NIR Diffuse Correlation Spectroscopy And Tomography For Tissue Blood Flow Monitoring And Imaging

Blood flow is an important physiological indicator that advances the understanding of disease.Near-infrared diffuse correlation spectroscopy(DCS)and tomography(DCT)have grown in interest over the past decade as non-invasive techniques for tissue blood flow measurement and imaging using photon autocorrelation signals,DCS/DCT provides It offers many advantages,such as non-invasiveness,versatility,portability and large penetration depth(up to several centimeters).Tissue hemodynamics,including blood flow,oxygenation,and oxygen metabolism,are closely related to many diseases.As one of the optical techniques to study human physiology for disease diagnosis,near-infrared diffuse spectroscopy(NIRS)has been developed for quite some time for the estimation of tissue oxygenation.Recently,a novel NIRS technique,diffuse light correlation spectroscopy and tomography(DCS/DCT),has emerged as a convenient tool for tissue blood flow measurement.This paper first introduces the basic principles and instruments of DCS/DCT,followed by the clinical application examples of DCS/DCT for the diagnosis and treatment of diseases of various organs/tissues,including mental,skeletal muscle,and cancer.The clinical application results demonstrate the advantages and flexibility of DCS/DCT in providing important information for disease diagnosis and treatment evaluation.First,this paper briefly introduces the basic principles and calculations of static NIRS and dynamic NIRS(i.e.,DCS/DCT).Then,the NIRS instrumentation,including hardware and software design,is introduced,the extension from spectroscopy to imaging is described,and recent advances in DCS/DCT are also described in detail.It concludes with a review of future plans for NIRS/DCS and its use in emergency and surgery.For specific applications,we extract the blood flow index(BFI)from the raw photoelectric field data.To represent tissue heterogeneity,we propose a Nth-order linear algorithm(NL),in which the DCS signal is processed by iterative computation.Under the structure of the NL calculation,this paper proposes and evaluates three methods to perform linear iterations to obtain the BFI index.The three techniques are least squares minimization(L2 criterion),least absolute value minimization(L1 criterion),and support vector machine methods(SVR),where L2 criterion is the conventional way to do linear regression.Regression methods developed in recent years for L1 and SVR.We programmed and used photon autocorrelation signals collected from liquid phantom experiments and human tissue to evaluate these three methods.The results show that the BFI noise is removed by the SVR method,and the best effect is obtained,with an error rate of 2.23% under the 3.0 cm light source-detection distance.The L1 standard method gave a moderate error of 2.81%.In contrast,the L2 standard technique achieved the largest error(3.93%)in removing BFI noise.The results obtained from this study will be very useful for tissue blood flow estimation,which is where this paper is innovative.