How AI Could Help Doctors Read Heart MRIs Faster

CLEVELAND - Reading cardiac MRI images can be time consuming, but a new AI program could help speed up the process.

“If you can imagine, the heart is this moving object that's in three dimensions. What MRI does is it cuts it in 2D slices, so you end up with more than a thousand images during your whole scan. And just think of a human having to read through a thousand images and look for potentially very, very small disease, maybe appearing in only like two or three images at a time. That's very time-consuming. And just the training needed to do that accurately and efficiently, it just takes a really long time,” explained David Chen, PhD, one of the researchers involved with developing the AI program for Cleveland Clinic.

Dr. Chen said the technology works by analyzing all of the heart images at once and then compiles a report within a matter of seconds.

It’s also trained to look for a variety of heart conditions unlike other AI programs that only focus on a certain type.

He said their goal isn’t just to speed up the time it takes to read cardiac MRI images but to also provide more accessibility.

“AI, at least to me, is not something that replaces human beings. It's to make people more efficient and also bring them back to the practice of medicine, bring them back to getting you in front of people, bring it back to really taking care of the patient,” he noted.

Dr. Chen said the technology is still in the research phase and more work needs to be done before it can be used on real world patients.