Skip to content
Source: ScienceDaily

🫀UC Berkeley AI Spots Hidden Heart Risk Signal in Routine EKG Scans

Your routine EKG may predict sudden cardiac death

TL;DR

UC Berkeley researchers trained an AI on 440,000 Swedish EKGs and found a warning signal doctors routinely miss. It flagged a group with a 7% annual death rate, and it works on EKG images hospitals already collect. That makes it a cheap candidate for smarter defibrillator decisions.

Here's the punchline: an AI can spot hidden sudden cardiac death risk in EKG scans hospitals already take every day. UC Berkeley researchers published the work in Nature, and it's the kind of result that could change who gets an implanted defibrillator. Why should you care? Because the current screening problem is brutal. Doctors often can't tell who needs a defibrillator before it's too late, and sudden cardiac arrest is not a niche issue. If you build health tech, work on clinical data pipelines, or train models on medical imaging, this is a clean example of finding signal in data that was already sitting there. The numbers: the team trained on over 440,000 EKGs from Sweden, then validated against thousands of patient records from the United States and Taiwan. The model isolated a group with a 7% annual death rate. It also flagged a broader group, most of whom looked low risk under conventional methods. The catch? About two-thirds of patients labeled high risk never need their defibrillator to fire a shock.

UC Berkeley AI Spots Hidden Heart Risk Signal in Routine EKG Scans — ScienceDaily

Key Points

1

The AI was trained on over 440,000 electrocardiograms collected in Sweden, then tested on thousands of additional patient records from the United States and Taiwan.

Keep reading — it's free

Enter your email to keep reading — plus the best of AI & tech, daily. Free, forever.

Also get
or

Already a member? Sign in

healthcare-aiekgcardiologymedical-mlnature-research

Comments

Subscribe to join the conversation...

Be the first to comment