胸部食管癌共面调强放射治疗肺受量预测模型研究

Lung dose prediction model in coplanar intensity-modulated radiotherapy for thoracic esophagealcarcinoma

目的:结合胸部食管癌靶区与肺体积和空间解剖位置关系,建立胸部食管癌共面调强放射治疗肺受量预测模型。方法:选取28例胸部食管癌病例,选择其靶区(PGTV,PCTV)处方剂量、靶区体积、肺体积,以及靶区和肺的空间位置关系为输入量,肺受量评价指标(V5,V10,V20,V30及Dmean)作为预测输出量,经主成分分析处理后进行多元非线性回归分析,求出回归方程建立预测模型。结果:肺受量的预测值跟实际值较为接近。其中V5相对误差在-4.37%~2.08%,绝对误差小于2.3%;V10相对误差在-6.53%~1.09%,绝对误差小于2.3%;V20相对误差在-8.99%~7.17%,绝对误差小于1.6%;V30相对误差在-15.93%~11.75%,但绝对误差小于0.9%;Dmean相对误差在-2.73%~2.68%,绝对误差小于31.7 cGy。结论:该模型对胸部食管癌放射治疗中肺受量预测的准确度较高,可作为食管癌放疗计划参数指导及计划评估工具。

Objective To construct a lung dose prediction model in coplanar intensity-modulated radiotherapy for thoracic esophageal carcinoma by comprehensively considering planning target volume and lung volume,and the spatial and anatomical relationships between target areas and lungs.Methods The prescription doses of target areas(PGTV,PCTV),target volume,lung volume,and the spatial and position relationships between target areas and lungs in 28 patients with thoracic esophageal carcinoma were selected as the input,while lung dosimetric indexes(V5,V10,V20,V30 and Dmean)were selected as the predicted output.After principal component analysis,multiple nonlinear regression analysis was carried out for obtaining the regression equation to establish the prediction model.Results The predicted values of lung doses were close to the actual values.The relative errors of V5,V10,V20,V30 and Dmeanwere from-4.37%to 2.08%,-6.53%to 1.09%,-8.99%to 7.17%,-15.93%to 11.75%,and-2.73%to 2.68%,respectively,and the corresponding absolute errors were less than 2.3%,2.3%,1.6%,0.9%and 31.7 cGy,respectively.Conclusion The prediction model which has a high accuracy in lung dose prediction in coplanar intensity-modulated radiotherapy for thoracic esophageal carcinoma can be used as a guidance and planning evaluation tool for esophageal carcinoma radiotherapy planning.