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

目的:探讨~(18)F-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.

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

Purpose: To explore the clinical value of the combination of prior knowledge and texture measurement for detecting lung tumors based on ~(18)F- FDG PET/CT. Methods: One hundred and thirty patients underwent ~(18)F- FDG PET/CT.Physician' s knowledge and texture features (coarseness, con-trast, 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 patholog-ical 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.