Detection of Whole Blood Glucose Concentration by Fourier-Transform Raman Spectroscopy and Artificial Neural Networks
Detection of Whole Blood Glucose Concentration by Fourier-Transform Raman Spectroscopy and Artificial Neural Networks
复制标题
DOI：
10.2991/icaset-17.2017.3
复制
发表时间：
2017-05
期刊：
影响因子：
--
通讯作者：
Qiaoyun Wang;Guangfei Wu;Peng Shan;Zhigang Li;Sheng Hu;Zhenhe Ma
中科院分区：
文献类型：
--
作者：
Qiaoyun Wang;Guangfei Wu;Peng Shan;Zhigang Li;Sheng Hu;Zhenhe Ma
目的：糖尿病是由血糖紊乱引起的一种重要的代谢性疾病。为了避免糖尿病相关的并发症，如失明和肢体丧失，需要调整饮食和胰岛素治疗，每天至少检查四到五次血糖水平。本文的目的是探讨傅里叶变换拉曼光谱作为诊断工具，监测血糖，甘油三酯和总胆固醇水平的潜力。研究方法：从116名糖尿病患者、患者和健康患者中获得拉曼光谱，以便深入了解用于糖尿病诊断的生化变化的测定。在1064 nm激发激光光源下检测人全血。结果：提出了一种改进的样本选择算法（K-Means）KM，并应用于全血的拉曼光谱分析。该算法已成功地应用于提高预测精度。该方法不需要复杂的样品制备过程，就可以直接测定人体全血中的葡萄糖浓度。结论：实验结果表明，拉曼光谱技术作为糖尿病疾病的快速诊断工具具有巨大的临床潜力。
Objectives: Diabetes, which is caused by the disorder of blood glucose, has been one of the most important metabolic diseases worldwide. In order to avoid the diabetes-related complications, such as blindness and loss of limbs, the diet adjustment and insulin therapy are need to check glucose level at least four to five times a day. The aim of this paper was to investigate the potential of Fourier-Transform Raman spectroscopy as a diagnostic tool to monitor the blood glucose, triglycerides and total cholesterol levels. Methods: Raman spectroscopy were acquired from 116 individuals with diabetes patients, patients and healthy patients in order to gain an insight into the determination of biochemical changes for the diabetes diagnose. The human whole blood was examined at 1064nm excitation laser source. Results: A new method with modified sample selection algorithm named (K-Means) KM is developed and applied to the Raman spectroscopy of whole blood. The algorithms have been successfully applied to improve the prediction accuracy. This method can detect the glucose concentration in the human whole blood without complicated sample preparation procedures. Conclusions: The experimental results show that the Raman spectroscopy technology has enormous clinical potential as a rapid diagnostic tool for diabetes diseases.