Researchers at Queen Mary University of London have created an AI tool that makes realistic heart scarring models to help treat atrial fibrillation (AF). These computer-made heart models can help doctors plan better treatments for people with AF, a common condition that causes irregular heartbeats.
Scarring in the heart (called fibrosis) happens due to aging, stress, or AF itself. These scars make it harder for the heart to beat properly. Doctors usually find this scarring using special MRI scans, but it’s hard to get enough of these scans for AI training.
AF is often treated using ablation, a procedure where doctors create tiny scars to block the signals causing irregular heartbeats. However, this treatment doesn’t always work, and it’s hard to know which method will work best for each patient.
To solve this, the researchers trained an AI tool using 100 real MRI scans from AF patients. The AI then made 100 fake—but realistic—scarring patterns. These were used in digital 3D heart models to test how different treatments might work.
Dr. Alexander Zolotarev, the lead researcher, said their AI tool created heart scar patterns that were very close to real ones. When they tested these in computer models, the results were almost as accurate as using real patient data. This approach also protects patient privacy and allows more testing than usual methods.
Dr. Zolotarev said the goal is not to replace doctors, but to give them a helpful tool. Doctors can use the digital heart models to try different treatments before doing the real procedure. This work is part of a larger project led by Dr. Caroline Roney to create “digital twin” hearts for personalized AF care.
Dr. Roney explained that the AI tool helps overcome the problem of limited patient data. It allows large-scale digital testing and helps build treatments that are more suited to each person.
With around 1.4 million people in the UK affected by AF—and half of ablation treatments not working the first time—this AI tool could reduce the number of repeat treatments. It also helps solve two big problems in healthcare: not having enough patient data and the need to keep medical information private.
Source: news-medical
