AI system reads heart scans as well as specialist doctors — and generates reports automatically

Researchers built an AI agent that interprets cardiac MRI scans end-to-end: it segments heart structures, measures function, identifies tissue damage, and writes clinical reports. On internal tests it matched specialist radiologists; on external hospital data it remained accurate enough to be clinically useful.

Cardiac MRI is the gold standard for heart disease diagnosis but severely underused because interpretation requires hours of specialized expertise per patient. If this system can reliably handle that work at scale, it removes a hard constraint on how many patients can be diagnosed accurately. The real question is whether hospitals actually deploy it — not because the AI is novel, but because it solves a genuine bottleneck in clinical workflow.