AUTHOR=Yang Hong , Wang Lin , Shao Guoliang , Dong Baiqiang , Wang Fang , Wei Yuguo , Li Pu , Chen Haiyan , Chen Wujie , Zheng Yao , He Yiwei , Zhao Yankun , Du Xianghui , Sun Xiaojiang , Wang Zhun , Wang Yuezhen , Zhou Xia , Lai Xiaojing , Feng Wei , Shen Liming , Qiu Guoqing , Ji Yongling , Chen Jianxiang , Jiang Youhua , Liu Jinshi , Zeng Jian , Wang Changchun , Zhao Qiang , Yang Xun , Hu Xiao , Ma Honglian , Chen Qixun , Chen Ming , Jiang Haitao , Xu Yujin 

TITLE=A combined predictive model based on radiomics features and clinical factors for disease progression in early-stage non-small cell lung cancer treated with stereotactic ablative radiotherapy

JOURNAL=Frontiers in Oncology

VOLUME=Volume 12 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.967360

DOI=10.3389/fonc.2022.967360

ISSN=2234-943X

ABSTRACT=Purpose: To accurately assess disease progression after stereotactic ablative radiotherapy (SABR) of early-stage non-small cell lung cancer（NSCLC）, a combined predictive model based on pre-treatment CT radiomics features and clinical factors was established. 
Methods: This study retrospectively analyzed the data of 96 patients with early-stage NSCLC treated with SABR. Clinical factors included general patient information (sex, age, KPS, Charlson score, lung function, smoking status), relevant information of lesions (diameter, location, pathological type, T stage), radiotherapy parameters (Biological Effective Dose),  types of peritumoral radiation–induced lung injury. Independent risk factors were screened by logistic regression analysis. Radiomics features were extracted from pre-treatment CT. Minimum redundancy maximum relevance (MRMR) and least absolute selector operator (Lasso) were adopted for the dimensionality reduction and feature selection. According to the weight coefficient of the features, the radiomic score was calculated and the radiomic model was constructed. Multiple logistic regression analysis was used to establish the combined model based on radiomic features and clinical factors. Receiver operating characteristic curve (ROC)，DeLong test，Hosmer-Lemeshow test, and decision curve analysis（DCA） were used to evaluate the diagnostic efficiency and clinical practicability of the model. 
Results: After the median follow-up time of 59.7 months, 29 patients experienced progression within 2 years. Among the clinical factors, the type of peritumoral radiation–induced lung injury was the only independent risk factor for progression (P < 0.05). Eleven features were selected from 1781 features to construct a radiomics model. For predicting disease progression after SABR, the area under curve（AUC）of training and validation cohorts in the radiomics model were 0.88（95%CI 0.80-0.96）and 0.80（95%CI 0.62-0.98），AUC of training and validation cohorts in the combined model were 0.88（95%CI 0.81-0.96）and 0.81（95%CI 0.62-0.99）. Both the radiomics model and the combined model have good prediction efficiency in the training cohort and the validation cohort, but DeLong test shows that there is no difference between them. 
Conclusions: Compared with the clinical model, the radiomics model and the combined model can better predict the disease progression of early-stage NSCLC after SABR, which might contribute to making individualized follow-up plans and treatment strategies.