Technion-Israel Institute of Technology researchers have developed a way to detect heart problems based on routine electrocardiography (ECG).

They have also built a database and an analytic tool to provide an automated diagnosis of eight common types of heart disease, quickly and with unprecedented accuracy.

ECG is the most common test of heart function. The noninvasive exam measures the electrical signals generated by the heart muscle tissue through electrodes placed in 12 locations on the skin.

However, ECG results must be interpreted by a cardiologist, making analysis subjective. Even with the introduction of artificial intelligence (AI), analysis is done one printout at a time and is not highly accurate.

The researchers say their new technology “demonstrates unprecedented accuracy in the interpretation of numerous ECG results and delivers a diagnosis of different cardiac disorders simultaneously.”

The analysis is based augmented neural networks, a form of AI that learns patterns by being trained on numerous samples – in this case, some 40,000 standard ECG recordings from more than 6,800 patients in 11 hospitals – that were reviewed, interpreted and classified by cardiologists in categories of normal heart function and eight clinical disorders.

The new system demonstrated 96 percent average accuracy in diagnosing the various conditions, compared to 80% for currently available algorithms.

Ventricular fibrillation, for example, was diagnosed with 98% accuracy compared to 92% for currently available algorithms. Left bundle branch block (LBBB) – blockage of electrical impulses to the heart’s left ventricle – was diagnosed with 100% accuracy versus 85%.

Doctoral student Vadim Gliner. Photo courtesy of Technion Spokesperson’s Office

Published in Nature Research’s Scientific Reports, the research project was headed by Prof. Yael Yaniv, director of the Bioelectric and Bio-energetic Systems Laboratory in the Faculty of Biomedical Engineering at the Technion; Prof. Assaf Schuster, co-chair of the Technion Center for Machine Learning and Intelligent Systems; and doctoral student Vadim Gliner of the Taub Faculty of Computer Science.

Vladimir Markov, head of the Laboratory for Systems Programming at Novgorod University and Arutyun Avetisyan, director of the Institute of Systems Programming at the Russian Academy of Sciences in Moscow partnered in the study.

The researchers emphasize that their system would not replace the physician, but rather provide a “second opinion” likely to detect problems that a human review might miss.