肺部孤立性占位CT灌注与常规和计算机辅助诊断分析  

The Evaluation of CT Perfusion,Routine and Computer Aided Diagnosis of Solitary Pulmonary Occupation

作　　者：顾艳[1] 周胜利[1] 袁刚[1] 黄连庆[1] 

机构地区：[1]徐州医学院附属连云港市第一人民医院放射科,江苏连云港222002

出　　处：《临床放射学杂志》2013年第7期963-967,共5页Journal of Clinical Radiology

基　　金：江苏省自然科学基金(SBK201122841)资助项目

摘　　要：目的探讨应用计算机辅助诊断(CAD)、常规形态学诊断和CT灌注对鉴别肺部孤立性占位良恶性的诊断价值。方法选择同时行CT平扫和CT灌注,并经临床和手术病理证实的肺部孤立性占位100例,首先将形态学征象进行分类统计分析,运用CAD中的最大似然判别法,将其各种征象转化为计分值,以分值的大小来判定肺结节所属的类型;然后分析良恶性占位CT灌注参数,评价灌注参数的诊断试验效能,并将CAD、常规形态学诊断和CT灌注的诊断正确率进行比较。结果 CAD对肺内孤立性占位的诊断正确率(总符合率)为80%,高于常规形态学诊断正确率(77%),但两组间差异无统计学意义(P=0.606);最能提示为恶性肿瘤的征象依次为空泡征、深分叶征、胸膜凹陷征、棘突征、血管集束征和毛刺征。最能提示为良性占位的征象为钙化。CT灌注对肺内孤立性占位的诊断正确率达91%,明显高于CAD和常规形态学诊断,并具有统计学意义(P值分别为0.027和0.007)。结论 CAD对肺内孤立性占位的鉴别诊断正确率与常规形态学诊断法相当,可用于指导日常阅片工作,并可以弥补低年资医师经验的不足;CT肺灌注成像技术能准确地反映肿块内部血管特征,稳定性好,较CAD正确率更高。Objective To evaluate the diagnostic values of computer aided diagnosis（ CAD）, routine diagnostic method and CT perfusion in differentiating the solitary benign or malignant pulmonary occupations. Methods The clinical data of 100 cases with solitary pulmonary occupations confirmed by pathology were retrospectively analyzed, all cases received CT plain scan and perfusion. The HRCT manifestations were studied statistically with the discrimination method of large log likelihood as one of the methods of CAD. CT signs of the 2 kinds of solitary pulmonary occupations were analyzed and then translated them into probability scores to identify the occupations category. Subsequently,perfusion parameters of BF（blood flow） , BV（ blood volume） , MTT（ mean transit time） , PS （capillary permeability surface area product） were compared be- tween benign and malignant occupations, and the efficiency of these perfusion parameters were evaluated as diagnostic tests. The whole diagnostic accuracy of the discrimination methods of CAD, routine diagnostic method and CT perfusion were compared. Results The whole diagnostic accuracy with discrimination method of CAD was 80% , which was higher than that（77% ） of routine diagnostic method, but there was no statistical difference （P = 0.606）. The most probable CT signs for lung cancer were vacuole sign, deep lobulated sign, pleural indentation, spinous process sign, vascular convergence and spicule sign; The most probable CT sign for benign occupations was calcification; The whole diagnostic accuracy with dis- crimination method of CT perfusion was 91% , which was higher than that of CAD and routine diagnostic method, and both of them were statistically difference（P = 0. 027, P = 0. 007）. Conclusion Comparing with the routine diagnostic method, CAD resuks have similar accurate rate for the 2 kinds of solitary pulmonary occupations, and can be used to guide daily work. However, CT perfusion can reflect the dynamic blood flow characteristics of solitary pulmonar

关 键 词：肺癌 肺部孤立性占位 计算机辅助诊断 最大似然法 体层摄影术 X线计算机 

分 类 号：R816.41[医药卫生—放射医学]