Comprehensive evaluation of a deep learning model for automatic organs at risk segmentation on heterogeneous computed tomography images for abdominal radiotherapy (Accepted to International Journal of Radiation Oncology Biology Physics).

• This work was modified from nnUNet.
• Due to data privacy protection, we can not release all-used hospital datasets, but we released 170 cases for academic research: please contact Xiangde (luoxd1996 AT gmail DOT com) for the dataset, please check the access requirement of this dataset in Here.

• PyTorch version >=1.8
• Some common python packages such as Numpy, SimpleITK, OpenCV, Scipy......

• Download the trained model (trained based our proposed method) from Google Drive.
• Now, you can use the following code to generate 16 OARs delineation.

from InferRobustABOD import Inference3D
Inference3D(rawf="liver_70_img.nii.gz", save_path="liver_70_pred.nii.gz") # rawf is the path of input image; save_path is the path of prediction.

• This project was originally developed for our previous work AbsegNet, if you find it's useful for your research, please consider to cite the followings:

    @article{liao2023AbsegNet,
    title={Comprehensive evaluation of a deep learning model for automatic organs at risk segmentation on heterogeneous computed tomography images for abdominal radiotherapy},
    author={Liao, Wenjun, Luo, Xiangde, He, Yuan, Dong, Ye, Li, Churong, Li, Kang, Zhang, Shichuan, Zhang, Shaoting, Wang, Guotai, and Jianghong Xiao.},
    journal={ International Journal of Radiation Oncology Biology Physics},
    DOI={https://doi.org/10.1016/j.ijrobp.2023.05.034},
    year={2023},
    publisher={Elsevier}
    }
  

or

Liao, Wenjun, Luo, Xiangde, He, Yuan, Dong, Ye, Li, Churong, Li, Kang, Zhang, Shichuan, Zhang, Shaoting, Wang, Guotai, and Jianghong Xiao. "Comprehensive evaluation of a deep learning model for automatic organs at risk segmentation on heterogeneous computed tomography images for abdominal radiotherapy." International Journal of Radiation Oncology*Biology*Physics, (2023). Accessed May 26, 2023. https://doi.org/10.1016/j.ijrobp.2023.05.034.

If you have any question, please contact Xiangde Luo.