AI model developed to detect cardiac fibrotic anomalies

Researchers from the COR group at the Institute of Information and Communication Technologies (ITACA) of the Polytechnic University of Valencia (UPV) have developed an Artificial Intelligence model capable of identifying and quantifying fibrotic anomalies related to atrial cardiomyopathy.

Using Graph Neural Networks (GNNs) to analyze electrical recordings from the body surface, the system achieved 89% accuracy in detecting affected areas and 84% accuracy in determining the extent of damage to atrial tissue. The model also demonstrated the ability to analyze anatomical structures that were not included during its training phase.

Currently, the study serves as a proof of concept developed using simulated data and does not yet have immediate clinical application. The next phase of research will involve validating the model using recordings from actual patients. If clinical trials confirm these results, the technology could facilitate the development of non-invasive tools for characterizing atrial tissue and improving the treatment of related pathologies. The study was published in the journal Discover Computing.