Conventional finite element analysis of the lumbar spine is strictly limited to the static geometry obtained via computed tomography with the patient in a static, supine position. Direct calculation of the stressstrain state for alternative physiological postures in real time is unfeasible due to high computational demands and the lack of computational meshes corresponding to various body positions (such as forward bending, lateral bending, or load-bearing) and loading scenarios. This article describes the development of a hybrid biomechanical approach that utilizes resting-state computed tomography data to predict changes in the stress-strain state resulting from variations in body position and loading conditions. Gradient boosting algorithms – specifically XGBoost and CatBoost – are employed to approximate the physical fields. The proposed approach enables the identification of zones with critical stress and strain concentrations during changes in body position and allows for the modeling of biomechanical states that are otherwise inaccessible to direct numerical analysis.

1. Guseva V. R., Nikitina V. V. (2023) Degenerative-Dystrophic Diseases of the Spine, Pathogenesis, Clinical Manifestations, Diagnosis, Treatment. Flagship of Science. 55–64 (in Russian).

2. Kok J., Shcherbakova Y. M., Schlоsser T. P., Seevinck P. R., Velden T. A., Castelein R. M., et al. (2023) Automatic Generation of Subject-Specific Finite Element Models of the Spine from Magnetic Resonance Images. Frontiers in Bioengineering and Biotechnology. 11.

3. Tsitko E. L., Smeyanovich A. F., Astapovich E. S., Tsitko E. V. (2018) Stress-Strain State of the Lumbar Spine at Stage III of the Degenerative Process. Multidisciplinary Approach to the Diagnosis and Treatment of Comorbid Pathology, Collection of Scientific Papers from the Republican Scientific and Practical Conference with International Participation, Gomel, Nov. 29–30. Gomel, Gomel State Medical University. 506–509 (in Russian).

4. Kurachka K., Wang X. (2025) Automation of Primary Diagnostics of Diseases of the Human Lumbar Spine Using Intelligent Analysis of CT Images. Open Semantic Technologies for Intelligent Systems, Research Papers Collection. (9). 355–360.

5. Semenchenya T. S., Kurochka K. S. (2025) Mathematical Model of the Stress-Strain State of the Human Lumbar Spine Based on the Analysis of CT Images. Information Technologies and Systems, Proceedings of the International Scientific Conference, Minsk, Nov. 19. Minsk, Belarusian State University of Informatics and Radioelectronics. 177–178 (in Russian).

6. Hurtado-Aviles J., Roca-Gonzalez J., Kurochka K. S., Manuel Sanz-Mengibar J., Santonja-Medina F. (2021) Developing of a Mathematical Model to Perform Measurements of Axial Vertebral Rotation on Computer-Aided and Automated Diagnosis Systems, Using Raimondi’s Method. Radiology Research and Practice. 1, 55–64.

7. Yang H., Jekir M. G., Davis M. W., Keaveny T. M. (2016) Effective Modulus of the Human Intervertebral Disc and Its Effect on Vertebral Bone Stress. Journal of Biomechanics. 49, 1134–1140.

8. Chumachenko E. N., Logashina I. V. (2014) Calculation of the Stress-Strain State of the Spinal Motion Segment Under Loads. Aerospace and Environmental Medicine. 48 (5), 51–57 (in Russian).

9. Massaron L., Ryan M. (2026) Machine Learning on Tabular Data: XGBoost, Deep Learning, and AI. Saint Petersburg, Piter Publ. (in Russian).

10. Probst P., Boulesteix A.-L., Bischl B. (2019) Tunability: Importance of Hyperparameters of Machine Learning Algorithms. Journal of Machine Learning Research. 20 (53), 1–32.