Artificial intelligence tools for phenotyping patients with anaphylaxis
Nikitina E.A. Dushkin A.D. Lebedkina M.S. Mukhina O.A. Kruglova T.S. Andreev S.S. Nurtazina A.Yu. Leonova E.S. Volkova P.A. Pampura A.N. Karaulov A.V. Lysenko M.A. Fomina D.S.
2025Pharmarus Print Media
Russian Journal of Allergy
2025#22Issue 11 - 10 pp.
BACKGROUND: Anaphylaxis is the most severe manifestation of immediate systemic hypersensitivity reactions. In recent years, there has been an increase in the number of reported cases of anaphylaxis. Given the variety of clinical manifestations behind the diagnosis of “anaphylaxis”, one of the most urgent issues at present is the task of phenotyping and endotyping of patients with this life-threatening pathology as part of implementation of the individual approach, both during treatment in the acute period and diagnosis in the period of recovery. AIM: To find the main clinical phenotypes of patients with immediate hypersensitivity reactions, which will further allow stratifying patients by risk groups bringing therapeutic and diagnostic algorithms to the modern level of personalized medicine. MATERIALS AND METHODS: The study was conducted based on retrospective stepwise analysis of patient medical record data from 2019 to 2022. To identify relatively homogeneous groups of patients based on clinical, agglomerative clustering was performed on 56 variables followed by 2 principal component extraction using the dimensionality reduction method with t-distributed stochastic neighbor embedding. Agglomerative clustering divided patients into 4 major clinical phenotypes, and each patient was assigned a corresponding phenotype. RESULTS: Based on this analysis, 4 phenotypes of patients with severe immediate-type hypersensitivity reactions were identified. CONCLUSION: A new classification based on the use of phenotypes, endotypes and biomarkers is currently being developed to broaden our understanding of anaphylactic reactions. Given the limitations of the study (patients were not subjected to additional examinations in the current study), it is not possible to reliably identify endophenotypic differences in different clusters. Additional studies are needed to identify the correlation between anaphylactic reaction phenotypes and its pathophysiological mechanisms of development.
anaphylactic reaction , anaphylaxis , artificial intelligence , Daria agglomerative clustering
Text of the article Перейти на текст статьи
Moscow City Hospital 52, Moscow, Russian Federation
The First Sechenov Moscow State Medical University (Sechenov University), Moscow, Russian Federation
National Medical Research Center for High Medical Technologies, Central Military Clinical Hospital named after A.A. Vishnevsky, Moscow, Russian Federation
Morozov Children’s City Clinical Hospital, Moscow, Russian Federation
N.I. Pirogov Russian National Research Medical University, Moscow, Russian Federation
LIFT Center LLC, Moscow, Russian Federation
Astana Medical University, Astana, Kazakhstan
Moscow City Hospital 52
The First Sechenov Moscow State Medical University (Sechenov University)
National Medical Research Center for High Medical Technologies
Morozov Children’s City Clinical Hospital
N.I. Pirogov Russian National Research Medical University
LIFT Center LLC
Astana Medical University
10 лет помогаем публиковать статьи Международный издатель
Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026