CBCT影像组学联合构建Nomogram模型预测食管癌放疗患者放射性肺炎

A nomogram model based on cone beam CT radiomics combined with clinical features and dosimetric parameters predicting radiation pneumonitis in patients with esophageal cancer receiving radiotherapy

目的通过对放疗疗程中不同时段CBCT图像的影像组学分析,寻找早期定量预测食管癌放疗放射性肺炎(RP)的参数,结合临床特征和肺剂量体积参数建立联合Nomogram模型并探讨这一模型对食管癌RP的预测价值。方法回顾分析2017—2019年间临床资料、剂量学参数、CBCT图像资料完整的96例胸中段食管鳞癌调强放疗患者资料,每例患者均分别获取放疗期间3个不同时段的肺CBCT图像。全组病例随机分成训练集(67例)和验证集(29例),以CBCT上双肺实质作为感兴趣区,运用3D-Slicer软件进行图像分割和特征提取,经LASSO-Logistics回归分析方法进行特征参数筛选并构建影像组学标签(Rad-score)。从3个不同时段建立的RP预测模型中选择最优模型联合经回归分析获得的最佳临床及剂量学参数,建立联合Nomogram模型,并进行受试者工作特征曲线分析,基于曲线下的面积(AUC)验证其诊断效能。结果第一时段的影像组学预测模型优于其他两个时段,在训练集中的AUC值为0.700(95%CI为0.568~0.832),敏感性和特异性分别为61.5%、75.0%;在验证集中的AUC值为0.765(95%CI为0.588~0.941),敏感性和特异性分别为84.6%、64.7%。影像组学联合临床及剂量学构建的Nomogram模型在训练集中的AUC值为0.836(95%CI为0.700~0.918),敏感性和特异性分别为96.0%、54.8%;在验证集中的AUC值为0.905(95%CI为0.799~1.000),敏感性和特异性分别为92.9%、73.3%。联合Nomogram模型诊断效能最佳。结论基于放疗早期肺CBCT影像组学特征构建的模型,对于食管癌RP具有一定的预测效能,Rad-score联合肺V5Gy、肺Dmean、肿瘤分期建立的Nomogram模型具有更好的预测准确性,可作为一种定量预测模型用于RP的预测。

Objective To develop and validate a nomogram model for predicting radiation-induced pneumonitis in esophageal cancer based on CBCT radiomics characteristics combined with clinical characteristics and lung dosimetric parameters.Methods Clinical data,dosimetric parameters and CBCT images of 96 patients with thoracic middle esophageal squamous cell carcinoma treated by intensity-modulated radiation therapy(IMRT)from 2017 to 2019 were analyzed retrospectively.The CBCT images of each patient in three different time periods were obtained.All patients were assigned randomly into the primary cohort(n=67)and validation cohort(n=29).Double lungs were selected as the region of interest(ROI),and 3D-slicer software was used for image segmentation and feature extraction.The LASSO regression were applied to identify candidate radiomic features and construct the Rad-score.The optimal time period,clinical and dosimetric parameters were selected to construct the nomogram model,and then the area under the receiver operating characteristic curve(AUC)was used to evaluate the prediction effect of the model.Results The predictive capacity of the model in the first time period was the highest.In the primary cohort,the AUC was 0.700(95%CI:0.568-0.832),the sensitivity was 61.5%,and the specificity was 75.0%.In the validation cohort,the AUC was 0.765(95%CI:0.588-0.941),the sensitivity was 84.6%and the specificity was 64.7%,respectively.In the combined nomogram model,the AUC in the primary cohort was 0.836(95%CI:0.700-0.918),the sensitivity was 96.0%and the specificity was 54.8%.In the validation cohort,the AUC was 0.905(95%CI:0.799-1.000),the sensitivity was 92.9%and the specificity was 73.3%,respectively.The diagnostic efficiency of combined nomogram model was the best.Conclusions The nomogram model based on early lung CBCT radiomics has certain predictive efficiency for RP.The model of lung CBCT radiomics in early stage of radiotherapy can predict RP of esophageal cancer.The nomogram model based on Rad-score combined with V5Gy,MLD and tumor stage yields better predictive accuracy,which can be used as a quantitative prediction model for RP.