information about the patient. Its interpretation becomes a difficult and time-consuming task for the doctor who analyzes them because every heartbeat needs to be classified, thus requiring highly accurate methods for automatic interpretation. In this paper, we present a three-stage process for analysing Holter recordings with robustness to noisy signal. First stage is a segmentation neural network (NN) with encoderdecoder architecture which detects positions of heartbeats. Second stage is a classification NN which will classify heartbeats as wide or narrow. Third stage in gradient boosting decision trees (GBDT) on top of NN features that incorporates patient-wise features and further increases performance of our approach. As a part of this work we acquired 5095 Holter recordings of patients annotated by an experienced cardiologist. A committee of three cardiologists served as a ground truth annotators for the 291 examples in the test set. We show that the proposed method outperforms the selected baselines, including two commercial-grade software packages and some methods previously published in the literature.