Convolutional neural network based dose prediction method for intensity-modulatedradiotherapy of cervical cancer(PDF)
《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]
- Issue:
- 2025年第4期
- Page:
- 421-428
- Research Field:
- 医学放射物理
- Publishing date:
Info
- Title:
- Convolutional neural network based dose prediction method for intensity-modulatedradiotherapy of cervical cancer
- Author(s):
- WU Xiaojuan1; 2; ZHANG Yibao3; REN Hongru1; MENG Lingjun2
- 1. School of Automation, Guangdong University of Technology, Guangzhou 510000, China; 2. Department of Radiation Oncology,Dezhou Hospital, Qilu Hospital of Shandong University, Dezhou 253000, China; 3. Department of Radiation Oncology, PekingUniversity Cancer Hospital & Institute/Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education), Beijing100142, China
- Keywords:
- cervical cancer; dose prediction; convolutional neural network; intensity-modulated radiotherapy
- PACS:
- R318；R815.2；TP391.4
- DOI:
- 10.3969/j.issn.1005-202X.2025.04.001
- Abstract:
- Objective To develop a convolutional neural network based model for predicting the dose distribution of intensitymodulated radiotherapy (IMRT) in cervical cancer, and to evaluate its potential applications in automated treatment planning.Methods The pelvic IMRT plans for 100 female patients were collected, with 80 cases in the training set, 10 in the validationset, and 10 in the test set. A dose prediction model was built based on the three-dimensional (3D) residual network forforecasting 3D dose distribution. Masks for organs-at-risk and planning target areas were extracted from CT images and RTStructure files. Density values were assigned to different structures according to a density map, and the resulting CT mapswere used as input images for model training. The optimal model was used to predict the 3D dose distribution, and thepredicted results were compared with the dose distribution from manual treatment planning in terms of dosimetricparameters. Results The experimental results on the 10-case test set demonstrated that dosimetric parameter differences wereinsignificant and within clinically acceptable ranges. The mean absolute error, average Dice similarity coefficient, and 95%Hausdorff distance for 10 cases in test set were (0.58±0.16) Gy, 0.90±0.03, and (10.61±7.17) mm, respectively. Comparedwith manual planning, prediction model showed slightly decreased rectal V45, small bowel D2 cc, and the V20of bilateralfemoral heads was reduced. The predicted D95 of planning target area was lower than manual planning, but the differences inD90, homogeneity index, and conformity index were trivial. There were minor differences in 3D dose distributions betweenthe two, and the dose distribution generated by prediction model met clinical requirements. Conclusion The convolutional neural network based dose prediction model can accurately forecast the dose distribution for cervical cancer IMRT, exhibitingthe potential to be used in automated treatment planning and quality evaluation.
Last Update: 2025-04-30