A test invented over 100 years ago can suddenly detect diseases it was never designed to see, because AI learned to read between the lines.

Here's the full story, via Dataconomy.

"Researchers at Imperial College London said an AI system can read an electrocardiogram in under two seconds and detect signs of heart failure and valve disease that clinicians cannot identify from the same trace." The results were presented at the European Society of Cardiology congress in Munich, in a trial covering 67,000 patients in the United States.

Dr Ahmed El-Medany, a British Heart Foundation clinical research fellow at Imperial, "described the system as 'superhuman AI.' He said the signal is present in the recording, but clinicians cannot reliably extract it from the ECG on their own."

"The tool identified up to 81% of heart failure cases and up to 90% of valve disease cases, despite using a test that was not originally designed to detect either condition." Across the wider research program, the team "reported accuracies of 83% to 93% for heart disease and 70% to 80% for the other conditions," including early signals for arrhythmias, diabetes, and kidney disease.

"The researchers said the system is intended to help triage patients rather than replace clinicians. It is designed to identify people more likely to have a structural heart problem so they can be prioritized for ultrasound scans." Dr Sonya Babu-Narayan, a consultant cardiologist at the British Heart Foundation, put it plainly: "Technology like the AI ECG in this research could be a solution to help fast-track patients most likely to have a heart abnormality."

"'Patients can often wait several months for a heart ultrasound scan after being referred for one by their doctor,' said Professor Fu Siong Ng, professor of cardiology at Imperial." An echocardiogram "requires trained staff, specialized equipment and an appointment, while an ECG can be recorded with 10 electrodes and a nurse."

Ng's group "trained its models on 1.6 million ECGs from Brazil linked to patient records, along with several million additional recordings from the United States," using the Brazilian dataset to help the models learn which ECG patterns preceded later diagnoses.

The work is "being commercialized through a spinout called Cardiovolt.ai, with Ng as chief medical officer and Dr Arunashis Sau as chief scientific officer," and the team's next step is handheld ECG devices with the AI built into the workflow.

"The current system has not yet completed the clinical validation and regulatory approval needed for routine use in British hospitals. The trial also did not show whether using the AI in clinical practice leads to earlier treatment or better patient outcomes." A genuinely promising early signal, not yet a finished product.

Source: Dataconomy (Aytun Çelebi): https://dataconomy.com/2026/09/02/ai-ecg-heart-failure-signs-identification/

Want practical AI guidance for parents and educators every week? Subscribe: https://www.aibyage.com/?modal=signup&utm_source=beehiiv&utm_medium=newsletter