电子鼻筛查白色念珠菌尿路感染

Screening of Candida albicans urinary tract infections by electronic nose

电子鼻筛查白色念珠菌尿路感染

Afonso, Helga, Escola Superior de Saúde - IPPorto; Faculdade de Medicina da Universidade do Porto, Portugal
Mota, Inês, Instituto de Ciências Biomédicas Abel Salazar, Portugal
Sousa, Ana, Escola Superior de Saúde - IPPorto, Portugal
Vieira, Mónica, Escola Superior de Saúde - IPPorto; Ciências Químicas e das Biomoléculas/CISA, Portugal
Rufo, João C, ISPUP; Faculdade de Medicina da Universidade do Porto, Portugal

Abstract

Fungus caused urinary tract infections (UTIs) are often misdiagnosed. Since culturing procedures are time-consuming, practitioners prophylactically prescribe antibiotics, which are ineffective. Hence, there is a need for reliable methods to detect infectious concentrations of fungi in urine. Since Candida spp is the main responsible for mycological UTI, this study focuses on the detection of candiduria by analysing volatile compounds (VOCs) profiles released by the fungus´ metabolic activity in urine.

The aim of this study was to develop and internally validate a detection algorithm for identifying the presence of pathological levels of Candida albicans in urine, using an electronic nose. To identify the VOCs profiles, the Cyranose 320 (Sensigent, USA) eNose, composed of 32 conducting polymer sensors, was used. Firstly, to optimize the eNose settings, urinary VOCs emissions were tested in terms of substrate heating temperature, as well as acquisition and purging times. Subsequently, 10 glass assay tubes containing urine from a healthy donor and 10 tubes containing urine inoculated with infectious levels (2.3 x 10^7 CFU/mL) of Candida albicans were analysed, in duplicate, with the eNose and resulting data were used to build the detection algorithm through recursive partitioning regression trees. The algorithm was then internally validated and efficacy measurements were retrieved. The Mann-Whitney test was then used to study the hypothesis of sensor 6 (S6) response between the groups.

本研究的目的是开发并内部验证一种检测算法，用于使用电子鼻识别尿液中白色念珠菌的病理水平。为了确定VOCs分布，使用了由32个导电聚合物传感器组成的Cyranose 320（Sensingent，USA）eNose。首先，为了优化eNose设置，根据基质加热温度以及采集和净化时间测试了尿液VOCs排放。随后，对10根含有健康供体尿液的玻璃试管和10根含有感染水平（2.3 x 10^7 CFU/mL）的白色念珠菌的尿液的试管进行了一式两份的分析，并使用eNose和所得数据通过递归分区回归树构建检测算法。然后对算法进行内部验证，并检索疗效测量值。然后使用Mann-Whitney检验研究各组之间传感器6（S6）响应的假设

There was clear differentiation between healthy and infected urine samples (Figure 1). The algorithm reported optimal discrimination of samples using S6 with a cut-off sensor response of 239 x10^-6, with a sensitivity of 85.0%, a specificity of 90.0% and an accuracy of 87.5%. The S6 response was significantly different between groups (p<0.001).

健康和感染的尿液样本之间有明显的区别（图1）。该算法报告了使用S6对样本的最佳识别，截止传感器响应为239 x10^-6，灵敏度为85.0%，特异性为90.0%，准确度为87.5%。各组之间的S6反应显著不同（p<0.001）

In conclusion, this study is promising and, in the future, with further validation using real UTI patients, it may contribute for better diagnosis.

总之，这项研究是有潜力的，在未来，随着使用真实UTI患者的进一步验证，它可能有助于更好的诊断。

*本文内容非商业广告，仅供专业人士参考!