PET/CT显像联合融合图像纹理特征分析在肺癌鉴别诊断中的应用价值探讨  

Diagnostic Value of Combination of PET/CT and Texture Features for Lung Cancer

作　　者：王长梅[1,2] 管一晖[1] 张文强[3] 施翔[3] 潘岩[3] 左传涛[1] 华逢春[1] 陈丽敏[1] 

机构地区：[1]复旦大学华山医院PET中心 [2]复旦大学金山医院核医学科 [3]复旦大学计算机科学技术学院

出　　处：《中国医学计算机成像杂志》2010年第2期147-151,共5页Chinese Computed Medical Imaging

基　　金：上海市科委项目(课题编号为08511501300)~~

摘　　要：目的:探讨18F-FDG PET/CT显像医师先验知识、PET/CT融合图像纹理特征参数联合应用对肺癌的诊断价值。方法:130例患者行PET/CT检查,分别由医师先验知识和4项融合图像纹理特征参数(粗糙度、对比度、频度、角二阶矩)判断肺部病灶的良恶性。经病理检查或治疗后影像学随访确诊肺部良恶性病变。计算每种方法及联合应用诊断肺癌的灵敏度和特异性。结果:130例中,77例诊断为肺癌。医师先验知识诊断肺癌的灵敏度为87.0%、特异性为81.1%,纹理特征参数诊断肺癌的灵敏度为80.5%、特异性为71.7%,两者联合的灵敏度为97.4%、特异性为60.4%。结论:基于PET/CT,两者联合应用在判断肺部病灶良恶性中起到积极作用。Purpose：To explore the clinical value of the combination of prior knowledge and texture measurement for detecting lung tumors based on 18F-FDG PET/CT. Methods：One hundred and thirty patients underwent 18F-FDG PET/CT.Physician＇s knowledge and texture features （coarseness,contrast,busyness and angular second moment） of PET/CT fusion images were used to differenciate the malignant from benign lesions. The differences between the two methods and post-operative pathological results or follow-up imagings after treatments were analyzed.Results：Seventy-seven patients were diagnosed with lung cancer.The sensitivity and specificity of the diagnosis obtained with prior knowledge,texture features and the combination of the two methods were 87.0%,81.1%,80.5%,71.7%,97.4%,60.4%,respectively. Conclusion：The result implies that the combination of the two methods may play a positive role for differentiating benign and malignant lesions in future.

关 键 词：体层摄影术 发射型计算机 体层摄影术 X线计算机 图像特征分析 先验知识 

分 类 号：R734.2[医药卫生—肿瘤] TP391.41[医药卫生—临床医学]