Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networks

By Pranav Rajpurkar, Awni Y. Hannun, Masoumeh Haghpanahi, Codie Bourn, Andrew Y. Ng. arXiv.org, 2017

873 citations recorded by the source corpus.

We develop an algorithm which exceeds the performance of board certified cardiologists in detecting a wide range of heart arrhythmias from electrocardiograms recorded with a single-lead wearable monitor. We build a dataset with more than 500 times the number of unique patients than previously studied corpora. On this dataset, we train a 34-layer convolutional neural network which maps a sequence of ECG samples to a sequence of rhythm classes. Committees of board-certified cardiologists annotate a gold standard test set on which we compare the performance of our model to that of 6 other individual cardiologists. We exceed the average cardiologist performance in both recall (sensitivity) and precision (positive predictive value).

Subfields: Artificial Intelligence, Cardiology and Cardiovascular Medicine, Cognitive Neuroscience

View Paper: https://www.semanticscholar.org/paper/b96de5f21c0449b23e7e256affc794f4f0404a5d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cardiology and Cardiovascular Medicine

In Cognitive Neuroscience

Citations in the corpus

Clinical applications of machine learning in cardiovascular disease and its relevance to cardiac imaging. (2018, methodology): “41 Echocardiography Two-dimensional echocardiography (2DE) is ubiquitously used to guide the diagnosis and down-stream management of numerous cardiac pathologies.”
BeatGAN: Anomalous Rhythm Detection using Adversarially Generated Time Series (2019, methodology): “We use CNNs because many studies show that CNN networks are more robust than LSTMs for time series [Rajpurkar et al., 2017].”
LSTM-Based ECG Classification for Continuous Monitoring on Personal Wearable Devices (2018, background): “Previous works have lower classification performance [7]–[19], are not suitable for continuous execution on wearable devices due to high computational intensity [20]–[23], do not include all the standard AAMI classes [26]–[31], or focus on other problems related to processing of ECG signals [32]–[42”
CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning (2017, background): “, 2017), arrhythmia detection (Rajpurkar et al., 2017), and hemorrhage identification (Grewal et al.”
EEG-Based Spatio–Temporal Convolutional Neural Network for Driver Fatigue Evaluation (2019, positive): “14-layer network, adopting the CNN architecture from electrocardiogram arrhythmias detection in [36].”
ECG Heartbeat Classification: A Deep Transferable Representation (2018, positive): “This approach provides us with a more accurate representation of ECG signal using which the machine can compete with a human cardiologist in analyzing the signal [11].”
COVID-19 Image Data Collection (2020, methodology): “Eventually, our model could take inspiration from work by Rajpurkar et al. (2017), which could predict pneumonia, as well as Cohen et al. (2019), which deployed such models.”
Computer-aided detection in chest radiography based on artificial intelligence: a survey (2018, background): “including pneumonia diagnosis [13], diabetic retinopathy detection [14], skin cancer classification [15], arrhythmia detection [16], and bleeding identification [17].”
Dropout Feature Ranking for Deep Learning Models (2017, background): “Deep neural networks (DNNs) have started to come out as top performers in biology and healthcare including genomics [1], medical imaging [2], EEG[3] and EHR [4].”
An Efficient Deep Learning Approach to Pneumonia Classification in Healthcare (2019, background): “Latest improvements in deep learning models and the availability of huge datasets have assisted algorithms to outperform medical personnel in numerous medical imaging tasks such as skin cancer classification [11], hemorrhage identification [12], arrhythmia detection [13], and diabetic retinopathy de”
Deep Learning in Pharmacogenomics: From Gene Regulation to Patient Stratification (2018, background): “developed a 34-layer convolutional neural network which exceeded the performance of board certified cardiologists in detecting a wide range of heart arrhythmias from electrocardiograms recorded with a single-lead wearable monitor [101].”
Artificial intelligence in retina. (2018, methodology): “Recently, a CNN trained to detect pneumonia-like features or abnormalities on chest and musculoskeletal radiographs performed on par with practicing radiologists, as reported in a preprint (Rajpurkar et al., 2017b).”
ECG Arrhythmia Classification Using STFT-Based Spectrogram and Convolutional Neural Network (2019, background): “This problem creates a gap between the dataset size and deep features, since the datasets that are publicly available in this domain lack in volume [30].”