DeepViewer软件自动勾画乳腺癌保乳术后肿瘤靶区的初步研究  

The Initial Study on Automatic Delineation of Gross Target Volume after Breast-Conserving Surgery by DeepViewer Software

作　　者：田书畅 时飞跃[2,3] 赵紫婷 王敏[2] 秦伟[2] 吴倩倩[1] 秦航[1] 蒋红兵[1,4] TIAN Shuchang;SHI Feiyue;ZHAO Ziting;WANG Min;QIN Wei;WU Qianqian;QIN Hang;JIANG Hongbing(Department of Medical Equipment,Nanjing Hospital Affiliated to Nanjing Medical University(Nanjing First Hospital),Nanjing Jiangsu 210006,China;Department of Tumor Radiotherapy,Nanjing Hospital Affiliated to Nanjing Medical University(Nanjing First Hospital),Nanjing Jiangsu 210006,China;Department of Medical Physics,Nanjing Medical University,Nanjing Jiangsu 210029,China;Nanjing Emergency Medical Center,Nanjing Jiangsu 210003,China)

机构地区：[1]南京医科大学附属南京医院(南京市第一医院)医疗设备处,江苏南京210006 [2]南京医科大学附属南京医院(南京市第一医院)肿瘤放疗中心,江苏南京210006 [3]南京医科大学医学物理研究中心,江苏南京210029 [4]南京市急救中心,江苏南京210003

出　　处：《中国医疗设备》2020年第12期29-32,共4页China Medical Devices

基　　金：南京市医学科技发展资金“青年工程”人才培养专项经费资助项目(QRX11033);南京医科大学科技发展基金(NMUB2018338)。

摘　　要：目的评估DeepViewer软件自动勾画乳腺癌保乳术后患者肿瘤靶区的可行性和准确性。方法选取20例乳腺癌保乳术后患者进行研究,其中左侧乳腺癌12例,右侧乳腺癌8例。将20例患者的定位CT图像从Eclipse治疗计划系统传输至DeepViewer软件,使用DeepViewer软件自动勾画靶区,然后将勾画好的结构文件传输并导回Eclipse治疗计划系统。比较该软件自动勾画和手工勾画靶区的体积差异、形状一致性和勾画时间,并进行统计学分析,评价DeepViewer软件对放疗靶区勾画的可行性和准确性。结果对20例乳腺癌患者的数据进行统计分析。自动和手工勾画两种方式心脏结构的百分体积差异(△V%)为(-21.43±15.29)%,相似性指数(DSC)值为(0.68±0.11),总位置差异(ΔL)为(1.76±1.23)cm。独立样本t检验的结果显示,左右两侧数据没有统计学差异(P>0.05)。结论使用DeepViewer软件进行乳腺癌保乳术后靶区的勾画,可以达到基本满意的效果,后期还有待进一步优化和完善。Objective To evaluate the feasibility and accuracy of DeepViewer software in automatically delineating gross target volume after breast-conserving surgery for breast cancer patients.Methods Twenty patients with breast cancer after breastconserving were selected for study,including 12 cases of left breast cancer and 8 cases of right breast cancer.The localized CT images of 20 patients were transferred from the Eclipse treatment planning system to the DeepViewer software,and the target areas were automatically delineated by using DeepViewer software.Then transmitted the sketched structure files back to the Eclipse treatment planning system.The difference of volume,shape consistency,and sketching time of the target areas between automatically delineating and manually delineating were compared,and statistical analysis was performed to evaluate the feasibility and accuracy of DeepViewer software in sketching the radiotherapy target area.Results Statistical analysis was performed on the data of 20 breast cancer patients.The percentage volume difference(△V%)of the heart structure between the automatic and manual sketching methods was(-21.43±15.29)%,the similarity index(DSC)value was 0.68±0.11.The total position difference(ΔL)was(1.76±1.23)cm.The results of independent sample t test showed that there was no statistical difference between the two side groups of data(P>0.05).Conclusion Using DeepViewer software to sketch the target area after breast-conserving,the result is basically satisfactory,which needs further optimization and improvement.

关 键 词：放疗 乳腺癌 保乳术 靶区 自动勾画 肿瘤靶区 百分体积差 相似性指数 勾画时间 

分 类 号：R730.56[医药卫生—肿瘤]