Dose prediction of CyberKnife Monte Carlo plan for lung cancer patients based on deep learning: robust learning of variable beam configurations

Abstract Background Accurate calculation of lung cancer dose using the Monte Carlo (MC) algorithm in CyberKnife (CK) is essential for precise planning. We aim to employ deep learning to directly predict the 3D dose distribution calculated by the MC algorithm, enabling rapid and accurate automatic pl...

Full description

Saved in:
Bibliographic Details
Main Authors: Yuchao Miao, Jiwei Li, Ruigang Ge, Chuanbin Xie, Yaoying Liu, Gaolong Zhang, Mingchang Miao, Shouping Xu
Format: Article
Language:English
Published: BMC 2024-11-01
Series:Radiation Oncology
Subjects:
Online Access:https://doi.org/10.1186/s13014-024-02531-5
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850107376270573568
author Yuchao Miao
Jiwei Li
Ruigang Ge
Chuanbin Xie
Yaoying Liu
Gaolong Zhang
Mingchang Miao
Shouping Xu
author_facet Yuchao Miao
Jiwei Li
Ruigang Ge
Chuanbin Xie
Yaoying Liu
Gaolong Zhang
Mingchang Miao
Shouping Xu
author_sort Yuchao Miao
collection DOAJ
description Abstract Background Accurate calculation of lung cancer dose using the Monte Carlo (MC) algorithm in CyberKnife (CK) is essential for precise planning. We aim to employ deep learning to directly predict the 3D dose distribution calculated by the MC algorithm, enabling rapid and accurate automatic planning. However, most current methods solely focus on conventional intensity-modulated radiation therapy and assume a consistent beam configuration across all patients. This study seeks to develop a more versatile model incorporating variable beam configurations of CK and considering the patient’s anatomy. Methods This study proposed that the AB (anatomy and beam) model be compared with the control Mask (only anatomy) model. These models are based on a 3D U-Net network to investigate the impact of CK beam encoding information on dose prediction. The study collected 86 lung cancer patients who received CK′s built-in MC algorithm plans using different beam configurations for training/validation (66 cases) and testing (20 cases). We compared the gamma passing rate, dose difference maps, and relevant dose-volume metrics to evaluate the model’s performance. In addition, the Dice similarity coefficients (DSCs) were calculated to assess the spatial correspondence of isodose volumes. Results The AB model demonstrated superior performance compared to the Mask model, particularly in the trajectory dose of the beam. The DSCs of the AB model were 20–40% higher than that of the Mask model in some dose regions. We achieved approximately 99% for the PTV and generally more than 95% for the organs at risk (OARs) referred to the clinical planning dose in the gamma passing rates (3 mm/3%). Relative to the Mask model, the AB model exhibited more than 90% improvement in small voxels (p < 0.001). The AB model matched well with the clinical plan’s dose-volume histograms, and the average dose error for all organs was 1.65 ± 0.69%. Conclusions Our proposed new model signifies a crucial advancement in predicting CK 3D dose distributions for clinical applications. It enables planners to rapidly and precisely predict MC doses for lung cancer based on patient-specific beam configurations and optimize the CK treatment process.
format Article
id doaj-art-84bbbacda5d84b0384a2203aee292961
institution OA Journals
issn 1748-717X
language English
publishDate 2024-11-01
publisher BMC
record_format Article
series Radiation Oncology
spelling doaj-art-84bbbacda5d84b0384a2203aee2929612025-08-20T02:38:35ZengBMCRadiation Oncology1748-717X2024-11-0119111510.1186/s13014-024-02531-5Dose prediction of CyberKnife Monte Carlo plan for lung cancer patients based on deep learning: robust learning of variable beam configurationsYuchao Miao0Jiwei Li1Ruigang Ge2Chuanbin Xie3Yaoying Liu4Gaolong Zhang5Mingchang Miao6Shouping Xu7National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeACCURAY, China National Nuclear CorporationDepartment of Radiation Oncology, The First Medical Center of the People’s Liberation Army General HospitalDepartment of Radiation Oncology, The First Medical Center of the People’s Liberation Army General HospitalSchool of Physics, Beihang UniversitySchool of Physics, Beihang UniversityDepartment of Radiation Oncology, the Fourth Hospital of Hebei Medical UniversityNational Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeAbstract Background Accurate calculation of lung cancer dose using the Monte Carlo (MC) algorithm in CyberKnife (CK) is essential for precise planning. We aim to employ deep learning to directly predict the 3D dose distribution calculated by the MC algorithm, enabling rapid and accurate automatic planning. However, most current methods solely focus on conventional intensity-modulated radiation therapy and assume a consistent beam configuration across all patients. This study seeks to develop a more versatile model incorporating variable beam configurations of CK and considering the patient’s anatomy. Methods This study proposed that the AB (anatomy and beam) model be compared with the control Mask (only anatomy) model. These models are based on a 3D U-Net network to investigate the impact of CK beam encoding information on dose prediction. The study collected 86 lung cancer patients who received CK′s built-in MC algorithm plans using different beam configurations for training/validation (66 cases) and testing (20 cases). We compared the gamma passing rate, dose difference maps, and relevant dose-volume metrics to evaluate the model’s performance. In addition, the Dice similarity coefficients (DSCs) were calculated to assess the spatial correspondence of isodose volumes. Results The AB model demonstrated superior performance compared to the Mask model, particularly in the trajectory dose of the beam. The DSCs of the AB model were 20–40% higher than that of the Mask model in some dose regions. We achieved approximately 99% for the PTV and generally more than 95% for the organs at risk (OARs) referred to the clinical planning dose in the gamma passing rates (3 mm/3%). Relative to the Mask model, the AB model exhibited more than 90% improvement in small voxels (p < 0.001). The AB model matched well with the clinical plan’s dose-volume histograms, and the average dose error for all organs was 1.65 ± 0.69%. Conclusions Our proposed new model signifies a crucial advancement in predicting CK 3D dose distributions for clinical applications. It enables planners to rapidly and precisely predict MC doses for lung cancer based on patient-specific beam configurations and optimize the CK treatment process.https://doi.org/10.1186/s13014-024-02531-5Automatic planningDeep learningDose predictionMonte CarloCyberKnife
spellingShingle Yuchao Miao
Jiwei Li
Ruigang Ge
Chuanbin Xie
Yaoying Liu
Gaolong Zhang
Mingchang Miao
Shouping Xu
Dose prediction of CyberKnife Monte Carlo plan for lung cancer patients based on deep learning: robust learning of variable beam configurations
Radiation Oncology
Automatic planning
Deep learning
Dose prediction
Monte Carlo
CyberKnife
title Dose prediction of CyberKnife Monte Carlo plan for lung cancer patients based on deep learning: robust learning of variable beam configurations
title_full Dose prediction of CyberKnife Monte Carlo plan for lung cancer patients based on deep learning: robust learning of variable beam configurations
title_fullStr Dose prediction of CyberKnife Monte Carlo plan for lung cancer patients based on deep learning: robust learning of variable beam configurations
title_full_unstemmed Dose prediction of CyberKnife Monte Carlo plan for lung cancer patients based on deep learning: robust learning of variable beam configurations
title_short Dose prediction of CyberKnife Monte Carlo plan for lung cancer patients based on deep learning: robust learning of variable beam configurations
title_sort dose prediction of cyberknife monte carlo plan for lung cancer patients based on deep learning robust learning of variable beam configurations
topic Automatic planning
Deep learning
Dose prediction
Monte Carlo
CyberKnife
url https://doi.org/10.1186/s13014-024-02531-5
work_keys_str_mv AT yuchaomiao dosepredictionofcyberknifemontecarloplanforlungcancerpatientsbasedondeeplearningrobustlearningofvariablebeamconfigurations
AT jiweili dosepredictionofcyberknifemontecarloplanforlungcancerpatientsbasedondeeplearningrobustlearningofvariablebeamconfigurations
AT ruigangge dosepredictionofcyberknifemontecarloplanforlungcancerpatientsbasedondeeplearningrobustlearningofvariablebeamconfigurations
AT chuanbinxie dosepredictionofcyberknifemontecarloplanforlungcancerpatientsbasedondeeplearningrobustlearningofvariablebeamconfigurations
AT yaoyingliu dosepredictionofcyberknifemontecarloplanforlungcancerpatientsbasedondeeplearningrobustlearningofvariablebeamconfigurations
AT gaolongzhang dosepredictionofcyberknifemontecarloplanforlungcancerpatientsbasedondeeplearningrobustlearningofvariablebeamconfigurations
AT mingchangmiao dosepredictionofcyberknifemontecarloplanforlungcancerpatientsbasedondeeplearningrobustlearningofvariablebeamconfigurations
AT shoupingxu dosepredictionofcyberknifemontecarloplanforlungcancerpatientsbasedondeeplearningrobustlearningofvariablebeamconfigurations