medalphabetМедицинский алфавитMedical alphabet2078-56312949-2807ООО «Альфмед»10.33667/2078-5631-2026-3-110-115medalphabet-4963Research ArticleСтатьиРазработка алгоритма оценки биологического возраста на базе системы объяснимого искусственного интеллекта с использованием лабораторных биомаркеров и показателей биоимпедансаDevelopment of an algorithm for assessing biological age based on an explainable artificial intelligence system using laboratory biomarkers and bioimpedance indicatorshttps://orcid.org/0009-0001-7183-8965БеланК. С.BelanK. S.

Белан Кирилл Сергеевич, руководитель направления превентивной медицины

Москва

Belan Kirill S., head of Preventive Medicine 

Moscow 

kirbelan@gmail.comЛембергК. А.LembergK. A.

Лемберг Ксения Александровна, генеральный директор 

Москва

Lemberg Ksenia A., general director 

Moscow 

IR_FATHUTDINOV@protek.ruФатхутдиновИ. Р.FatkhutdinovI. R.

Москва

Moscow 

АнтоновК. К.AntonovK. K.

Антонов Константин Константинович, к.м.н. заведующий клиниками Атлас 

Москва

Antonov Konstantin K., PhD Med Sci, head of Atlas Clinics 

Moscow 

antonov.kk@atlasclinic.ruhttps://orcid.org/0009-0004-6695-4509ТихоноваО. Е.TikhonovaO. E.

Тихонова Олеся Евгеньевна, системный аналитик 

Москва

Tikhonova Olesya E., systems analys 

Moscow 

lesya.pashkovski@gmail.comhttps://orcid.org/0000-0003-3437-5233ГрязеваН. В.GryazevaN. V.

Грязева Наталья Владимировна, д.м.н., профессор кафедры дерматовенерологии и косметологии

Москва

Gryazeva Natalya V., Dr Med Sci (habil.), professor at Dept of Dermatovenereology and Cosmetology

Moscow 

tynrik@yandex.ruООО «Медицинский центр Атлас»РоссияAtlas Medical Center LLCRussian FederationФГБУ ДПО «Центральная государственная медицинская академия» Управления делами Президента Российской ФедерацииРоссияCentral State Medical Academy of the Administrative Department of the President of Russian FederationRussian Federation20261404202603Дерматология (1)110115Copyright © Белан К.С., Лемберг К.А., Фатхутдинов И.Р., Антонов К.К., Тихонова О.Е., Грязева Н.В., 20262026Белан К.С., Лемберг К.А., Фатхутдинов И.Р., Антонов К.К., Тихонова О.Е., Грязева Н.В.Belan K.S., Lemberg K.A., Fatkhutdinov I.R., Antonov K.K., Tikhonova O.E., Gryazeva N.V.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://www.med-alphabet.com/jour/article/view/4963

По мере динамичного развития исследований в области медицины долголетия биологический возраст все чаще рассматривается как наиболее перспективный интегральный показатель, объединяющий сигнатуры биомаркеров и обладающий клинической значимостью. В данной работе описан процесс разработки системы искусственного интеллекта для оценки биологического возраста на основе мультимодальных биомедицинских данных. Исследованы восемь категорий биомаркеров: биоимпедансометрия, общий и биохимический анализ крови, коагулограмма, липидный профиль, эндокринный статус, маркеры воспаления и простатический специфический антиген. Для более точной интерпретации применены глубокие нейросетевые архитектуры (MLP, DANet, FT-Transformer, GANDALF) и методы объяснимого ИИ (SHAP). Агрегированная модель, использующая несколько категорий показателей, достигла средней абсолютной ошибки 6.7 года и коэффициента корреляции Пирсона 0,83 с хронологическим возрастом. Система обеспечивает интерпретацию вклада отдельных биомаркеров в возрастную акселерацию и стратификацию пациентов по рискам, что открывает возможности для персонализированной медицины и превентивных стратегий. Результаты валидированы на российской популяции.

As research in longevity medicine rapidly develops, biological age is increasingly viewed as the most promising integrated indicator combining biomarker signatures and possessing clinical significance. This paper describes the development of an artificial intelligence system for estimating biological age based on multimodal biomedical data. Eight categories of biomarkers were analyzed: bioimpedance analysis, complete blood count and biochemistry, coagulation profile, lipid profile, endocrine status, inflammatory markers, and prostate-specific antigen. Deep neural network architectures (MLP, DANet, FT-Transformer, GANDALF) and explainable AI methods (SHAP) were applied to achieve a more accurate interpretation. The aggregated model, using multiple categories of indicators, achieved a mean absolute error of 6.7 years and a Pearson correlation coefficient of 0,83 with chronological age. The system enables the interpretation of the contribution of individual biomarkers to agerelated acceleration and patient risk stratification, opening up opportunities for personalized medicine and preventive strategies. The results have been validated in the Russian population.

старениебиомаркеры старениябиологический возрастискусственный интеллектздоровое долголетиеagingaging biomarkersbiological ageartificial intelligencehealthy longevityReferencesLevine M.E. et al. An epigenetic biomarker of aging for lifespan and healthspan // Aging (Albany NY). 2018. Vol. 10, № 4. P. 573–591.Levine M.E. et al. An epigenetic biomarker of aging for lifespan and healthspan // Aging (Albany NY). 2018. Vol. 10, № 4. P. 573–591.Drewelies J. et al. Using blood test parameters to define biological age among older adults: association with morbidity and mortality independent of chronological age validated in two separate birth cohorts // GeroScience. 2022. Vol. 44, № 6. P. 2685–2699.Drewelies J. et al. Using blood test parameters to define biological age among older adults: association with morbidity and mortality independent of chronological age validated in two separate birth cohorts // GeroScience. 2022. Vol. 44, № 6. P. 2685–2699.Mamoshina P. et al. Population Specific Biomarkers of Human Aging: A Big Data Study Using South Korean, Canadian, and Eastern European Patient Populations // J Gerontol A Biol Sci Med Sci. 2018. Vol. 73, № 11. P. 1482–1490.Mamoshina P. et al. Population Specific Biomarkers of Human Aging: A Big Data Study Using South Korean, Canadian, and Eastern European Patient Populations // J Gerontol A Biol Sci Med Sci. 2018. Vol. 73, № 11. P. 1482–1490.Bortz J. et al. Biological age estimation using circulating blood biomarkers // Commun Biol. 2023. Vol. 6. P. 1089.Bortz J. et al. Biological age estimation using circulating blood biomarkers // Commun Biol. 2023. Vol. 6. P. 1089.Jia Q. et al. A biological age model based on physical examination data to predict mortality in a Chinese population // iScience. 2024. Vol. 27, № 3. P. 108891.Jia Q. et al. A biological age model based on physical examination data to predict mortality in a Chinese population // iScience. 2024. Vol. 27, № 3. P. 108891.Putin E. et al. Deep biomarkers of human aging: Application of deep neural networks to biomarker development // Aging (Albany NY). 2016. Vol. 8, № 5. P. 1021–1030.Putin E. et al. Deep biomarkers of human aging: Application of deep neural networks to biomarker development // Aging (Albany NY). 2016. Vol. 8, № 5. P. 1021–1030.Li X. et al. Application of biological age assessment of Chinese population in potential anti-ageing technology // Immun Ageing. 2018. Vol. 15. P. 33.Li X. et al. Application of biological age assessment of Chinese population in potential anti-ageing technology // Immun Ageing. 2018. Vol. 15. P. 33.Bae C.-Y. et al. Biological age and lifestyle in the diagnosis of metabolic syndrome: the NHIS health screening data, 2014–2015 // Sci Rep. 2021. Vol. 11. P. 444.Bae C.-Y. et al. Biological age and lifestyle in the diagnosis of metabolic syndrome: the NHIS health screening data, 2014–2015 // Sci Rep. 2021. Vol. 11. P. 444.Kang Y.G. et al. Models for estimating the metabolic syndrome biological age as the new index for evaluation and management of metabolic syndrome // Clin Interv Aging. 2017. Vol. 12. P. 253–261.Kang Y.G. et al. Models for estimating the metabolic syndrome biological age as the new index for evaluation and management of metabolic syndrome // Clin Interv Aging. 2017. Vol. 12. P. 253–261.Kang Y.G. et al. Biological age as a health index for mortality and major age-related disease incidence in Koreans: National Health Insurance Service – Health screening 11-year follow-up study // Clin Interv Aging. 2018. Vol. 13. P. 429–436.Kang Y.G. et al. Biological age as a health index for mortality and major age-related disease incidence in Koreans: National Health Insurance Service – Health screening 11-year follow-up study // Clin Interv Aging. 2018. Vol. 13. P. 429–436.Chen L. et al. Modeling biological age using blood biomarkers and physical measurements in Chinese adults // eBioMedicine. 2023. Vol. 89. P. 104458.Chen L. et al. Modeling biological age using blood biomarkers and physical measurements in Chinese adults // eBioMedicine. 2023. Vol. 89. P. 104458.Verschoor C.P. et al. Comparing Biological Age Estimates Using Domain-Specific Measures From the Canadian Longitudinal Study on Aging // J Gerontol A Biol Sci Med Sci. 2020. Vol. 76, № 2. P. 187–194.Verschoor C.P. et al. Comparing Biological Age Estimates Using Domain-Specific Measures From the Canadian Longitudinal Study on Aging // J Gerontol A Biol Sci Med Sci. 2020. Vol. 76, № 2. P. 187–194.Zhong X. et al. Estimating Biological Age in the Singapore Longitudinal Aging Study // J Gerontol A Biol Sci Med Sci. 2020. Vol. 75, № 10. P. 1913–1920.Zhong X. et al. Estimating Biological Age in the Singapore Longitudinal Aging Study // J Gerontol A Biol Sci Med Sci. 2020. Vol. 75, № 10. P. 1913–1920.Fermín-Martínez C.A. et al. AnthropoAge, a novel approach to integrate body composition into the estimation of biological age // Aging Cell. 2022. Vol. 22, № 1. P. e13756.Fermín-Martínez C.A. et al. AnthropoAge, a novel approach to integrate body composition into the estimation of biological age // Aging Cell. 2022. Vol. 22, № 1. P. e13756.Bae C.-Y. et al. A model for estimating body shape biological age based on clinical parameters associated with body composition // Clin Interv Aging. 2013. Vol. 8. P. 11–18.Bae C.-Y. et al. A model for estimating body shape biological age based on clinical parameters associated with body composition // Clin Interv Aging. 2013. Vol. 8. P. 11–18.Maaten L., Hinton G.E. Visualizing Data using t-SNE // Journal of Machine Learning Research. 2008. Vol. 9, № 86. P. 2579–2605.Maaten L., Hinton G.E. Visualizing Data using t-SNE // Journal of Machine Learning Research. 2008. Vol. 9, № 86. P. 2579–2605.Pearson K. LIII. On lines and planes of closest fit to systems of points in space // The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science. Taylor & Francis, 1901. Vol. 2, № 11. P. 559–572.Pearson K. LIII. On lines and planes of closest fit to systems of points in space // The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science. Taylor & Francis, 1901. Vol. 2, № 11. P. 559–572.Tenenbaum J.B., Silva V. de, Langford J.C. A Global Geometric Framework for Nonlinear Dimensionality Reduction // Science. American Association for the Advancement of Science, 2000. Vol. 290, № 5500. P. 2319–2323.Tenenbaum J.B., Silva V. de, Langford J.C. A Global Geometric Framework for Nonlinear Dimensionality Reduction // Science. American Association for the Advancement of Science, 2000. Vol. 290, № 5500. P. 2319–2323.Kruskal J.B. Nonmetric multidimensional scaling: A numerical method // Psychometrika. 1964. Vol. 29, № 2. P. 115–129.Kruskal J.B. Nonmetric multidimensional scaling: A numerical method // Psychometrika. 1964. Vol. 29, № 2. P. 115–129.Kruskal J.B. Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis // Psychometrika. 1964. Vol. 29, № 1. P. 1–27.Kruskal J.B. Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis // Psychometrika. 1964. Vol. 29, № 1. P. 1–27.Dasgupta S., Gupta A. An elementary proof of a theorem of Johnson and Lindenstrauss // Random Structures & Algorithms. 2003. Vol. 22, № 1. P. 60–65.Dasgupta S., Gupta A. An elementary proof of a theorem of Johnson and Lindenstrauss // Random Structures & Algorithms. 2003. Vol. 22, № 1. P. 60–65.Achlioptas D. Database-friendly random projections: Johnson-Lindenstrauss with binary coins // Journal of Computer and System Sciences. 2003. Vol. 66, № 4. P. 671–687.Achlioptas D. Database-friendly random projections: Johnson-Lindenstrauss with binary coins // Journal of Computer and System Sciences. 2003. Vol. 66, № 4. P. 671–687.

The authors declare that there are no conflicts of interest present.