探讨Bayes判别分析在食管癌淋巴结转移CT诊断中的价值

发布时间：2018-05-17 19:49

本文选题：体层摄影术 + X线计算机　；参考：《放射学实践》2015年08期

【摘要】：目的:利用Bayes判别分析初步建立诊断食管癌淋巴结转移的各种CT征象的联合诊断模型。方法:共搜集胸段食管癌208例,将其分为实验组即训练样本(166例,80%)及检验组即验证样本(42例,20%),实验组用以建立诊断方程,检验组用以验证方程。结果:经计算获得的方程为:Y0=-7.499X1+7.957X2+167.761X3-0.087X4+0.459X5-5.528X6-0.711X7-66.080,Y1=-6.697X1+8.231X2+181.686X3-0.106X4+2.219X5-3.331X6-0.124X7-83.183,其中Y0为非转移组,Y1为转移组,X1为淋巴结最大短径,X2为淋巴结最大长径,X3为横纵比,X4为最大截面积,X5为是否边缘模糊,X6为是否中央低密度,X7为是否成簇分布。利用自身检验法所得诊断模型的符合率为87.7%,误判率为12.3%,交叉检验法与自身检验法所得结果相近。当利用验证样本数据代入方程,所得模型诊断符合率为84.7%,误判率为15.3%。结论:通过Bayes判别分析法所建立的不同CT征象对食管癌淋巴结转移的联合诊断模型具有一定的诊断价值,但诊断模型还有待进一步完善。
[Abstract]:Objective: to establish a combined diagnostic model of various CT signs for lymph node metastasis of esophageal carcinoma by Bayes discriminant analysis. Methods: 208 cases of thoracic esophageal carcinoma were divided into two groups: experimental group (166 cases) and test group (42 cases). The experimental group was used to establish the diagnosis equation and the test group to verify the equation. Results: the calculated equation is: 1: Y0- 7.499X1 7.957X2 167.761X3-0.087X4 0.459X5-5.528X6-0.711X7-66.080 8.231X2 181.686X3-0.106X4 2.219X5-3.331X6-0.124X7-83.183, in which Y0 is the non-metastatic group, Y1 is the largest short diameter of lymph node, X2 is the largest length of lymph node, X3 is transverse and longitudinal, and X4 is the largest cross-sectional area. X 6 is whether the central low density X 7 is clustered or not. The coincidence rate and misjudgment rate of the diagnostic model obtained by self-test method were 87.7and 12.3 respectively. The results obtained by cross-test method were similar to those obtained by self-test method. When the validation sample data is substituted into the equation, the diagnostic coincidence rate of the model is 84.7, and the misjudgment rate is 15.3. Conclusion: the different CT signs established by Bayes discriminant analysis have certain diagnostic value for the combined diagnosis model of lymph node metastasis of esophageal carcinoma, but the diagnostic model still needs to be further improved.
【作者单位】：北京协和医学院中国医学科学院肿瘤医院影像诊断科;
【基金】：国家自然科学基金资助项目(61372192)
【分类号】：R735.1;R730.44

【参考文献】