急性阑尾炎的多排螺旋CT影像特点比较与机器学习分析  

Comparison and Machine Learning Analysis of Multi-slice Spiral CT Image Features of Acute Appendicitis

作　　者：米华 贺文[2] Mi Hua;He Wen(Department of Radiology,Beijing Xuanwu Hospital of Traditional Chinese Medicine,Beijing 100050;Department of Radiology,Beijing Friendship Hospital Affiliated to Capital Medical University,Beijing 100050)

机构地区：[1]北京市宣武中医医院放射科,北京100050 [2]首都医科大学附属北京友谊医院放射科,北京100050

出　　处：《现代医用影像学》2021年第4期658-661,共4页Modern Medical Imageology

摘　　要：目的:探讨各种类型急性阑尾炎的CT影像表现特点及意义。方法:回顾性分析了87例经手术病理证实为急性阑尾炎的CT影像,卡方检验分析各征象在不同类型急性阑尾炎之间的差异,随机森林法机器学习判断区分化脓性和坏疽性阑尾炎的CT征象的重要性。结果:单纯性阑尾炎、化脓性阑尾炎和坏疽性阑尾炎发生率分别为24.1%,46.0%和29.9%。在三种类型之间,阑尾有无粪石没有显著差异(P>0.05),其他征象均有显著(P<0.05)或极显著(P<0.01)差异。机器学习发现肠壁和肠管外气体积聚、阑尾增强程度减低是坏疽性阑尾炎区别于化脓性阑尾炎的重要征象。结论:多排螺旋CT可以揭示各型急性阑尾炎的形态特点,对选择治疗方式能够提供有价值的信息。Objective:To explore the CT image features and significance of various types of acute appendicitis.Methods:A retrospective analysis of the CT images of 87 cases of acute appendicitis confirmed by operation and pathology.Chi square test was used to analyze the differences of the signs among different types of acute appendicitis.The importance of CT signs in differentiating purulent and gangrenous appendicitis by random forest machine learning.Results:The incidence of simple,purulent and gangrenous appendicitis were 24.1%,46.0%and 29.9%,respectively.Among the three types,there was no significant difference in the presence or absence of fecal stones in the appendix(P>0.05).In other signs,there were significant(P<0.05)or extremely significant(P<0.01)differences.Machine learning found the decrease of the enhancement degree of the appendix were the important signs that distinguish gangrenous appendicitis from purulent appendicitis.Conclusion:Multi-slice spiral CT can reveal the morphological characteristics of various types of acute appendicitis,and can provide valuable information for the choice of treatment.

关 键 词：单纯性阑尾炎 化脓性阑尾炎 坏疽性阑尾炎 随机森林机器学习 X线计算机体层摄影术 

分 类 号：R656.8[医药卫生—外科学]