Researchers at Cleveland Clinic have developed an artificial intelligence system that can accurately interpret cardiac MRI scans — and the key innovation is that it doesn't rely on images alone.
According to Cleveland Clinic, the system works by combining the visual data from cardiac MRI images with written clinical impressions — the notes and observations that physicians record alongside scans. By fusing these two types of information, the AI is able to read heart imaging with a level of accuracy that neither source could achieve on its own.
Cardiac MRI is one of the most detailed tools doctors have for assessing heart structure and function, but interpreting the scans is time-consuming and requires specialized expertise that isn't always readily available. Radiologists and cardiologists must review both what they see in the image and the broader clinical context to reach a reliable conclusion — a task the Cleveland Clinic system now attempts to replicate.
The approach reflects a broader trend in medical AI: moving beyond single-modality models that process only images or only text, toward systems that integrate multiple streams of information the way a human clinician would.
If validated at scale, a tool like this could help hospitals manage growing imaging workloads, reduce delays in diagnosis, and extend specialist-level cardiac care to facilities that lack dedicated heart imaging experts — potentially reaching more patients sooner.