American Association for Cancer Research
10780432ccr201264-sup-240646_3_supp_6424433_qdgfd0.pdf (884.49 kB)

Figure S5 from A Radiomics Model for Predicting the Response to Bevacizumab in Brain Necrosis after Radiotherapy

Download (884.49 kB)
journal contribution
posted on 2023-03-31, 22:07 authored by Jinhua Cai, Junjiong Zheng, Jun Shen, Zhiyong Yuan, Mingwei Xie, Miaomiao Gao, Hongqi Tan, Zhongguo Liang, Xiaoming Rong, Yi Li, Honghong Li, Jingru Jiang, Huiying Zhao, Andreas A. Argyriou, Melvin L.K. Chua, Yamei Tang

Figure S5. Model comparisons using ROC curve analyses.


National Key R&D Program of China

National Natural Science Foundation of China

Science and Technology Program of Guangzhou

National Medical Research Council Clinician-Scientist Award



Bevacizumab is considered a promising therapy for brain necrosis after radiotherapy, while some patients fail to derive benefit or even worsen. Hence, we developed and validated a radiomics model for predicting the response to bevacizumab in patients with brain necrosis after radiotherapy. A total of 149 patients (with 194 brain lesions; 101, 51, and 42 in the training, internal, and external validation sets, respectively) receiving bevacizumab were enrolled. In total, 1,301 radiomic features were extracted from the pretreatment MRI images of each lesion. In the training set, a radiomics signature was constructed using the least absolute shrinkage and selection operator algorithm. Multivariable logistic regression analysis was then used to develop a radiomics model incorporated in the radiomics signature and independent clinical predictors. The performance of the model was assessed by its discrimination, calibration, and clinical usefulness with internal and external validation. The radiomics signature consisted of 18 selected features and showed good discrimination performance. The model, which integrates the radiomics signature, the interval between radiotherapy and diagnosis of brain necrosis, and the interval between diagnosis of brain necrosis and treatment with bevacizumab, showed favorable calibration and discrimination in the training set (AUC 0.916). These findings were confirmed in the validation sets (AUC 0.912 and 0.827, respectively). Decision curve analysis confirmed the clinical utility of the model. The presented radiomics model, available as an online calculator, can serve as a user-friendly tool for individualized prediction of the response to bevacizumab in patients with brain necrosis after radiotherapy.