基于气敏传感器阵列的呼气内源性有机挥发物预诊断研究

The VOCs of human breath are complicated. The recent developments of researches find the close relations between these VOCs and liver diseases、kidney diseases、endocrine diseases and several cancers. Comparing with the normal medical diagnosis based on body fluid, the medical diagnosis based on VOCs of breath has several advantages: large amounts of hearth information, detecting more diseases, no intrusive testing, low pain, convenient sampling, rapid data processing, and so on. It has been confirmed that some VOCs can reflect several kidney diseases, and the forming mechanism is comparatively clear. Therefore, kidney disease is chosen as entry point in this report. Main study on pre-diagnosis in endogenous VOCs of breath is about kidney disease. There some features in gas analyzing of pre-diagnosis: complicated components, complicated pathological characteristics of VOCs, synchronized quantitative and qualitative detection, individual difference and Segmentation. According to above characteristics, this paper suggests that the gas sensor array is used for the VOCs analysis and building pathological fingerprint. The main research work will focus on 5 areas: breath acquisition and preprocessing technology optimization, sensitivity of gas sensor array optimization, gas sensitive film shape optimization, gas sensor signal transmitting of review more parameters detection, parameter estimation and pattern recognition of response signals from gas sensor array. I expect some breakthroughs in key technologys of breath VOCS pre-diagnostics based on gas sensor array.

人所呼出的气体混合物所包含的VOCs种类繁多构成复杂，且与肝病、肾病、内分泌系统疾病以及某些癌症具有密切的联系。与基于体液、组织样本的常规医学诊断相比基于呼气VOCS成分的疾病预诊断具有信息量大、识别病症多、非侵入性、低痛苦、采样方便、处理迅速等特点。目前已知的肾病病征气体较多且形成机理较为明晰，因此计划以肾病为切入点进行呼气疾病预诊断研究。疾病诊断的呼气分析具有组分异常复杂、疾病特征呼气成分复杂、同时需要定性定量分析、具有个体差异及分段性等特点。针对这些特点申请书提出基于气敏传感器阵列进行呼气内源性VOCs成分分析构建疾病表征的气体病理指纹。研究工作主要围绕呼气采集与预处理技术的优化、气敏阵列敏感性能的优化、气敏薄膜形态的优化、多参量复合检测的气敏信号变送方案、气敏阵列响应信号的模式识别与参数估计五个方面展开。旨在面向呼气预诊断的气敏传感器阵列技术关键问题上有所突破。

摘要.本项目的主要研究目的是研究针对呼出的气体混合物中所包含的与内源性疾病有密切联系的挥发性有机物的检测与识别手段，从而实现对病症的早期与诊断。主要围绕呼气采集与预处理技术的优化、气敏阵列敏感性能的优化、多参量复合检测的气敏信号变送方案、气敏阵列响应信号的模式识别与参数估计几个方面展开。经过三年的研究，目前项目组围绕相关内容取得了丰富的研究成果：呼吸预处理部分解决了分段呼气采样问题，并经过大量实验确定了呼气样本收集过程中的最优干燥处理方法；通过实验验证，选用卟啉、酞菁结合锌、钴等金属离子组成的金属卟啉制成传感器敏感单元，并在对不同浓度的三甲胺、氨气、苯酚的定性定量识别中取得了较好的识别成果；通过添加碳元素的方法改变敏感薄膜的电导电性，使其在敏感材料与目标气体反应后得到可观测的电学特性改变；在自制传感器阵列基础上采用基于神经网络的识别模型，实现了对多种胺类气体浓度标定试验、通过乳蛋白腐败过程中有机挥发混合物识别分析腐败程度的电子鼻定性定量实验以及对呼气混合物中几种胺类气体的定性定量实验。本项目的研究成果对基于呼气混合物预诊断的后续研究，以及相关成果的实际应用提供了部分关键科学问题的解决方案与探索成果。