Development and validation of AI delineation of the thoracic RTOG organs at risk with deep learning on multi-institutional datasets

Xianghua Ye; Dazhou Guo; Lujun Zhao; Congying Xie; Dandan Zheng; Haihua Yang; Xiangzhi Zhu; Xin Sun; Pingping Dong; Huanhuan Li; Weiwei Kong; Jianzhong Cao; Honglei Chen; Juntao Ran; Kai Ren; Hongxin Su; Hao Hu; Cuimeng Tian; Tianlu Wang; Qiang Zeng

doi:10.1016/j.intonc.2024.12.001

您当前的位置：

首页 >

文章列表页 >

Development and validation of AI delineation of the thoracic RTOG organs at risk with deep learning on multi-institutional datasets

Research article | 更新时间：2026-06-03

- Development and validation of AI delineation of the thoracic RTOG organs at risk with deep learning on multi-institutional datasets
- Development and validation of AI delineation of the thoracic RTOG organs at risk with deep learning on multi-institutional datasets
- 智能肿瘤学 2025年1卷第1期页码：61-71
- 作者机构：
  
  1. Department of Radiation Oncology, The First Affiliated Hospital, Zhejiang University,Zhejiang,Hangzhou,China,310003
  2. DAMO Academy, Alibaba Group,NY,New York,USA,10014
  3. Department of Radiation Oncology, Tianjin Medical University Cancer Institute and Hospital,Tianjin,China,300070
  4. Department of Radiation Oncology, The First Affiliated Hospital, Wenzhou Medical University,Zhejiang,Wenzhou,China,325000
  5. Department of Radiation Oncology, Taizhou Hospital,Zhejiang,Taizhou,China,317000
  6. Department of Radiation Oncology, Jiangsu Cancer Hospital,Jiangsu,Nanjing,China,210019
  7. Department of Radiation Oncology, Nanjing Drum Tower Hospital,Jiangsu,Nanjing,China,210008
  8. Department of Radiation Oncology, Shanxi Cancer Hospital,Shanxi,Taiyuan,China,030013
  9. Department of Radiation Oncology, Jinan JunXin Med,Shandong,Jinan,China,250001
  10. Department of Radiation Oncology, The First Hospital of Lanzhou University,Gansu,Lanzhou,China,730000
  11. Department of Radiation Oncology, Kiang Wu Hospital, Macau SAR,China,999078
  12. Department of Radiation Oncology, Beijing Chest Hospital, Capital Medical University,Beijing,China,101149
  13. Department of Radiotherapy, Cancer Hospital & Institute
  14. Cancer Hospital of China Medical University
  15. Department of Radiotherapy, Cancer Hospital of Dalian University of Technology
  16. Faculty of Medicine, Dalian University of Technology,Liaoning,Shenyang,China,110042
- 作者简介：
- 基金信息：
- DOI：10.1016/j.intonc.2024.12.001
  中图分类号：
- 收稿：2024-09-05，
  
  网络首发：2024-12-07，
  
  纸质出版：2025-01-15
- 稿件说明：
移动端阅览
Xianghua Ye, Dazhou Guo, Lujun Zhao, 等. Development and validation of AI delineation of the thoracic RTOG organs at risk with deep learning on multi-institutional datasets[J]. 智能肿瘤学, 2025,1(1):61-71.

Xianghua Ye, Dazhou Guo, Lujun Zhao, et al. Development and validation of AI delineation of the thoracic RTOG organs at risk with deep learning on multi-institutional datasets[J]. Intelligent Oncology, 2025, 1(1): 61-71.
Xianghua Ye, Dazhou Guo, Lujun Zhao, 等. Development and validation of AI delineation of the thoracic RTOG organs at risk with deep learning on multi-institutional datasets[J]. 智能肿瘤学, 2025,1(1):61-71. DOI： 10.1016/j.intonc.2024.12.001.

Xianghua Ye, Dazhou Guo, Lujun Zhao, et al. Development and validation of AI delineation of the thoracic RTOG organs at risk with deep learning on multi-institutional datasets[J]. Intelligent Oncology, 2025, 1(1): 61-71. DOI： 10.1016/j.intonc.2024.12.001.

摘要

IntroductionAccurate contouring of thoracic organs at risk (OARs) is essential for minimizing complications in radiation treatment. Manual contouring of thoracic OARs is not only time-consuming but also prone to substantial user variation. To enhance the efficiency and consistency

we developed a unified deep learning (DL) OAR contouring model

DeepOAR

that was trained using multiple partially labeled datasets for segmenting a comprehensive set of thoracic OARs following the Radiation Therapy Oncology Group (RTOG)-guided OAR atlas. This DL model supports the segmentation of six required and eight optional OARs guided by the NRG-RTOG 1106 trial

providing precise and reproducible OARs contouring that are ready to be used in radiotherapy practice.Materials and methodsFollowing the OAR contouring recommendation of the NRG-RTOG 1106 trial

we collected and curated three private datasets and two public datasets

comprising a total of 531 patients with partially annotated thoracic OARs. These partially annotated datasets were utilized to develop DeepOAR

which consisted of a shared encoder and 14 separate decoders

with each decoder dedicated to one specific OAR. For model training

we utilized all patients from the two public datasets and 75% of the patients from the private datasets. We reserved the remaining 25% of the private datasets for independent testing. A multi-user study involving 21 radiation oncologists was conducted on 40 randomly selected patients from the independent testing dataset to evaluate the clinical applicability of DeepOAR. The Dice coefficient score (DSC) and average surface distance (ASD) were computed to evaluate the quantitative delineation performance of the model.ResultsDeepOAR outperformed nnUNet (the benchmark medical segmentation model) across all 14 OARs

achieving mean DSC and ASD values of 88.4% and 1.0 mm

respectively

in the independent testing set. Multi-user validation demonstrated that 89.7% of DeepOAR-generated OARs were clinically acceptable or required only minor revisions. A comparison using two randomly selected patients showed that the delineation variability of DeepOAR was significantly smaller than the inter-user variation among radiation oncologists. Human editing of DeepOAR’s predictions could further improve OAR delineation accuracy by an average of 3% increase in DSC and 40% reduction in ASD while significantly reducing the workload of radiation oncologists for contouring 14 thoracic OARs by an average of 77.0%.ConclusionWe developed DeepOAR

a DL-based unified contouring model trained using multiple partially labeled datasets

to delineate a comprehensive set of 14 thoracic OARs following the RTOG-guided OAR atlas. Both qualitative and quantitative results demonstrated the strong clinical applicability of DeepOAR for the OAR delineation process in thoracic cancer radiotherapy workflows

along with improved efficiency

comprehensiveness

and quality.

Abstract

we developed a unified deep learning (DL) OAR contouring model

DeepOAR

providing precise and reproducible OARs contouring that are ready to be used in radiotherapy practice.Materials and methodsFollowing the OAR contouring recommendation of the NRG-RTOG 1106 trial

we collected and curated three private datasets and two public datasets

comprising a total of 531 patients with partially annotated thoracic OARs. These partially annotated datasets were utilized to develop DeepOAR

which consisted of a shared encoder and 14 separate decoders

with each decoder dedicated to one specific OAR. For model training

achieving mean DSC and ASD values of 88.4% and 1.0 mm

respectively

a DL-based unified contouring model trained using multiple partially labeled datasets

along with improved efficiency

comprehensiveness

and quality.

关键词

Keywords

references

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

From morphology to function: Advances in multimodal imaging for pulmonary function prediction

AI-based diagnosis of clear-cell renal cell carcinoma based on non-contrast CT

Deep learning-based segmentation of small-volume brain metastases in lung cancer patients

Multimodal medical imaging AI for breast cancer diagnosis: A comprehensive review

Deep learning-based localization of preoperative parathyroid glands in secondary hyperparathyroidism patients using dual-phase unenhanced and contrast-enhanced computed tomography data