The study aims to develop a computer-assisted decision support based on cerebrospinal fluid (CSF) and blood findings to improve their value and ease the diagnostic procedure of chronic inflammatory diseases (CIDs) of central nervous system (CNS). Data were collected from patients suffering from multiple sclerosis (MS, n = 73), from another CID of the CNS (n = 22), or a psychiatric disease (control group, CTRL, n = 12). Univariate and multivariate analyses were performed using multiple logistic regression and artificial neural networks. Differentiating between MS and CID, no parameter could be disclosed that could provide a meaningful decision support. However, multivariate analysis obtained a statistically significant classification (sensitivity = 84.9%, specificity = 54.5%, p < 0.001). On the contrary, multivariate analysis based on the differentiation between MS vs. CTRL, gave good results (sensitivity = 95.9%, specificity = 83.3%, p < 0.001). It became evident from standard laboratory findings that there is a significant potential for computer-aided decision support.

中文翻译：

---

多发性硬化症的计算机辅助诊断

这项研究旨在基于脑脊液（CSF）和血液检查结果开发计算机辅助决策支持，以提高其价值并简化中枢神经系统（CNS）慢性炎症性疾病（CID）的诊断程序。数据来自多发性硬化症（MS，n  = 73），中枢神经系统的另一CID（n  = 22）或精神病患者（对照组，CTRL，n = 12）。使用多元逻辑回归和人工神经网络进行单因素和多因素分析。区分MS和CID，无法公开任何可以提供有意义的决策支持的参数。但是，多变量分析获得了统计学上显着的分类（敏感性= 84.9％，特异性= 54.5％，p  <0.001）。相反，基于MS与CTRL之间差异的多变量分析得出了良好的结果（灵敏度= 95.9％，特异性= 83.3％，p  <0.001）。从标准实验室的发现中可以明显看出，计算机辅助决策支持的巨大潜力。