TOMSK, Jul 29 – RIA Tomsk. A group of scientists from Tomsk State University (TSU), Siberian State Medical University (SSMU) and the Institute of Cardiology developed a non-invasive approach to diagnosing myocardial infarction based on the analysis of biomarkers in exhaled air; they also "taught" the computer to recognize these markers, TSU press service told RIA Tomsk.
It was previously reported that scientists of TSU and Cardiology Research Institute several years ago were developing a technique to assess immune inflammation in the heart, which would allow timely detection of patients with complications after myocardial infarction. The researchers planned to use biomarkers to form diagnostic panels that would allow them to identify a risk group among patients and prevent the development of complications.
Detect an infarction by air
According to her, the issue of prompt diagnosis of myocardial infarction is very relevant, since this pathology is one of the three most frequent causes of death. Typically, acute myocardial infarction is detected using electrocardiogram (ECG) and highly specific for acute myocardial infarction (AMI) markers in the blood, but these methods are not always convenient.
"Specialists of the laboratory of laser molecular imaging and machine learning of TSU, SSMU, and Cardiology Research Institute have developed a new noninvasive approach for diagnosing myocardial infarction. It is based on the method of analysis of biomarkers in the exhaled air, previously created in TSU for the diagnosis of serious pathologies", the interlocutor of the agency said.
With reference to the developers, she added that exhaled air contains many molecules, which are products of metabolism, and serve as indicators of certain changes in the human body. For example, in AMI. Myocardial cells experience oxygen starvation accompanied by lipid peroxidation. This chemical reaction is also manifested in the exhaled air.
How the study was conducted
During the work, scientists of Cardiology Research Institute took samples from 30 patients with primary myocardial infarction. As a result, they identified six volatile molecular biomarkers, which were used to train a computer model so that the artificial intelligence learned to identify this disease by biological markers of people, the interlocutor of the agency explained.
She noted that a peculiarity of this work is that two alternative approaches are used at creation of diagnostic models: the first one is based on revealing chemical patterns of exhaled air sample, the second one - on registration of group of pre-detected molecular biomarkers.
The researchers outlined the results of their collaboration in a scientific paper, which was published in the Journal of Breath Research.