AI tool detects heart failure from a single heartbeat

Researchers develop a neural diagnostic tool which can detect cardiac anomalies with 100% accuracy

Artificial intelligence (AI) has now taken another step forward in healthcare, as a new study reveals a neural network tool that can identify heart failure from just a single heartbeat.
A new model called Convolutional Neural Networks (CNN) was designed by researchers at Universities of Surrey, Warwick, and Florence that was found to detect heart failure instantly with 100% accuracy.

Currently, most practitioners use electrocardiograms (ECG) to determine heart conditions. However, this method can take several days of analysis and testing to determine a diagnosis. A new technology may just be the answer to making quicker and more accurate cardiac diagnostics.

Longevity.Technology: AI focused diagnostic tools are a growing trend with an increased demand for better efficiency and accuracy. This new technology paves the way for life-saving diagnostic technology that could cut down costs and provide a quicker and easier way to extend healthspan. 

The TRL score for this Longevity.Technology domain is currently set at: ‘Technology completes secondary trials and provides further evidence for safety and efficacy.
TRL7 orange

The TRL score for the technology addressed in this article is: ‘Late proof of concept demonstrated in real life conditions.

Congestive Heart Failure (CHF) is a condition affecting how well the heart can pump blood around the body and is most common in adults over the age of 65. According to the European Society of Cardiology, the condition affects 26 million people worldwide. CHF was responsible for $31 billion in healthcare expenditures in the US in 2012, with the costs predicted to rise by 127% by 2030 [2].

The new model is based on an AI approach, which makes use of hierarchical neural networks that are highly effective at recognising complex data patterns. The CNN combines the use of advanced signal processing and machine learning on the raw ECG data to improve detection results.

“We trained and tested the CNN model on large publicly available ECG datasets featuring subjects with CHF as well as healthy, non-arrhythmic hearts,” said Dr Sebastiano Massaro, the associate professor of organizational neuroscience at the University of Surrey.

The researchers noted that although their results have greater accuracy than current methods, they might not be as precise for patients with milder forms of CHF. The data in the study consisted of ECG readings from healthy patients or those with severe CHF, this means that although the research is promising, a broader spectrum of CHF diagnosis needs to be researched before this new technology can be used in clinical practices.

Yet, despite the need for further research, this new technology is another addition to more AI-driven Longevity tools that can improve healthspan and lifespan, as well as attracting significant investment. AI focused healthcare is a growing market that is expected to grow to $36 billion by 2025 [4].

However, this CNN model is not the first, there have been several predecessors that have not been as successful. Massaro and his team created the new model to be less complex and more suitable for mobile applications. The vision is ultimately to adapt the technology to a wearable or mobile device which could monitor patients in real-world conditions. This would allow the patients themselves, their GPs or caregivers to detect CHF and relieve some of the financial burden on healthcare businesses.