C反应蛋白
白细胞
医学
内科学
胃肠病学
接收机工作特性
逻辑回归
炎症
作者
Mengjie Yu,Tianxin Xiang,Xiaoping Wu,Shouhua Zhang,Wenlong Yang,Yu Zhang,Qiang Chen,Shuilin Sun,Baogang Xie
标识
DOI:10.1515/cclm-2017-0858
摘要
Abstract Background: The discovery of new metabolic markers may be helpful for early diagnosis of acute pediatric appendicitis (APA). However, no studies have been reported regarding identification of potential metabolic markers for the APA diagnosis by metabonomics. Methods: Serum samples of APA (n=32), non-appendicitis inflammation (NAI, n=32) and healthy children (HS, n=65) were analyzed by the 1 H NMR-based metabonomics. A logistic regression model was established to screen the most efficient markers combinations for classification. Forty double-blind samples were further validated the model. Results: Nine blood metabolites that were different in the APA group from other groups were identified. To differentiate APA from HS, single variable of acetate, formate, white blood cell (WBC) and C-reactive protein (CRP) showed a high diagnostic value (area under the receiver operating characteristic [AUROC]<0.92), while they had a weak diagnostic value (AUROC<0.77) for identifying the APA and NAI. By contrast, the AUROC values of leucine (0.799) were higher than that of WBC and CRP. A combination of five variables, i.e. leucine, lactate, betaine, WBC and CRP, showed a high diagnostic value (AUROC=0.973) for the APA discriminating from the NAI, and the sensitivity and specificity were 93.8% and 93.7%, respectively. Further double-blind sample prediction showed that the accuracy of the model was 85% for 40 unknown samples. Conclusions: The current study provides useful information in our understanding of the metabolic alterations associated with APA and indicates that measurement of these metabolites in serum effectively aids in the clinical identification of APA.
科研通智能强力驱动
Strongly Powered by AbleSci AI