Prediction of Radiation-induced Lung Injury in Postoperative Breast Cancer Patients with Intensity-modulated Radiotherapy based on Radiomics and Dosiomics

DOI: https://doi.org/10.62517/jmhs.202405102

Author(s)

Li Mo, Ling Lin, Liuke Liang, Qingguo Fu, Suning Huang, Haiming Yang*

Affiliation(s)

Department of Radiotherapy, Guangxi Medical University Cancer Hospital, Nanning, Guangxi, China *Corresponding Author.

Abstract

This study aims to establish a prediction model for radiation-induced lung injury in postoperative breast cancer patients with intensity-modulated radiotherapy (IMRT) by combining CT-based radiomics with dosiomics. A retrospective analysis was conducted on 79 female breast cancer patients diagnosed and treated at the Affiliated Tumor Hospital of Guangxi Medical University from September 2019 to October 2022. CT images, treatment plans, and clinical follow-up data of these patients were collected. The affected lung was taken as the region of interest (ROI), from which 3384 radiomic and dose radiomic features were extracted. 25 prediction models were constructed using 5 feature extraction methods and 5 classifiers. The ‘fsv-Logistic Regression model’ emerged as the optimal combination, with the training set AUC values of 0.822, 0.844and 0.898, and validation set AUC values of 0.752, 0.752and 0.914. The combined model demonstrated superior predictive performance compared to individual radiomics or dosiomics models. The results of this study show that, CT-based radiomic features combined with dosiomics enhance the predictive efficacy of radiation-induced lung injury in postoperative breast cancer patients with IMRT.

Keywords

Breast Cancer; Radiation-Induced Lung Injury; Radiomics; Dosiomics

References

[1]Wen G, Tan YT, Lan XW, et al. New clinical features and dosimetric predictor identification for symptomatic radiation pneumonitis after tangential irradiation in breast cancer patients. Cancer, 2017, 8(18): 3795-3802. [2]Liu Jing, Zhou Dong, He Jiabei, et al. Research Progress on Factors Affecting Radiation Pneumonia in Breast Cancer Radiotherapy. Journal of Dalian Medical University, 2022, 44(5): 440-445. [3]Gillies RJ, Kinahan PE, Hricak H. Radiomics images are more than pictures, they are data. Radiology, 2016, 278(2): 563-577. [4]Van Timmeren JE, Leijenaar RT, Van Elmpt W, et al. Survival prediction of non-small cell lung cancer patients using radiomics analyses of cone-beam CT images. Radiother Oncol, 2017, 123(3): 363-369. [5]Rossi L, Bijman R, Schillemans W, et al. Texture analysis of 3D dose distributions for predictive modelling of toxicity rates in radiotherapy. Radiother Oncol, 2018, 129(3): 548‐553. [6]CoxJ D, Stetz J, Pajak TF. Toxicity criteria of the Radiation Therapy Oncology Group (RT0G) and the European Organization for Research and Treatment of Cancer (EORTC). Int J Radiat Oncol Biol Phys. 1995, 31(5): 1341-1346. [7]Peng Lin, Peng-Fei Yang, Shi Chen, et al. A Delta-radiomics model for preoperative evaluation of Neoadjuvant chemotherapy response in high-grade osteosarcoma. Cancer Imaging, 2020, 20(7): 1-12. [8]Zhou Lu, Wang Linjing, Zhang Guoqian, et al. Prediction of radiation pneumonitis based on radiomics and dosiomics. Chinese Journal of Medical Physics, 2023, 40(7): 808-813. [9]Roffo G, Melzi S, Castellani U, et al. Infinite Latent Feature Selection: A Probabilistic Latent Graph-Based Ranking Approach. IEEE, 2017.