الملخص
Objective:To explore the feasibility and validity of constructing an intensity-modulated radiotherapy gamma pass rate prediction model after combining the SHAP values with the extreme gradient boosting tree (XGBoost) algorithm feature selection technique, and to deliver corresponding model interpretation.Methods:The dose validation results of 196 patients with pelvic tumors receiving fixed-field intensity-modulated radiotherapy using modality-based measurements with a gamma pass rate criterion of 3%/2 mm and 10% dose threshold in Hunan Provincial Tumor Hospital from November 2020 to November 2021 were retrospectively analyzed. Prediction models were constructed by extracting radiomic features based on dose files and using SHAP values combined with the XGBoost algorithm for feature filtering. Four machine learning classification models were constructed when the number of features was 50, 80, 110 and 140, respectively. The area under the receiver operating characteristic curve (AUC), recall rate and F1 score were calculated to assess the classification performance of the prediction models.Results:The AUC of prediction model constructed with 110 features selected based on the SHAP-valued features was 0.81, the recall rate was 0.93 and the F1 score was 0.82, which were all better than the other 3 models.Conclusion:For intensity-modulated radiotherapy of pelvic tumor, SHAP values can be used in combination with the XGBoost algorithm to select the optimal subset of radiomic features to construct predictive models of gamma pass rates, and deliver an interpretation of the model output by SHAP values, which may provide value in understanding the prediction by machine learning-dependent models.
الملخص
Objective:To realize the automatic evaluation of the target conformity, dose overflow, dose drop and other indicators in stereotactic body radiation therapy (SBRT) plan for the early non-small cell lung cancer (NSCLC) in the Eclipse planning system.Methods:Eclipse Scripting API application development interface and C# programming language were used to develop it with script plug-ins combined with independent programs. According to the requirements of the NSCLC SBRT plan in Radiation Therapy Oncology Group (RTOG) 0915 report, the automatic evaluation of related indicators was realized. A rule of equal interval sampling and double thresholds was designed during the period, which could more accurately and conveniently evaluate the dose drop outside the target area.Results:13 clinical cases of NSCLC SBRT plans were randomly selected, the method in this study and the module equipped with the Eclipse system to were utilized to evaluate and compare the results. It was concluded that the results of the two methods were in line with the clinical requirements, and the former was more efficient ( P<0.05). Conclusions:The NSCLC SBRT plan evaluation software in this article has a friendly interface. It can not only realize the automatic evaluation of target conformity, dose overflow and dose drop, but also effectively improve work efficiency and better serve the clinical and scientific research work of radiotherapy.
الملخص
Objective:To realize the interactive form evaluation of radiotherapy dose and the automatic calculation of conformity index (CI) and heterogeneity index (HI) in the Eclipse planning system.Methods:The Eclipse Scripting API application development interface and C# programming language were employed to develop it with script plug-ins combined with independent programs. The visual interface programming and call the relevant dose query function in the API were utilized to realize the interactive table dose evaluation and the automatic calculation of HI. The functions of calling dosage structure creation and structure Boolean operation in the API were adopted to realize the automatic calculation of CI.Results:15 clinical radiotherapy cases of nasopharyngeal carcinoma were selected. The dose evaluation results were statistically compared between this method and the module equipped in the Eclipse system. The results showed that the accuracy was consistent between these two methods ( P>0.05), whereas the evaluation efficiency of this method was significantly higher compared with that of the Eclipse system module ( P<0.05). Conclusions:The interactive evaluation form in this article has a friendly interface, which allows users to more conveniently perform dose assessment, multi-plan comparison, and calculation of CI and HI, which effectively improves work efficiency and can better serve the clinical and scientific research work of radiotherapy.