A team of researchers out of Japan has created a new artificial intelligence solution to improve the diagnosis of heart arrhythmic disorders.
The tool relies on deep learning to gain insights from both electrocardiography (ECG) data and X-ray images, Kobe University Hospital physicians explained in Scientific Reports.
In doing so, the AI is able to predict the location of "accessory pathways" inside the heart that can cause irregular heart pulses—a condition known as Wolff-Parkinson-White disorder.
This approach solves a previously unresolved problem, the authors explained, and should allow doctors to offer their patients a more accurate explanation of their condition before determining proper treatment, according to Makoto Nishimori, MD, of Kobe Hospital’s Division of Cardiovascular Medicine and colleagues.
Historically, a conventional 12-lead ECG is used to pinpoint accessory pathway locations before patients undergo catheter ablation. The procedure can completely cure this disorder, but the current ECG method is not 100% accurate which makes it difficult to communicate the plan with patients.
With this in mind, Nishimori and co-authors developed their AI to try and do the job. While each repeated attempt did grow more accurate, the platform was ultimately unable to make consistently correct predictions using only ECG data.
They solved the problem, however, after feeding the tool additional information such as heart size, from chest X-rays, along with traditional ECG output. And with this, the researchers’ algorithm vastly improved.
Given their success, the scientists hope to extend their methodology to other disorders.
Read the entire study published on April 13 here.
Matt joined Chicago’s TriMed team in 2018 covering all areas of health imaging after two years reporting on the hospital field. He holds a bachelor’s in English from UIC, and enjoys a good cup of coffee and an interesting documentary.