BOSTON (WHDH) - Researchers at Brigham and Women’s Hospital are revolutionizing the way doctors can detect a common heart disease. Atrial septal defect, or ASD, is a congenital heart disease and could be detected with artificial intelligence.
Shinichi Goto, a doctor and researcher, said the symptoms that go along with ASD are typically mild, so the disease often goes undiagnosed. However, if not treated, ASD can cause serious problems.
“You get irreversible complications such as atrial fibrillation, stroke, heart failure, and so on,” Goto said. “It’s missed in most of the cases.”
Goto said the reason ASD is often undiagnosed is because screening for it is expensive. It could also be inaccurate, Goto said.
“The problem of detecting ASD is that the gold standard, or the best way, to detect this atrial septal defect is to use echocardiogram which is a non-invasive modality, but it’s time-consuming and labor-consuming.”
In an effort to make screening for the disease more accessible, researchers at Brigham and Women’s working with doctors at Keio University in Japan have developed a deep-learning artificial intelligence model to screen an electrocardiogram, or ECG, for signs of ASD.
“This AI takes the same input, the ECG, but it could detect ASD with a much higher sensitivity,” Goto said.
According to Goto, the AI model correctly detected ASD 93.7% of the time during its test phase. He said detecting this disease early is crucial. The model is promising and could become a key screening tool, he said.
“Once you close it before the symptoms develop, you can stop the symptoms from developing,” Goto said. “The number is large enough that we believe the AI model is detecting something that is truly useful.”
(Copyright (c) 2024 Sunbeam Television. All Rights Reserved. This material may not be published, broadcast, rewritten, or redistributed.)