THE MILAD STASH

Heart of the matter: Berkeley AI spots hidden EKG clue to sudden cardiac death

SUMMARY

UC Berkeley researchers used artificial intelligence to find a previously overlooked warning signal in routine heart scans that may predict sudden cardiac death. The AI was trained on more than 440,000 electrocardiograms from Sweden, matched against death certificate data. It was then tested on thousands of records from the United States and Taiwan. The team reports it identified vulnerable patients better than conventional screening, which largely measures how much blood the heart pumps. More than 300,000 Americans die of sudden cardiac arrest each year.

Why it matters: Sudden cardiac arrest kills over 300,000 Americans yearly, and AI reading routine EKGs may identify at-risk patients better than current pumping-based screening, potentially improving prevention, though the findings still need wider validation.

sciencedaily.com · Nov 15, 2026, 7:46 AM PT
Read it at sciencedaily.com

Comments

No comments yet.

Comments show once Milad approves them.

Back to the front page