摘要
目的应用^(18)F脱氧葡萄糖(^(18)F-FDG)正电子发射计算机断层扫描(PET/CT)图像勾画食管癌原发肿瘤区(GTV),探讨不同方法勾画的准确性。方法选取10例未手术的胸段食管癌患者的PET/CT,分别采用标准摄取值(SUV)2.5、最大SUV(SUV_(max))的40%、PET-edge以及医生通过视觉观察法手动勾画四种方法勾画食管癌GTV,分别命名为GTV2.5、GTV40%、GTV_(edge)以及GTV_(man)。以GTV_(man)作为标准,采用GTV的体积、Hausdorff距离(HD)以及形状相似性指数(DSC)3个指标评价自动识别与手动勾画的几何学一致性。结果与GTV_(man)体积相比,所有患者的GTV2.5均偏大(24.90±7.61)%,所有患者的GTV40%均偏小(26.81±9.63)%,GTV_(edge)有的偏大有的偏小,差异为(12.51±14.28)%(均值使用偏差比率的绝对值)。三种勾画方式的DSC分别为0.82、0.78、0.85,GTV40%与GTV_(edge)勾画的DSC有统计学差异(P=0.015),其余对比无统计学差异。HD分别为6.87mm、6.92mm、6.25mm,差异均无统计学意义。结论使用2.5的阈值以及PET-edge自动识别可用于PET/CT勾画食管癌GTV,推荐二者结合使用指导GTV的勾画,满足临床应用的同时,提高勾画的准确度和效率。
Objective To explore the accuracy of different methods for delineating the gross tumor volumes(GTV)of the primary esophageal cancer using PET/CT images.Methods PET/CT images of 10 patients with thoracic esophageal cancer without surgery were selected.The GVTs of esophageal cancer were delineated by standard uptake value(SUV)2.5,maximum SUV(SUV_(max))40%,PET-edge,and manual delineation by visual,named GTV2.5,GTV40%,GTV_(edge)and GTV_(man)respectively.Using GTV_(man)as the standard,three indexes:volume of GTV,distances of Hausdorff(HD)and dice similarity coefficient(DSC)were used to evaluate the geometric consistency between automatic identifications and manual delineations.Results Compared with the volume of GTV_(man),GTV2.5 were larger in all patients(24.90±7.61)%,GTV40%were smaller in all patients(26.81±9.63)%,and GTV_(edge)were larger in some patients and smaller in others(12.51±14.28)%(absolute value of the deviation ratio of the mean).The DSC of the three delineation methods were 0.82,0.78,0.85,which delineated by GTV40%and GTV_(edge)were statistically different(P=0.015),and the rest have no statistically differences.HD was 6.87mm,6.92mm and 6.25mm,the differences were not statistically significant.Conclusion Standard uptake value(SUV)2.5 and automatic PET-edge recognition can be used to delineate esophageal cancer GTV in PET/CT images.It is recommended that the two methods should be combined to guide the delineation of GTV,which can meet the needs of clinical application,and improve the accuracy and efficiency of delineation.
作者
石鑫磊
葛小林
狄晓珂
王沛沛
SHI Xin-lei;GE Xiao-lin;DI Xiao-ke;WANG Pei-pei(Department of Radiology,The First Affiliated Hospital of Nanjing Medical University,Nanjing 210029,Jiangsu Province,China;Center of Radiation Oncology,The First Affiliated Hospital of Nanjing Medical University,Nanjing 210029,Jiangsu Province,China)
出处
《中国CT和MRI杂志》
2023年第7期77-79,共3页
Chinese Journal of CT and MRI
基金
国家自然科学基金(82003228)。