中枢神经系统CT影像诊断专家系统的设计与实现

Design and Implementation of the Central Nervous System CT Imaging Diagnosis Expert System

目的:探讨专家系统在辅助中枢神经系统疾病的CT影像诊断中的价值。方法:根据影像诊断观察习惯,选取20个指标,建立中枢神经系统影像诊断知识库,并根据公认的发生概率设置各征象的初始值。采用VFP9.0数据库语言编程。选取三甲医院有完整临床和影像学资料并经病理证实的术前误诊病例共173例,分别由两名三甲医院副主任医师(第一组)共同阅片,根据经验进行讨论并达成一致,作出诊断;由一名副主任医师(第二组)和一名住院医师(第三组)分别将上述病例的相关信息输入专家系统,记录所得结果。各组间阅片结果采用字2检验。结果:第一组诊断准确率为37.57%(65/173),第二组诊断准确率为63.01%(109/173),第三组诊断准确率为46.89%(81/173)。经统计学处理表明,第二组与另外两组间差异有统计学意义(P<0.05)。第一组与第三组间差异无统计学意义(P>0.05)。结论:将专家系统融入中枢神经系统CT影像诊断具有重要的实用价值。

Objective： To evaluate the clinical value of expert system of computer, which assisted by CT diagnosis in the central nervous system diseases. Method： According to imaging diagnosis observation habits, 20 indicators were selected to establish the central nervous system imaging diagnosis library,and the initial values of the signs were seted in accordance with accepted probability of occurrence. The software of computer which assisted CT diagnosis in the central nervous system diseases was made by VFP9.0.173 misdiagnoses were selected from grade three and first-class hospitals which were proved by surgery and pathology. Two senior radiologists worked together as the first group, the CT images without the software were analyzed. The relative data were puted into the software by a senior radiologist as the second group and a junior radiologist as the third group,then the results were geted from them. The statistical method 字2 test was used in different group. Result： Diagnostic accuracy of the first group was 37.57%（65/173）, the second group was 63.01%（109/173）, and the third group was 46.89%（81/173）. Statistical analysis showed that between the second group and other two groups the difference were statistically significant（P〈0.05）.Between the first group and the third group, the difference was no statistically significant（P〉0.05）.Conclusion： Expert system has important and practical value with the central nervous system CT imaging diagnostic.