17 Starshinova A. Sabirova, A. Kudryavtsev I. Rubinstein A. Churilov L.P. Belyaeva E. Anastasia K. Sharipov, R.A. Tukfatullin, R.K. Nikolenko N. Dovgalyuk I. Kudlay D. tuberculosis infection risk groups early diagnosis latent tuberculosis infection immunological tests T-cell response machine learning Scopus Web of Science Белый список 2026 2026-06-01 10.3389/fmed.2025.1710960 FRONTIERS IN MEDICINE Introduction: Tuberculosis infection remains one of the most dangerous and difficult to diagnose diseases. To date, issues related to the early diagnosis of tuberculosis remain unresolved, which is particularly important for its detection in high-risk groups. The detection of latent tuberculosis infection (LTBI) is necessary to control the spread of tuberculosis infection. The diagnosis of LTBI is indirect and based on the detection of an immune response to mycobacterial antigens. Currently, LTBI diagnosis is recommended in highrisk groups. However, diagnosis is difficult and not always straightforward with the use of various immunological tests. The aim of this study is to conduct a systematic review of scientific publications focused on the application of immunological tests and machine learning technologies for the early detection of latent tuberculosis infection in high-risk populations. Material and Methods: We analyzed articles for the period from 2015 to 2025, published in international databases (Medline, PubMed, Scopus). The keywords we used were “tuberculosis infection,” “risk groups,” “early diagnosis,” “latent tuberculosis infection,” “immunological tests,” “T-cell response,” and “machine learning.” The narrative review was carried out in accordance with the PRISMA protocol (http://www.prisma-statement.org). Results: A descriptive research method was used to compile the review, followed by systematization of the information and formulation of the main conclusions. The data obtained allow us to assert that the use of a comprehensive approach in the diagnosis of LTBI, namely the simultaneous use of several immunological tests in combination with laboratory and instrumental research methods in the same individuals, can be considered justified. Conclusion: The creation of a strategy for detecting LTBI in individuals from risk groups can facilitate the detection of infection and play an important role in preventing the development of tuberculosis. The possibility of using machine learning and artificial intelligence will allow the risk of developing active tuberculosis to be determined based on the use of immunological tests. https://repo.bashgmu.ru/publication/5451 https://repo.bashgmu.ru/files/5645