AI Detects Congestive Heart Failure from Single Heartbeat
By HospiMedica International staff writers Posted on 30 Sep 2019 |
Image: A new study claims that heart failure can now be detected from a single heartbeat (Photo courtesy University of Surrey).
A new study reports successful identification of severe chronic heart failure (CHF) in 100% of cases using a single heartbeat recording from an electrocardiogram (ECG).
Researchers at the University of Surrey (Guildford, United Kingdom), the University of Warwick (Coventry, United Kingdom), and The Organizational Neuroscience Laboratory (OneLab; London, United Kingdom) developed the novel approach, which is based on artificial intelligence (AI). The technique uses a convolutional neural network (CNN) data tree to detect data patterns and structures at extremely high efficiency, which accurately identifies CHF on the basis of one raw ECG heartbeat.
The CNN allows time series sub-sequences that serve as the input data to be visualized, thus allowing it to discriminate between CHF and healthy subjects, which not only allows faster detection, but also helps to understand how certain tissue behavior is related to the signals recorded. The researchers tested the model on ECG datasets comprising 490,505 heartbeats; it achieved 100% CHF detection accuracy, and also identified class-discriminative heartbeat sequences and specific ECG morphological characteristics. The study was published on September 3, 2019, in Biomedical Signal Processing and Control Journal.
“We trained and tested the CNN model on large publicly available ECG datasets featuring subjects with CHF as well as healthy, non-arrhythmic hearts. Our model delivered 100% accuracy; by checking just one heartbeat we are able detect whether or not a person has heart failure,” said study author Sebastiano Massaro, PhD, of University of Surrey. “Our model is also one of the first known to be able to identify the ECG' s morphological features specifically associated to the severity of the condition.”
“We envision the adaptation of this system to wearable devices that may be able to perform prediction and detection of CHF using interim ECG recordings, by looking at individual heartbeat morphology. This could allow not only cardiologists but even patients and their caregivers, nurses, trainees, and GPs to take part in the detection process,” concluded Dr. Massaro. “With the completion of the training phase, the network works very rapidly, making it fit for deployment into cloud systems or adaptation to mobile devices.”
CHF is a progressive pathophysiological condition responsible for chronic loss of pumping power in the heart. According to the European Society of Cardiology (ESC), around 26 million people worldwide are affected. Its prevalence increases quickly with age, and mortality rate is closely associated with the degree of severity, reaching peaks of 40% in the most serious events. CHF is also one of the foremost reasons for hospitalization in the elderly, and it is characterized by a resilient relapse rate, with half of the outpatients readmitted within a few months from hospital discharge.
Related Links:
University of Surrey
University of Warwick
The Organizational Neuroscience Laboratory
Researchers at the University of Surrey (Guildford, United Kingdom), the University of Warwick (Coventry, United Kingdom), and The Organizational Neuroscience Laboratory (OneLab; London, United Kingdom) developed the novel approach, which is based on artificial intelligence (AI). The technique uses a convolutional neural network (CNN) data tree to detect data patterns and structures at extremely high efficiency, which accurately identifies CHF on the basis of one raw ECG heartbeat.
The CNN allows time series sub-sequences that serve as the input data to be visualized, thus allowing it to discriminate between CHF and healthy subjects, which not only allows faster detection, but also helps to understand how certain tissue behavior is related to the signals recorded. The researchers tested the model on ECG datasets comprising 490,505 heartbeats; it achieved 100% CHF detection accuracy, and also identified class-discriminative heartbeat sequences and specific ECG morphological characteristics. The study was published on September 3, 2019, in Biomedical Signal Processing and Control Journal.
“We trained and tested the CNN model on large publicly available ECG datasets featuring subjects with CHF as well as healthy, non-arrhythmic hearts. Our model delivered 100% accuracy; by checking just one heartbeat we are able detect whether or not a person has heart failure,” said study author Sebastiano Massaro, PhD, of University of Surrey. “Our model is also one of the first known to be able to identify the ECG' s morphological features specifically associated to the severity of the condition.”
“We envision the adaptation of this system to wearable devices that may be able to perform prediction and detection of CHF using interim ECG recordings, by looking at individual heartbeat morphology. This could allow not only cardiologists but even patients and their caregivers, nurses, trainees, and GPs to take part in the detection process,” concluded Dr. Massaro. “With the completion of the training phase, the network works very rapidly, making it fit for deployment into cloud systems or adaptation to mobile devices.”
CHF is a progressive pathophysiological condition responsible for chronic loss of pumping power in the heart. According to the European Society of Cardiology (ESC), around 26 million people worldwide are affected. Its prevalence increases quickly with age, and mortality rate is closely associated with the degree of severity, reaching peaks of 40% in the most serious events. CHF is also one of the foremost reasons for hospitalization in the elderly, and it is characterized by a resilient relapse rate, with half of the outpatients readmitted within a few months from hospital discharge.
Related Links:
University of Surrey
University of Warwick
The Organizational Neuroscience Laboratory
Latest Critical Care News
- Stretchable Microneedles to Help In Accurate Tracking of Abnormalities and Identifying Rapid Treatment
- Machine Learning Tool Identifies Rare, Undiagnosed Immune Disorders from Patient EHRs
- On-Skin Wearable Bioelectronic Device Paves Way for Intelligent Implants
- First-Of-Its-Kind Dissolvable Stent to Improve Outcomes for Patients with Severe PAD
- AI Brain-Age Estimation Technology Uses EEG Scans to Screen for Degenerative Diseases
- Wheeze-Counting Wearable Device Monitors Patient's Breathing In Real Time
- Wearable Multiplex Biosensors Could Revolutionize COPD Management
- New Low-Energy Defibrillation Method Controls Cardiac Arrhythmias
- New Machine Learning Models Help Predict Heart Disease Risk in Women
- Deep-Learning Model Predicts Arrhythmia 30 Minutes before Onset
- Breakthrough Technology Combines Detection and Treatment of Nerve-Related Disorders in Single Procedure
- Plasma Irradiation Promotes Faster Bone Healing
- New Device Treats Acute Kidney Injury from Sepsis
- Study Confirms Safety of DCB-Only Strategy for Treating De Novo Left Main Coronary Artery Disease
- Revascularization Improves Quality of Life for Patients with Chronic Limb Threatening Ischemia
- AI-Driven Prediction Models Accurately Predict Critical Care Patient Deterioration