Spectrum News 1: Cedars-Sinai Researchers Create New AI-Based System to Read Echocardiograms
Spectrum News 1 recently interviewed cardiologist and researcher David Ouyang, MD, adjunct assistant professor in the Smidt Heart Institute at Cedars-Sinai, about a new artificial intelligence (AI) system that could help physicians more quickly and consistently evaluate heart disease—the leading cause of death in the U.S.
The AI model, called EchoPrime, was developed by Cedars-Sinai researchers in collaboration with other institutions. It quickly and efficiently analyzes images and video from an echocardiogram test, which uses ultrasound to assess the heart’s structure, size and function.
Ouyang said the model was trained on 12 million different videos from Cedars-Sinai and “is mapping what the different videos and images look like and what a clinician would say about the videos and images.”
He told Spectrum News 1 the technology offers significant time savings for healthcare providers: A typical echocardiogram study takes about one hour, while EchoPrime “can save 20 to 30 minutes out of that process.”
Ouyang said the model could also improve diagnostic accuracy and consistency.
“It’s very common that a patient can see two different doctors and come up with two different diagnoses,” he told Spectrum News 1. “By having AI assist with the preliminary assessment, this allows for more reliable and consistent diagnoses.”
Ouyang said this technology represents a significant advancement in cardiac imaging technology.
“It’s really at the intersection, both in cardiology and in AI and in computer vision, that can bring together a lot of tremendous improvements in the field,” he said.
Watch the complete segment from Spectrum News 1.