Machine Learning-Based Radiomic Signatures for Predicting Immunotherapy Response in Advanced Non-Small Cell Lung Cancer

• E.M.J. Karthikeyan

  Department of General Surgery, Vinayaka Mission’s Kirupananda Variyar Medical College & Hospitals; Vinayaka Mission’s Research Foundation (DU), Salem, Tamil Nadu, India

• Manoranjan Parhi

  Department of Center for Data Science, IMS and SUM Hospital, Siksha 'O' Anusandhan (Deemed to be University), Bhubaneswar, Odisha, India

• P. Grace Kanmani Prince

  Department of Biomedical Engineering, Sathyabama Institute of Science and Technology, Chennai, Tamil Nadu, India

• Arvindkumar M. Chauhan

  Faculty of Allied and Healthcare, Gokul Global University, Sidhpur, Gujarat, India

• Shitij Goyal

  Quantum University Research Center, Quantum University, Roorkee, Uttarakhand, 247667, India

• Prerak Sudan

  Center for Research Impact & Outcome, Chitkara University Institute of Engineering and Technology, Chitkara University, Rajpura, 140401, Punjab, India

Immunotherapy has brought about many advances in treatment options for patients with advanced non-small cell lung cancer (NSCLC), although there is great variability among patients' responses to the treatment; thus, there is a need for accurate predictive methods. This study seeks to analyze the predictive power of machine learning models using computed tomography (CT)-based radiomic features to identify responders to immunotherapy in patients with metastatic NSCLC. The use of pretreatment CT images and responses to treatment was analyzed in a retrospective cohort of 150 patients with advanced non-small cell lung cancer. The responders were those who demonstrated full or partial effectiveness of treatment, while the non-responders were the patients who had cancer stabilisation or progressive illness. Tumor segmentation was performed, and radiomic features of intensity, shape, texture, and higher-order features were extracted. Feature selection and dimensionality reduction were performed before constructing Support Vector Machine (SVM), Random Forest (RF), and Gradient Boosting (GB) classification models. The predictive performance of the models was evaluated using accuracy, sensitivity, specificity, and the area under the curve (AUC) of the receiver operating characteristic (ROC) curve. The RF classifier provided the best accuracy at 88.6%, sensitivity at 86.9%, specificity at 89.7%, and an AUC value of 0.92. In feature ablation, the accuracy rate obtained by CT combined radiomic features was 88.6%, and AUC was 0.92, which was superior to the texture features with accuracy and AUC of 83.7% and 0.87, respectively, and intensity features with 78.5% and 0.82, respectively. It is evident that CT radiomic features might be helpful for investigating the immune therapy response. However, further validation is needed.

[1] Tong H, Sun J, Fang J, Zhang M, Liu H, Xia R, Zhou W, Liu K, Chen X. A machine learning model based on PET/CT radiomics and clinical characteristics predicts tumor immune profiles in non-small cell lung cancer: a retrospective multicohort study. Frontiers in Immunology 2022; 13: 859323.

[2] Cenitta D, Patil S, Asha K, Anitha M, Sowmya T, Pai P, Arjunan RV. A Privacy-Preserving Cloud-Based System for Cardiovascular Disease Prediction Using Lightweight Machine Learning Models. J Internet Serv Inf Secur 2025; 15(3): 129-48.

[3] Cousin F, Louis T, Dheur S, Aboubakar F, Ghaye B, Occhipinti M, Vos W, Bottari F, Paulus A, Sibille A, Vaillant F. Radiomics and delta-radiomics signatures to predict response and survival in patients with non-small-cell lung cancer treated with immune checkpoint inhibitors. Cancers 2023; 15(7): 1968.

[4] Aerts HJ, Velazquez ER, Leijenaar RT, Parmar C, Grossmann P, Carvalho S, Bussink J, Monshouwer R, Haibe-Kains B, Rietveld D, Hoebers F. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nature Communications 2014; 5(1): 4006.

[5] Wu Y, Zhang W, Liang X, Zhang P, Zhang M, Jiang Y, Cui Y, Chen Y, Zhou W, Liang Q, Dai J. Habitat radiomics analysis for progression free survival and immune-related adverse reaction prediction in non-small cell lung cancer treated by immunotherapy. Journal of Translational Medicine 2025; 23(1): 393.

[6] Devi A, NirmalRaj T. SRD-GDT²: a SMOTE-regularized dual-guided deep tabular framework for accurate cardiovascular disease prediction. Int Acad J Sci Eng 2025; 12(3): 512-526.

[7] Kothari G. Role of radiomics in predicting immunotherapy response. Journal of medical imaging and radiation oncology 2022; 66(4): 575-91.

[8] Lambin P, Rios-Velazquez E, Leijenaar R, Carvalho S, Van Stiphout RG, Granton P, Zegers CM, Gillies R, Boellard R, Dekker A, Aerts HJ. Radiomics: extracting more information from medical images using advanced feature analysis. European Journal of Cancer 2012; 48(4): 441-6.

[9] Yolchuyeva S, Giacomazzi E, Tonneau M, Lamaze F, Orain M, Coulombe F, Malo J, Belkaid W, Routy B, Joubert P, Manem VS. Radiomics approaches to predict PD-L1 and PFS in advanced non-small cell lung patients treated with immunotherapy: a multi-institutional study. Scientific Reports 2023; 13(1): 11065.

[10] Gupta N, Verma A. The role of inflammation in cardiovascular disease. In: Medxplore: Frontiers in Medical Science. Periodic Series in Multidisciplinary Studies 2025; pp. 37-51.

[11] Peng J, Zou D, Zhang X, Ma H, Han L, Yao B. A novel sub-regional radiomics model to predict immunotherapy response in non-small cell lung carcinoma. Journal of Translational Medicine 2024; 22(1): 87.

[12] Zheng J, Xu S, Wang G, Shi Y. Applications of CT-based radiomics for the prediction of immune checkpoint markers and immunotherapeutic outcomes in non-small cell lung cancer. Frontiers in immunology 2024; 15: 1434171.

[13] Min J, Dong F, Chen Y, Li W, Wu Y, Tan Y, Yang F, Wu P, Chai Y. The NSCLC immunotherapy response predicted by tumor-infiltrating T cells via a non-invasive radiomic approach. Frontiers in immunology 2024; 15: 1379812.

[14] Chand G, Cook G, Hughes D, Meszaros L, Wong N, Ting H, Zhao L, Xing Y, Zhao J. Assessing PD-L1 expression in non-small cell lung cancer patients using68Ga-labeled anti-programmed death ligand-1 (PD-L1) single-domain antibody PET/CT.

[15] Yuan L, Wang Q, Sun F, Shi H. Deep learning and inflammatory markers predict early response to immunotherapy in unresectable NSCLC: A multicenter study. Biomolecules and Biomedicine 2025; 25(10): 2252.

[16] Wang J, Wang J, Huang X, Zhou Y, Qi J, Sun X, Nie J, Hu Z, Wang S, Hong B, Wang H. CT radiomics-based model for predicting TMB and immunotherapy response in non-small cell lung cancer. BMC medical imaging 2024; 24(1): 45.

[17] Yolchuyeva S, Giacomazzi E, Tonneau M, Ebrahimpour L, Lamaze FC, Orain M, Coulombe F, Malo J, Belkaid W, Routy B, Joubert P. A radiomics-clinical model predicts overall survival of non-small cell lung cancer patients treated with immunotherapy: a multicenter study. Cancers 2023; 15(15): 3829.

[18] Lin S, Ma Z, Yao Y, Huang H, Chen W, Tang D, Gao W. Automatic machine learning accurately predicts the efficacy of immunotherapy for patients with inoperable advanced non-small cell lung cancer using a computed tomography-based radiomics model. Diagnostic and Interventional Radiology 2025; 31(2): 130.

[19] Gong J, Bao X, Wang T, Liu J, Peng W, Shi J, Wu F, Gu Y. A short-term follow-up CT based radiomics approach to predict response to immunotherapy in advanced non-small-cell lung cancer. Oncoimmunology 2022; 11(1): 2028962.

[20] Benzekry S, Grangeon M, Karlsen M, Alexa M, Bicalho-Frazeto I, Chaleat S, Tomasini P, Barbolosi D, Barlesi F, Greillier L. Machine learning for prediction of immunotherapy efficacy in non-small cell lung cancer from simple clinical and biological data. Cancers 2021; 13(24): 6210.

• PDF

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.