"Superhuman" AI Spots Heart Disease in Under 2 Seconds

Doctors have developed an AI tool capable of spotting signs of heart disease in less than two seconds from a routine electrocardiogram, extracting diagnostic signals embedded in the heart's electrical readings that are completely invisible to the human eye, according to The Guardian's reporting on the breakthrough, presented to delegates at the European Society of Cardiology's annual congress in Munich.

The traditional ECG has been a vital diagnostic tool for over a century, recording the heart's electrical activity to detect heart attacks and abnormal rhythms. But it has never been able to detect heart disease itself, that has historically required an echocardiogram, an ultrasound scan that often carries months-long waiting lists, according to NewsBeep's reporting on the presentation. The new AI model changes that equation entirely by pulling additional diagnostic information out of the same routine ECG data already being collected.

Dr. Arunashis Sau, chief scientific officer of Cardiovolt.ai and a cardiology registrar at Imperial College Healthcare NHS Trust, described the team's specific ambition for the project: "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," according to the British Heart Foundation's reporting on the research. The models were trained on a data bank of more than 1.6 million ECGs from Brazil, each linked to a patient's full medical history, plus several million additional ECGs from the United States.

The AI Model's Diagnostic Performance

One of the most significant applications researchers highlighted isn't just faster diagnosis for suspected cases, it's catching disease in patients who had no reason to expect it. Dr. Ng, one of the researchers involved, explained the opportunistic screening potential directly: "Another potential application of this AI model is to opportunistically diagnose heart failure and heart valve disease in whom these conditions are not suspected. The AI model could be run on all ECGs done in a hospital to flag those at highest risk of these diseases, so that they can be diagnosed earlier," according to NewsBeep's reporting on his comments.

Dr. Ahmed El-Medany, a British Heart Foundation clinical research fellow who led the Imperial College London analysis, described the tool as a "superhuman AI" and identified the next major milestone: designing handheld AI-powered ECG readers healthcare professionals could use directly in the field, rather than relying solely on hospital-based equipment.

This breakthrough isn't happening in isolation. It joins a genuinely active pipeline of AI cardiac diagnostic tools moving toward clinical use, including a separate AI tool now rolling out across the NHS that analyzes heart MRI scans in just 20 seconds, compared to the 13 minutes or more a doctor would typically need, while detecting structural changes with 40% greater precision, according to earlier reporting from University College London Hospitals. Researchers in Munich also presented findings on AI-based analysis of facial videos capable of rapidly detecting undiagnosed high blood pressure and type 2 diabetes, according to NewsBeep's broader coverage of the congress.

This breakthrough is worth understanding for any business in healthcare technology, medical devices, or health insurance, since AI diagnostic tools that extract more value from existing, already-collected data represent a genuinely different and more scalable cost model than tools requiring entirely new equipment or testing procedures. A tool that works on data hospitals are already gathering has a meaningfully lower barrier to widespread adoption than technology requiring new infrastructure investment.

For businesses in health insurance and risk assessment specifically, AI models capable of opportunistically flagging undiagnosed disease from routine tests already performed for unrelated reasons could meaningfully shift how early intervention and risk stratification get approached across entire patient populations, not just those already flagged as high-risk.

What can this new AI tool detect from an ECG?
The AI tool detects signs of heart failure and heart valve disease, two of the most common forms of heart disease, extracting diagnostic signals embedded in ECG electrical readings that are invisible to the human eye, with results available in under two seconds.

How accurate is the AI heart disease detection tool?
The model achieved diagnostic accuracy between 83% and 93% for heart disease and 70% to 80% for non-cardiovascular conditions, based on testing using a data bank of more than 1.6 million ECGs from Brazil plus millions more from the United States.

Is this AI heart disease tool available for patients yet?
Not widely yet. Researchers presented the findings at the European Society of Cardiology congress, with the next planned step being the development of handheld AI-powered ECG readers for healthcare professionals to use directly.

Researchers unveiled a "superhuman" AI tool capable of detecting heart failure and valve disease from a routine ECG in under two seconds, extracting diagnostic signals invisible to the human eye, presented at the European Society of Cardiology's annual congress in Munich. The model, trained on more than 1.6 million ECGs from Brazil and millions more from the US, achieved 83% to 93% diagnostic accuracy for heart disease. Researchers say the tool could eventually be run on all ECGs performed in a hospital to opportunistically flag undiagnosed disease in patients tested for unrelated reasons, with handheld AI-powered ECG readers identified as the next development milestone.