CONCEPT AND ARCHITECTURE OF AN INTELLIGENT CLINICAL DECISION SUPPORT SYSTEM FOR LUNG CANCER IMMUNOTHERAPY WITH PATIENT DATA PRIVACY PRESERVATION
DOI:
https://doi.org/10.32689/2786-7161-2026-2-6Keywords:
non-small cell lung cancer, immunotherapy, immune checkpoint inhibitors, clinical decision support system, machine learning, PD-L1, radiomics, personalized oncologyAbstract
The object of the study is the process of automated prediction of clinical response to immune checkpoint inhibitor therapy in patients with non-small cell lung cancer. The scientific problem addressed lies in the insufficient predictive accuracy of existing approaches based exclusively on the expression level of the PD-L1 biomarker, which leads to unpredictable treatment outcomes, high financial expenses for patients, and risks of developing severe immunerelated toxicity. The essence of the obtained results consists in substantiating the concept and developing an original multi-level architecture of an intelligent clinical decision support system (ICDSS), comprising modules for data collection and integration, CT image radiomic feature extraction, prognostic modeling based on a stacking ensemble (Random Forest, XGBoost, and logistic regression), integration with NCCN clinical guidelines, and a visualization interface. Due to their specific features and characteristic distinctions, namely the transition to multimodal analysis (integrating clinico-pathological, molecular, laboratory, and radiomic predictors within a single prognostic model) and the use of explainable artificial intelligence (XAI) methods via SHAP (SHapley Additive exPlanations) values, these results allowed for overcoming the limitations of single-dimensional biological markers and ensuring high transparency of conclusions while maintaining confidentiality. The obtained results are explained by the synergistic effect of combining base algorithms into a stacking ensemble, which minimizes prediction variance and increases model robustness against overfitting on limited clinical samples, as well as the capability of radiomic features to objectively reflect the spatial heterogeneity of the tumor microenvironment. In practice, these solutions can be utilized in oncological facilities in the form of a web application, provided that patient DICOM CT files, laboratory, and immunohistochemical reports are available, offering flexible integration capabilities into current medical information systems via the HL7 FHIR interface while maintaining a balance between personal data protection and the needs of clinical research.
References
Sung, H., Ferlay, J., Siegel, R. L. et al. (2021). Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. Vol. 71, no. 3. P. 209–249. https://doi.org/10.3322/caac.21660
Reck, M., Rodríguez-Abreu, D., Robinson, A. G. et al. (2016). Pembrolizumab versus Chemotherapy for PD-L1–Positive Non–Small-Cell Lung Cancer. N Engl J Med. Vol. 375. P. 1823–1833. https://doi.org/10.1056/NEJMoa1606774
Mitra, J., Ghose, S., Thawani, R. (2025). Clinically Explainable Prediction of Immunotherapy Response Integrating Radiomics and Clinico-Pathological Information in Non-Small Cell Lung Cancer. Cancers. Vol. 17, no. 16. P. 2679. https://doi.org/10.3390/cancers17162679
Khunger, M., Hernandez, A. V., Pasupuleti, V. et al. (2017). Programmed Cell Death 1 (PD-1) Ligand (PD-L1) Expression in Solid Tumors As a Predictive Biomarker of Benefit From PD-1/PD-L1 Axis Inhibitors: A Systematic Review and Meta-Analysis. JCO Precis Oncol. No. 1. P. 1–15. https://doi.org/10.1200/PO.16.00030
Hosny, A., Parmar, C., Quackenbush, J. et al. (2018). Artificial intelligence in radiology. Nat Rev Cancer. Vol. 18, no. 8. P. 500–510. https://doi.org/10.1038/s41568-018-0016-5
Doroshow, D. B., Bhalla, S., Beasley, M. B. et al. (2021). PD-L1 as a biomarker of response to immune-checkpoint inhibitors. Nat Rev Clin Oncol. Vol. 18. P. 345–362. URL: https://doi.org/10.1038/s41571-021-00473-5
Diem, S., Schmid, S., Krapf, M. et al. (2017). Neutrophil-to-Lymphocyte ratio (NLR) and Platelet-to-Lymphocyte ratio (PLR) as prognostic markers in patients with non-small cell lung cancer (NSCLC) treated with nivolumab. Lung Cancer. Vol. 111. P. 176–181. https://doi.org/10.1016/j.lungcan.2017.07.024
Li, C., Zhou, Z., Hou, L. et al. (2024). A novel machine learning model for efficacy prediction of immunotherapy-chemotherapy in NSCLC based on CT radiomics. Comput Biol Med. Vol. 178. P. 108638. https://doi.org/10.1016/j.compbiomed.2024.108638
Ahluwalia, V. S., Parikh, R. B. (2025). Explainable Machine Learning to Predict Treatment Response in Advanced Non-Small Cell Lung Cancer. JCO Clin Cancer Inform. Vol. 9. P. e2400157. https://doi.org/10.1200/CCI-24-00157
Rakaee, M., Tafavvoghi, M., Ricciuti, B. et al. (2025). Deep Learning Model for Predicting Immunotherapy Response in Advanced Non-Small Cell Lung Cancer. JAMA Oncol. Vol. 11, no. 2. P. 109–118. https://doi.org/10.1001/jamaoncol.2024.5356
Chang, J. S., Kim, H., Baek, E. S. et al. (2025). Continuous multimodal data supply chain and expandable clinical decision support for oncology. npj Digit Med. Vol. 8, no. 1. P. 128. https://doi.org/10.1038/s41746-025-01508-2
Fridman, W. H., Pagès, F., Sautès-Fridman, C., Galon, J. (2012). The immune contexture in human tumours: impact on clinical outcome. Nat Rev Cancer. Vol. 12. P. 298–306. https://doi.org/10.1038/nrc3245
Collins, G. S., Reitsma, J. B., Altman, D. G., Moons, K. G. M. (2015). Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD): The TRIPOD Statement. Ann Intern Med. Vol. 162, no. 1. P. 55–63. https://doi.org/10.7326/M14-0697
Про захист персональних даних : Закон України. URL: https://zakon.rada.gov.ua/laws/show/2297-17
Думчиков, М. О., Малетова, О. С., Уткіна, М. С., Ободяк, В. К. (2026). Трансформація національних підходів щодо захисту персональних даних у кіберпросторі на основі стандартів ЄС: Посібник / за ред. М. О. Думчикова. Суми : СумДУ, 196 c.
Lu S., Zheng X., Wang Y. et al. (2025). Machine learning–based prognostic modeling for locally advanced non-small cell lung cancer treated with immuno-radiotherapy. Front Pharmacol. Vol. 16. P. 1723874. https://doi.org/10.3389/fphar.2025.1723874
Lundberg, S. M., Lee, S.-I. (2017). A Unified Approach to Interpreting Model Predictions. Advances in Neural Information Processing Systems. Vol. 30. URL: https://proceedings.neurips.cc/paper/2017/hash/8a20a8621978632d76c43dfd28b67767-Abstract.html
