Heart problem spotted in 2 secs by 'superhuman' AI? How this tool can see deep into your heart valves what doctors may miss

What if a routine heart test could reveal hidden heart disease in just seconds? Researchers at Imperial College London have developed an artificial intelligence system that can analyse a standard electrocardiogram (ECG) and identify signs of conditions such as heart failure and heart valve disease that may not be obvious to a doctor looking at the tracing alone.

The technology has been described as “superhuman” because it can detect patterns in ECG data that even expert cardiologists cannot identify simply by examining the heart tracing. According to The Guardian, the system can analyse an ECG in less than two seconds, potentially helping doctors identify patients who need further tests more quickly.

Read more - Barefoot mom in sari runs 3 km to win Rs 5,000 for daughters' education after sleepless night over daughter's food and rent

How does the AI heart disease detector work?

An ECG is a routine test that records the heart’s electrical activity. The test itself takes around 10 seconds and is commonly used to check heart rate and rhythm.

However, an ECG can contain information beyond what doctors traditionally use it for. The Imperial team trained AI models on millions of ECG recordings linked to patients’ medical records, allowing the system to learn subtle electrical patterns associated with different diseases.

Researchers used more than 1.6 million ECGs from Brazil, alongside large datasets from the US, to train and test the models.

The AI effectively searches for tiny digital signals hidden within the ECG that could indicate an underlying health problem.

Why is the AI being called ‘superhuman’?

The term does not mean that AI is replacing cardiologists. Instead, researchers use “superhuman” to describe tasks that cannot be performed by even highly experienced doctors simply by looking at an ECG.

“We looked at the ECGs to see if we could do things that are superhuman. Not things that clinicians can already do, but things that no cardiologist, no matter how expert, can do,” Dr Arunashis Sau of Cardiovolt.ai and Imperial’s National Heart and Lung Institute said.

The AI models achieved diagnostic accuracy of between 83% and 93% for different cardiovascular conditions across international datasets.

AI can detect hidden heart failure and valve disease

One of the most promising uses is identifying heart conditions that may otherwise go unnoticed.

In a US trial involving around 67,000 people, the system identified heart failure with up to 81% accuracy and heart valve disease with up to 90% accuracy, according to The Guardian.

The technology is not intended to replace an echocardiogram or provide a final diagnosis on its own. Instead, an abnormal AI reading could help doctors decide which patients should undergo more detailed testing.

Professor Fu Siong Ng, who led the research, said the aim is to identify underlying heart failure or valve disease that might otherwise be missed and then confirm suspected cases with an echocardiogram.

Could AI make heart disease diagnosis faster?

This is where the technology could have a major practical impact. Patients suspected of having heart disease may have to wait for additional investigations, including echocardiograms. An AI system that can screen an ECG almost immediately could help clinicians identify people who appear to be at higher risk and prioritise them for further tests.

In other words, the technology would work as a triage tool rather than a replacement for doctors.

That could be particularly useful in healthcare systems dealing with long waiting lists and large numbers of patients requiring cardiac investigations.

AI ECG could detect more than heart disease

The Imperial researchers say the information hidden in an ECG may extend beyond cardiovascular conditions.

Their models have also shown the ability to identify patterns associated with conditions such as diabetes and kidney disease. Imperial says the models achieved 70-80% accuracy for some non-cardiovascular conditions using a single 10-second ECG.

Researchers are investigating what these digital signals represent by comparing ECG information with imaging, genetic and protein data from resources such as the UK Biobank.

Cardiovolt.ai aims to bring the AI into hospitals

The research is now moving towards real-world medical use through Cardiovolt.ai, a spinout from Imperial College London.

The company was created to commercialise the AI models developed by the research team and has raised £1.4 million in funding. Its immediate focus is securing regulatory approval in the UK, EU and US.

The longer-term vision is to make AI analysis part of routine ECG testing in hospitals. For now, however, the technology remains a tool to support clinical decision-making rather than a standalone diagnostic system.

Could a 10-second ECG become a powerful health screening tool?

The humble ECG has traditionally been seen as a relatively simple test, mainly useful for checking the heart's rhythm and electrical activity.

AI could change that. By analysing patterns too subtle for humans to recognise, researchers believe a standard ECG could eventually become a much more informative early-warning tool, helping doctors identify hidden disease, prioritise further testing and potentially intervene earlier.

As Professor Ng put it, a 10-second ECG that was once considered a blunt screening tool could eventually become one of the most informative tests in medicine.


Inputs from agencies