MICROBIOME-ASSOCIATED METABOLOMIC IN THE CLINICAL LABORATORY DIAGNOSIS OF TONSILLITIS

The features of the functional activity of oropharyngeal microbiocenosis in tonsillitis were studied by the method of microbiome-associated metabolomic. A model of predictive diagnostics «Tonsillitis» was built, it was validated, and after combining the training and validation samples, the model «Tonsillitis 1.0» was built. It has been established that the concentrations of anaerobic microorganisms-propionic and butyric acid are included in the discriminant equation of the Tonsillitis model, and the concentration of valeric acid is also included in the Tonsillitis 1.0 model. The Tonsillitis model was built by the results of biochemical analysis of the oropharyngeal microflora of 40 patients, describes the unique ratio of the concentrations of propionic and butyric acids in saliva (the metabolic imprint of tonsillitis), expressed as diagnostic coefficients (linear equations), which allow diagnosing tonsillitis with 95% predictive accuracy. The predictive accuracy for the validation sample (30 people) is 100%. The predictive accuracy of the Tonsillitis 1.0 model (70 people) is 97.1%. An increase in the predictive accuracy of the model for predictive diagnosis of tonsillitis occurs due to an increase in sensitivity from 90% in the Tonsillitis model to 95% in the Tonsillitis 1.0 model. An increase in the number of groups contributes to an increase in the predictive accuracy of the model for predictive diagnosis of tonsillitis, increases the statistical significance of differences between the Tonsillitis groups and the comparison group. The model of predictive diagnosis «Tonsillitis 1.0» can be used in clinical laboratory diagnosis of tonsillitis by metabolic biomarkers in saliva using microbiome-associated metabolomic. © 2023, Klinichescheskaya Laboratornaya Diagnostika. All rights reserved.