METHOD FOR NONINVASIVE DIFFERENTIAL DIAGNOSIS OF DISEASES OF RESPIRATORY SYSTEM AND DEVICE FOR ITS IMPLEMENTATION Russian patent published in 2021 - IPC A61B5/93 G01N33/49 

FIELD: medicine.

SUBSTANCE: group of inventions relates to the field of medicine and can be used for non-invasive differential diagnosis of diseases of the respiratory system. The method includes preliminary sampling of exhaled air from patients and determination of a set of specific volatile markers characteristic of the disease, followed by data processing using a neural network. Sampling in the volume of 250 ml and analysis of exhaled air samples is carried out using the device. Identification of pathophysiological changes in exhaled air is carried out based on the calculation of the output neuron value of a trained artificial neural network above the threshold of separation of positive and negative samples. The probability of the presence of the disease is determined by the Area Under ROC curve (AUC) method. Values for each individual type of pathology are formed by a pre-trained neural network on the corresponding control set of patients without pathology, a number of 20 or more in a 50:50 ratio with the presence of pathology, and determining the type of disease of the respiratory system organs by determining the degree of similarity of the gas pattern of the patient examined during screening with the gas patterns of groups of patients of the training set of the artificial neural network and with values of the output neuron from 0.25 to 1, malignant neoplasms in the lungs are diagnosed, with values of the output neuron from 0.1 to 1, malignant formations of the oropharyngeal region and larynx, with values of 0.65 to 1, coronavirus infection (COVID-19) and community-acquired pneumonia. The device for non-invasive differential diagnosis of diseases of the respiratory system contains detectors for detecting a set of markers characteristic of the state of the object, an artificial neural network for signal processing, a sampling chamber with input and output valves, a measurement chamber, a measuring module, a microcontroller, a relay module, a compressor, an air filter and a personal computer with an artificial neural network. In the measuring module, a set of semiconductor gas sensors with different selectivity and sensitivity to reducing and oxidizing gases is used as detectors.

EFFECT: simplification of the method for noninvasive diagnosis and detection of malignant neoplasms in the lungs, oropharyngeal region and larynx, coronavirus infection (COVID-19), community-acquired pneumonia.

5 cl, 7 dwg, 4 tbl