A feasibility study of computer aided diagnosis in appendicitis.
A feasibility study of computer aided diagnosis in appendicitis.
复制标题
阑尾炎计算机辅助诊断的可行性研究。
DOI：
--
复制
发表时间：
1982
期刊：
影响因子：
--
通讯作者：
Elerding Sc
中科院分区：
文献类型：
--
作者：
Van Way Cw rd;Murphy;Dunn El;Elerding Sc
在这项研究中，假设计算机辅助诊断可以更准确地区分急性阑尾炎患者和腹痛但阑尾正常的患者。通过分析五年内476例急诊阑尾切除术患者的资料，建立了数据库。急性阑尾炎360例，占76%，正常阑尾98例，占20%，其他疾病18例，占4%。分析记录中的病史、体格检查和实验室检查结果。然后将数据库随机分为两部分。第1部分采用卡方检验进行单变量判别分析。阑尾炎和正常阑尾之间唯一显著不同的是性别、症状持续时间、厌食和呕吐。多变量判别分析被用来获得一个腹痛指数，区分阑尾炎和正常阑尾的敏感性为0.82和特异性为0.39。使用腹痛指数来评估数据库第2部分中的患者，58名阑尾正常的患者中有23或40%可以避免手术。然而，在169名阑尾炎患者中，有31人或18%的人没有接受手术;这31人中有3人阑尾穿孔。阑尾炎的计算机辅助诊断并不比临床诊断更有效。
In this study, the hypothesis that computer aided diagnosis could enable a more accurate differentiation between patients with acute appendicitis and those with abdominal pain but normal appendixes was examined. A data base was established by analyzing the records of 476 patients having an emergency measure appendectomy during a five year period. There were 360 or 76 per cent with acute appendicitis, 98 or 20 per cent with normal appendixes and 18 or 4 per cent with other diseases requiring operation. The records were analyzed with regard to history, physical examination and laboratory findings. The data base was then divided randomly into two parts. Part 1 was subjected to univariate discriminant analysis, using the chi-square test. The only quantities which were significantly different between appendicitis and a normal appendix were sex, duration of symptoms, anorexia and vomiting. Multivariate discriminant analysis was used to derive an abdominal pain index which discriminated between appendicitis and a normal appendix with a sensitivity of 0.82 and a specificity of 0.39. Using the abdominal pain index to evaluate the patients in part 2 of the data base, 23 or 40 per cent of the 58 patients with a normal appendix would have avoided operation. However, 31 or 18 per cent of the 169 patients with appendicitis would have not been operated upon; three of those 31 had perforated appendixes. Computer aided diagnosis was no more effective than unaided clinical diagnosis in appendicitis.