The Studies Of Noninvasive Blood Glucose Monitoring Using Optical Coherence Tomography And Factors Affecting Its Accuracy

Optical coherence tomography(OCT) is a noninvasive imaging technique with high-resolution at micrometer-scale spatial resolution and millimeter-scale imaging depth. However, the technique of using OCT for noninvasive blood glucose monitoring has not come into clinical application yet, because there are multiple complex physiology and environment factors which can influence the optical properties of measured tissues. Based on this, the methods on how to improve the accuracy of noninvasive blood glucose measurement against the influencing factors are studied in this thesis.Firstly, this dissertation investigates the effects of contact pressure on the skin from the OCT probe, the location of focal point of the optical beam, and the motion artifacts on the accuracy of noninvasive blood glucose monitoring. The algorithm that uses correlation analysis to calculate the depth regions which have the maximum correlation between blood glucose concentration and tissue optical properties is presented. The selected correlation regions also have the least noise interference from other skin layers. This method can improve the prediction accuracy of noninvasive blood glucose concentration monitoring.To investigate the lag of glucose concentration between interstitial fluid of human skin and blood(peripheral blood and venous blood) during rapid blood glucose variations, the OCT was used to measure the changes of attenuation coefficient of human skin caused by the change of blood glucose in oral glucose tolerance tests and glucose clamp tests in vivo. It is demonstrated experimentally that the lag time changes at different depths in human skin. A decrease of lag time with increased depth can be seen. The knowledge of such lag time of blood glucose can improve the accuracy and reliability of measured value in noninvasive blood glucose sensing.The effect of temperature on optical attenuation coefficients of in vivo human skin in different layers is also studied in this paper. A temperature control module with an integrated optical probe to precisely control the surface temperature of a section of human skin is designed. The attenuation coefficient closely follows temperature variations without hysteresis. A negative thermal coefficient of attenuation is observed in the epidermis of human skin. As for dermis, the signs of thermal coefficient are found to be different at different depths because of the variation of tissue structure. The correlation distributions with depth need to be calculated by correlation analysis algorithm.The influence of temperature on noninvasive glucose monitoring of in vivo human skin with OCT using multiple linear regression(MLR) model was investigated. On average, the changes in attenuation coefficient per 1°C of temperature will lead to 0.26 mmol/L prediction error of blood glucose concentration. An algorithm of temperature compensation based on correlation analysis is proposed to improve the accuracy of predicted glucose value. In addition, a meathod of rapid blood glucose calibration is also presented in this paper. The depth regions which attenuation coefficients have both high correlation with temperuature and blood glucose can be found with rapid temperutre modulation. These high correlation regions can be used to predict blood glucose concentration in the subsequent steps. The calibrated meathod can significantly shorten the calibration duration and relieve the pain of diabetics to some extent.