[1]韦尚炎,彭慧娟,刘倩,等.基于多任务学习的自动分割质量预测模型在直肠癌临床靶区中的应用[J].中国医学物理学杂志,2026,43(5):675-681.[doi:DOI:10.3969/j.issn.1005-202X.2026.05.016]
 WEI Shangyan,PENG Huijuan,et al.Application of a multi-task learning-based quality prediction model for automatic clinical target volume segmentation in rectal cancer[J].Chinese Journal of Medical Physics,2026,43(5):675-681.[doi:DOI:10.3969/j.issn.1005-202X.2026.05.016]
点击复制

基于多任务学习的自动分割质量预测模型在直肠癌临床靶区中的应用()

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

卷:
43卷
期数:
2026年第5期
页码:
675-681
栏目:
医学人工智能
出版日期:
2026-05-28

文章信息/Info

Title:
Application of a multi-task learning-based quality prediction model for automatic clinical target volume segmentation in rectal cancer
文章编号:
1005-202X(2026)05-0675-07
作者:
韦尚炎12彭慧娟2刘倩2唐源2陈辛元2门阔2全红1
1.武汉大学物理科学与技术学院, 湖北 武汉 430072; 2.国家癌症中心/国家肿瘤临床医学研究中心/中国医学科学院北京协和医学院肿瘤医院, 北京 100021
Author(s):
WEI Shangyan1 2 PENG Huijuan2 LIU Qian2 TANG Yuan2 CHEN Xinyuan2 MEN Kuo2 QUAN Hong1
1. School of Physics and Technology, Wuhan University, Wuhan 430072, China 2. National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100021, China
关键词:
直肠癌预测模型自动分割分割质量预测多任务学习
Keywords:
Keywords: rectal cancer prediction model automatic segmentation segmentation quality prediction multi-task learning
分类号:
R318;R811.1
DOI:
DOI:10.3969/j.issn.1005-202X.2026.05.016
文献标志码:
A
摘要:
【摘要】目的:针对放疗自动分割结果仍需人工逐层审核、质控效率较低的问题,提出一种自动分割质量预测模型,以提高自动分割结果的审核效率。方法:基于Transformer框架,并结合多任务学习策略,构建一个端到端的质量预测模型。该模型以CT图像与分割掩码为输入,输出对应分割图的Dice相似性系数(DSC)预测值。收集570例直肠癌患者的放疗数据,针对临床靶区(CTV)分割结果进行质量预测,采用平均绝对误差(MAE)作为主要评估指标。结果:多任务训练策略下模型预测CTV分割的DSC的MAE为0.035 4±0.042 2,优于单任务训练的0.040 2±0.053 6,表明多任务学习有助于提升预测精度。结论:本研究提出的多任务质量预测模型在直肠癌CTV分割质量评估中表现出良好的准确性,模型可用于辅助临床快速识别有问题的分割层面,为医生提供量化参考指标,从而提高分割审核效率,推动放疗流程的智能化与标准化。
Abstract:
Abstract: Objective Given that automatic segmentation results in radiotherapy still require manual slice-by-slice review and exhibit low quality control efficiency, this study proposes a quality prediction model for automatic segmentation, thus improving the review efficiency of automatic segmentation results. Methods An end-to-end quality prediction model was developed based on the Transformer architecture combined with a multi-task learning strategy. The model took CT images and corresponding segmentation masks as inputs, and output the predicted Dice similarity coefficient (DSC) for the given segmentation. A dataset comprising 570 rectal cancer patients was collected for quality prediction for clinical target volume (CTV) segmentation, with mean absolute error as the primary evaluation metric. Results Under the multi-task learning strategy, the model achieved a mean absolute error of 0.035 4±0.042 2 for CTV segmentation DSC prediction, outperforming the single-task training result of 0.040 2±0.053 6 and verifying the efficacy of multi-task learning in improving prediction accuracy. Conclusion The proposed multi-task quality prediction model demonstrates favorable accuracy in evaluating CTV segmentation quality for rectal cancer. It can assist clinicians in quickly identifying potentially inaccurate segmentation slices and provide quantitative references, thereby improving the efficiency of segmentation review and promoting the intelligent and standardized workflow in radiotherapy.

相似文献/References:

[1]张和华,吕 洋,苌飞霸,等.基于支持向量机的胸阻抗与胸外按压深度预测模型构建[J].中国医学物理学杂志,2015,32(03):343.[doi:10.3969/j.issn.1005-202X.2015.03.009]
[2]张武哲,陈志坚,彭 逊,等.不同角度固定野动态调强在直肠癌术前放疗中的剂量学比较[J].中国医学物理学杂志,2014,31(02):4735.[doi:10.3969/j.issn.1005-202X.2014.02.004]
[3]李忠伟,张九堂,吴智理,等.R620-ACF型腹部碳纤维托架对直肠癌调强放射治疗剂量的影响[J].中国医学物理学杂志,2014,31(05):5148.[doi:10.3969/j.issn.1005-202X.2014.05.012]
[4]吴虹,翟福山,刘明,等.应用CBCT研究膀胱体积大小对直肠癌术后放疗剂量学的影响[J].中国医学物理学杂志,2013,30(04):4253.[doi:10.3969/j.issn.1005-202X.2013.04.010]
[5]张艺宝,吴昊,李莎,等.临床前验证与几何对比分析基于图谱库的危及器官自动勾画[J].中国医学物理学杂志,2015,32(06):761.[doi:doi:10.3969/j.issn.1005-202X.2015.06.001]
 [J].Chinese Journal of Medical Physics,2015,32(5):761.[doi:doi:10.3969/j.issn.1005-202X.2015.06.001]
[6]王沛沛,刘哲铭,李彩虹,等.容积旋转调强与固定野动态调强在直肠癌新辅助放疗中的剂量学比较[J].中国医学物理学杂志,2015,32(06):901.[doi:doi:10.3969/j.issn.1005-202X.2015.06.030]
 [J].Chinese Journal of Medical Physics,2015,32(5):901.[doi:doi:10.3969/j.issn.1005-202X.2015.06.030]
[7]姚文良,许波,戚金凤,等.直肠癌术前不同照射技术剂量学比较[J].中国医学物理学杂志,2016,33(3):317.[doi:DOI:10.3969/j.issn.1005-202X.2016.03.021]
 [J].Chinese Journal of Medical Physics,2016,33(5):317.[doi:DOI:10.3969/j.issn.1005-202X.2016.03.021]
[8]黎旦,宾石珍,程品晶,等.非晶硅电子射野影像系统与Arc CHECK在直肠癌容积旋转调强剂量验证中的应用[J].中国医学物理学杂志,2016,33(7):674.[doi:10.3969/j.issn.1005-202X.2016.07.006]
 [J].Chinese Journal of Medical Physics,2016,33(5):674.[doi:10.3969/j.issn.1005-202X.2016.07.006]
[9]王贝,全红,邱杰,等. 容积旋转调强技术用于直肠癌时最佳机架角度设置研究[J].中国医学物理学杂志,2016,33(9):898.[doi:10.3969/j.issn.1005-202X.2016.09.006]
 [J].Chinese Journal of Medical Physics,2016,33(5):898.[doi:10.3969/j.issn.1005-202X.2016.09.006]
[10]孙永健,曹洋森,代智涛,等.Monaco设计直肠癌固定机架角调强和容积旋转调强治疗的剂量[J].中国医学物理学杂志,2017,34(2):161.[doi:10.3969/j.issn.1005-202X.2017.02.010]
 Dosimetric comparison of seven- field intensity modulated radiotherapy and volumetric modulated arc therapy for rectal cancer designed with Monaco[J].Chinese Journal of Medical Physics,2017,34(5):161.[doi:10.3969/j.issn.1005-202X.2017.02.010]

备注/Memo

备注/Memo:
【收稿日期】2026-01-05 【基金项目】国家自然科学基金(12175312) 【作者简介】韦尚炎,硕士研究生,研究方向:医学图像处理与人工智能在放射治疗中的应用,E-mail: weishangyan2023@163.com 【通信作者】全红,副教授,研究方向:医学影像、放射治疗、纳米技术在肿瘤放化疗中的应用,E-mail: 00007962@whu.edu.cn
更新日期/Last Update: 2026-05-29