江苏大学学报（医学版）2025，Vol.35Issue(4)：343-350,360,9.DOI:10.13312/j.issn.1671-7783.y240227
基于CT影像组学诺莫图模型预测食管鳞癌的放疗效果
Prediction of radiotherapy efficacy for esophageal squamous cancer using a CT-based radiomics nomogram model
孙静宇 1孙洋霖 2朱伶可 2倪青青 1张利利 2巩萍 3王强4
作者信息
- 1. 江苏大学医学院,江苏 镇江 212013
- 2. 徐州医科大学第一临床医学院,江苏 徐州 221004
- 3. 徐州医科大学医学影像学院,江苏 徐州 221004
- 4. 江苏大学医学院,江苏 镇江 212013||徐州市肿瘤医院放疗科,江苏 徐州 221005
- 折叠
摘要
Abstract
Objective:To develop and validate a nomogram model utilizing CT radiomics features for predicting the efficacy of radiotherapy in esophageal squamous cancer.Methods:A retrospective analysis was conducted for 201 patients diagnosed with esophageal squamous cancer at Xuzhou Cancer Hospital and Xuzhou Mining Group General Hospital between January 2019 and January 2024.The dataset was partitioned into the training set(n=79),the testing set(n=34),and the external validation set(n=88).Radiomics features were extracted from pre-radiotherapy CT images of patients,followed by dimensionality reduction.Univariate and multivariate Logistic regression analyses were employed to identify the significant clinical features of the constructed model.The support vector machine learning algorithm was utilized to construct three predictive models:the clinical model,the radiomics model,and the nomogram model.The nomogram model was visualized using a nomogram.The predictive performance of the constructed model was assessed through external validation.The model's clinical applicability were measured by dicision curve analysis(DCA).Results:A total of 1 834 CT radiomics features were extracted.Following feature screening and dimensionality reduction,eleven optimal radiomics features were selected for subsequent model construction.The area under curve(AUC)for the nomogram model in the training set was 0.980,which exceeded that of both the clinical model(AUC=0.859)and the radiomics model(AUC=0.926).In the testing set,the AUC for the nomogram model was 0.834,surpassing the clinical model(AUC=0.638)and the radiomics model(AUC=0.818).Furthermore,in the external validation set,the AUC for the nomogram model was 0.883,outperforming the clinical model(AUC=0.652)and the radiomics model(AUC=0.765).DCA results demonstrated that the nomogram model had the highest clinical net benefit in comparison to the alternative models.Conclusion:The developed nomogram model,which integrates CT radiomics and clinical features,may be applied for predicting the efficacy of radiotherapy for esophageal squamous cancer.关键词
食管癌/影像组学/放射治疗/诺莫图模型/CTKey words
esophageal cancer/radiomics/radiotherapy/nomogram model/CT分类
医药卫生引用本文复制引用
孙静宇,孙洋霖,朱伶可,倪青青,张利利,巩萍,王强..基于CT影像组学诺莫图模型预测食管鳞癌的放疗效果[J].江苏大学学报（医学版）,2025,35(4):343-350,360,9.
