Artificial Intelligence Model Detects Heart Tissue Damage via Surface Recordings
Researchers at the Universitat Politècnica de València have developed an artificial intelligence model that detects atrial cardiomyopathy tissue damage using body surface electrical recordings.
Researchers from the COR group at the ITACA institute within the Universitat Politècnica de València have developed an artificial intelligence model that localizes and quantifies tissue abnormalities associated with atrial cardiomyopathy using electrical recordings taken from the body surface.
Developed Using Graph Neural Networks
Based on graph neural network technology, this innovative system achieved an accuracy rate of 89 percent in locating the affected regions and 84 percent in determining the degree of tissue damage.
Success in Structures Not Used During Training
This specially developed artificial intelligence model successfully maintained the same detection and measurement capability even when analyzing anatomical structures that were not used during the training phase.
Next Steps and Scientific Publication
Carried out using simulated data at this stage, this scientific study serves as a proof of concept, and it was stated that the next step is to perform validation on patients.
Led by María Macarulla-Rodríguez, this important research was published in the respected scientific journal Discover Computing.