Atrial fibrillation is a heart disorder that can lead to critical situations such as stroke, and finally death. Artificial intelligence "read" the records of a million EKG records and with its help the researchers were able to discover early signs
Shani Beaton and Shaina Gendelman, graduate students in the Faculty of Biomedical Engineering at the Technion, wrote an algorithm that predicts the chances of an infected person developing atrial fibrillation in the next five years - a heart disorder that can lead to critical results. The research was conducted under the guidance of Dr. Joachim Bahar, Head The laboratory for artificial intelligence in medicine (in square meters).
Atrial fibrillation is a heart rhythm disorder that is not immediately life-threatening but significantly increases the risk of stroke and death. It is now known that certain behaviors such as a sedentary lifestyle, smoking, and obesity increase the risk of atrial fibrillation, so being warned of such a risk may allow a person to take risk-reducing measures and enter into a follow-up routine that will enable early detection of the problem.
The students trained a deep learning system (multiple neural network) using more than a million ECG recordings of more than 400,000 patients, and thus created a mechanism for predicting a person's chances of developing atrial fibrillation within a five-year period. They then combined the deep neural network with information Clinically on the patient, this model was able to correctly predict the risk of developing atrial fibrillation in 60%. of the cases, while maintaining a high specificity rate of 95% (that is, only 5% of the people identified as people at risk did not develop the disease).
According to Dr. Behar, "We do not intend to replace the human doctor, but we would like to give him better tools to support him in making decisions. Advanced computational tools know how to process data more efficiently and accurately than any human, and deep learning enables the identification of patterns that were unknown to us. Throughout history, doctors progressed from manual pulse measurement to a stethoscope and from there to an EKG, and we believe that EKG analysis based on machine learning is another important step that significantly improved the quality of diagnosis and prevention."
According to the researchers, since the ECG is a routine and relatively cheap test, it is possible to integrate the machine learning model into clinical practice and thus improve the management of healthcare services. Access to additional data sets will allow the algorithm to gradually improve as a risk prediction tool. Moreover, the model can be adapted to disease prediction other cardiovascular
The study was conducted in collaboration with Antonio Ribeiro from Uppsala University in Sweden and Gabriela Miana, Carla Moreira and Antonio Luis Ribeiro from the Universidade Federal de Minas Gerais in Brazil. The EKG and electronic medical records of the patients were provided by the Telehealth Network of Minas Gerais, a public telehealth system that assists most local authorities in the state of Minas Gerais, Brazil.
The study was published in European Heart Journal - Digital Health
Comments
It's strange that all the decoding of the EKG Still not done by AI. On the face of it, this seems like a small problem several orders of magnitude from protein folding, for example.
This is the future.