An implantable sensor is being created that allows measurement of blood glucose through fluorescent detection of an embedded chemical assay. The sensor is based on the competitive binding reaction between the protein Concanavalin A and various saccharide molecules, specifically a glycodendrimer and glucose. Previous studies have shown the ability of an embedded chemical assay using Con A and dextran with shorter wavelength dyes to both sense changes in glucose and generate sufficient fluorescent emission to pass through the dermal tissue. However, due to the chemical constituents of the assay, multivalent binding was evident resulting in poor spectral change due to glucose within the biological range. Use of a glycodendrimer and longer wavelength dyes has improved the sensor’s spectral change due to glucose and the overall signal to noise ratio of the sensor. In this work, a description of this sensor and the results obtained from it will be presented showing a large dynamic range of fluorescence with glucose.

A glucose affinity sensor based on a homogeneous fluorescence resonance energy transfer (FRET) assay system was developed to monitor the competitive binding between concanavalin A (con A), a sugar-binding protein labeled with acceptor fluorophore, Alexa Fluor 647 (AF647) and polysaccharides conjugated with donor fluorophore, Alexa Fluor 568 (AF568). Confounding factors such as: (i) the impact of scattering due to tissue optical properties; (ii) the reabsorption of propagated donor fluorescence by the acceptor fluorophore; (iii) photobleaching; and (iv) fluorophore loading must be accounted for before quantitative glucose measurements can be made from fluorescence intensity measurements. Fluorescence lifetime spectroscopy made in the frequency domain circumvents most of these artifacts by measuring phase-shift and modulation ratio related with the fluorophore lifetime change. Experiments were performed to assess the FRET effects of this affinity sensing system, using dextran (MW 2000K) labeled with donor molecule, AF568 (donor-dextran), and con A labeled with acceptor molecule, AF647 (acceptor- con A). Herein we demonstrate that the FRET decay kinetics can indicate changes in the competitive binding of 0.09 μM dextran as con A concentration (from 0 to 10.67 μM) and glucose concentration (from 0 to 224 mg/dL) are changed. Preliminary work also presented here shows that quantitative frequency-domain lifetime measurement of FRET changes could be achieved in tissue-like scattering media using the photon diffusion equation.

The increasing prevalence of diabetes in the United States has led many to pursue methods for non-invasive glucose detection using various optical approaches such as NIR absorption spectroscopy, Raman spectroscopy, fluorescence spectroscopy, and polarization. Polarization approaches using the aqueous humor as the sensing site have been previously shown to achieve 5 mg/dl accuracy in vitro, however accuracy in vivo has yet to be obtained due to motion induced birefringence changes in the cornea. A dual-wavelength close-looped system was developed to compensate for motion artifact. This method has shown 15 mg/dl accuracy in the presence of birefringence changes in the optical path in vitro similar to those that occur in the cornea -- something previous systems were not capable of doing.

Tight control of blood glucose levels has been shown to dramatically reduce the long-term complications of diabetes. Current invasive technology for monitoring glucose levels is effective but underutilized by people with diabetes because of the pain of repeated finger-sticks and the cost of reagent strips. Optical sensing of glucose could potentially allow more frequent monitoring and tighter glucose control for people with diabetes. The key to a successful optical non-invasive measurement of glucose is the collection of an optical spectrum with a very high signal-to-noise-ratio in a spectral region with significant glucose absorption. Unfortunately, the optical throughput of skin is very small due to absorption and scattering. To overcome these difficulties, we have developed a high-brightness tunable laser system for measurements in the 2.0-2.5 μm wavelength range. The system is based on a 2.3 micron wavelength, strained quantum-well laser diode incorporating GaInAsSb wells and AlGaAsSb barrier and cladding layers. Wavelength control is provided by coupling the laser diode to an external cavity that includes an acousto-optic tunable filter. Tuning ranges of greater than 110 nm have been obtained. Because the tunable filter has no moving parts, scans can be completed very quickly, typically in less than 10 ms. We describe the performance of the laser system and its potential for use in a non-invasive glucose sensor.

Near infrared spectroscopy has been proposed as an effective way for measuring blood glucose non-invasively. However the change of spectrum due to an increase in glucose level is very small compared to the changes due to other variations such as absorption of major blood components, skin surface reflectance, temperature and pressure and so on. So the complexity of spectrum makes it difficult to identify unique glucose information. In this paper, the effect of background correction is discussed firstly. Then a simple substitution is proposed to compute the net analyte signal of glucose using the subspace spanned by the background spectra. For the in vitro experiment, the net analyte signals of glucose using the traditional methods and the subspace spanned by background have the same peaks in the absorption peaks of glucose for the glucose aqueous solution. For in vivo experiment, there is significant spectral difference between the subject who took OGTT test and the subject who took no glucose or water. And the net analyte signal of glucose is computed for OGTT test based on the subspace spanned by the spectra of subject who didn’t take glucose. Results show that, the spectral information induced by glucose taking is quite significant but it does not have the same peak at the absorption peak of glucose in near-infrared region.

We investigate whether acousto-optic spectroscopy can be utilized for a non-invasive quantitative determination of chemical species in human blood or tissue. In a series of preliminary experiments in a model system consisting of a light absorbing structure buried in a light scattering environment we have found that it is possible to extract semi-quantitative information on the absorbance of this absorber. From these pilot experiments we conclude that our approach deserves further investigation.

Experimental methods of coherence-gated time domain imaging (TD-OCT), Doppler OCT and quasielastic laser light scattering are compared in terms of optimal data acquisition and processing. Low coherence methods are applied for flow visualization in hydrodynamic phantoms and in vivo. Low coherence reflection Doppler spectra are compared with laser Doppler spectra. Structural images of in vivo human subcutaneous veins with diameter of about 1 mm are demonstrated before and after optical clearing and raster averaging. Structural images of human thumb nail and ~1 mm in depth tissue underneath in the transitional OCT mode are presented. Low power rapid scanning optical delay line in the reference arm and low numerical aperture in the sample arm of the interferometer, along with tissue optical clearing were applied for increasing transcutaneous coherence probing depth up to 1.5 - 1.6 mm.

A blood perfusion and oxygenation sensor has been developed for in situ monitoring of transplanted organs. In processing in situ data, motion artifacts due to increased perfusion can create invalid oxygenation saturation values. In order to remove the unwanted artifacts from the pulsatile signal, adaptive filtering was employed using a third wavelength source centered at 810nm as a reference signal. The 810 nm source resides approximately at the isosbestic point in the hemoglobin absorption curve where the absorbance of light is nearly equal for oxygenated and deoxygenated hemoglobin. Using an autocorrelation based algorithm oxygenation saturation values can be obtained without the need for large sampling data sets allowing for near real-time processing. This technique has been shown to be more reliable than traditional techniques and proven to adequately improve the measurement of oxygenation values in varying perfusion states.

Near infrared integrating sphere spectroscopy and chemometric multivariate calibration were applied to determine hematocrit (HCT) and oxygen saturation (O2Sat) of circulating human blood. The diffuse reflectance were measured and the partial least square method (PLS) was used for calibration considering different wavelength ranges. The HCT and the O2Sat could be predicted with a root mean square error (PRMSE) of 1.9% and 2.8% respectively, using PLS. Each parameter was adjusted to various levels, and three measurement series from blood of three different donors were carried out for the calibration with the PLS. The calibration includes changes in hemolysis and osmolarity as well as inter-individual differences in cell dimensions and hemoglobin content. Prediction of hemolysis was also possible for one blood sample with a PRMSE of 0.8%.

Mature human red blood cells (RBCs) are light scatterers with homogeneous bodies enclosed by membranes and have attracted significant attention for optical diagnosis of disorders related to blood. RBCs possess viscoelastic structures and tend to deform from biconcave shapes isovolumetrically in blood flow in response to pressure variations. Elastic scattering of light by a deformed RBC provides a means to determine their shapes because of the presence of strong light scattering signals, and development of efficient modeling tools is important for developing bed-side instrumentation. The size parameters α, defined as α=2πα/λ with 2α as the characteristic size of the scatterer and λ as the light wavelength in the host medium, of the scatterer of RBCs are in the range of 10 to 50 for wavelengths of light in visible and near-infrared regions, and no analytical solutions have been reported for light scattering from deformed RBCs. We developed a parallel Finite-Difference-Time-Domain (FDTD) method to numerically simulate light scattering by a deformed RBC in a carrier fluid under different flow pressures. The use of parallel computing techniques significantly reduced the computation time of the FDTD method on a low-cost PC cluster. The deformed RBC is modeled in the 3D space as a homogeneous body characterized by a complex dielectric constant at the given wavelength of the incident light. The angular distribution of the light scattering signal was obtained in the form of the Mueller scattering matrix elements and their dependence on shape change due to pressure variation and orientation was studied. Also calculated were the scattering and absorption efficiencies and the potential for using these results to probe the shape change of RBCs will be discussed.

In the United States, approximately 100 patients develop fatal sepsis associated with platelet transfusions every year. Current culture methods take 24-48 hours to acquire results, which in turn decrease the shelf life of platelets. Many of the microorganisms that contaminate platelets can replicate easily at room temperature, which is the necessary storage temperature to keep platelets functional. Therefore, there is a need for in-situ quality control assessment of the platelet quality. For this purpose, a real time spectrophotometric technique has been developed. The Spectral Acquisition Processing Detection (SAPD) method, comprised of a UV-vis spectrophotometer and modeling algorithms, is a rapid method that can be performed prior to platelet transfusion to decrease the risk of bacterial infection to patients. The SAPD method has been used to determine changes in cell suspensions, based on size, shape, chemical composition and internal structure. Changes in these cell characteristics can in turn be used to determine microbial contamination, platelet aging and other physiologic changes. Detection limits of this method for platelet suspensions seeded with bacterial contaminants were identified to be less than 100 cfu/ml of sample. Bacterial counts below 1000 cfu/ml are not considered clinically significant. The SAPD method can provide real-time identification of bacterial contamination of platelets affording patients an increased level of safety without causing undue strain on laboratory budgets or personnel while increasing the time frame that platelets can be used by dramatically shortening contaminant detection time.

Worldwide, oral cancer is the sixth most common cancer for both sexes. In Singapore, the 5-year survival rate of oral cancer is about 50%. The high mortality rate has been attributed to the difficulties in detecting the disease in an early treatable stage. Currently, the standard screening procedures for oral cancer are histopathology examination of biopsied tissues and exfoliative cytological assessment. These techniques, unfortunately, are low in sensitivity. In this study, we exploit the high amplification factor of SERS to investigate on the possibility of utilising molecular vibrational information from saliva samples to detect oral cancer early. All raw saliva samples were centrifuged at 13,000 krpm for 5 minutes to remove unwanted particles prior to SERS measurements. The purified saliva samples were then applied directly on gold particle films, followed by excitation with a 633 nm HeNe laser. SERS spectrum can be obtained in less than 2 minutes for each sample. We have studied the saliva spectra acquired from 5 normal individuals and 5 patients with oral cancer. In addition, we also observe new peaks at 1097 cm^-1 and 1627 cm^-1 in some of the abnormal samples. These peaks are not present in the spectra acquired from the normal samples. Preliminary measurements will be presented. This study may lead to the development of a sensitive and portable diagnostics system for oral cancer.

Epithelial ovarian cancer has the highest mortality rate among the gynecologic cancers. Early detection would significantly improve survival and quality of life of women at increased risk to develop ovarian cancer. We have constructed a device to investigate endogenous signals of the ovarian tissue surface in the UV C to visible range and describe our initial investigation of the use of optical spectroscopy to characterize the condition of the ovary. We have acquired data from more than 33 patients. A table top spectroscopy system was used to collect endogenous fluorescence with a fiberoptic probe that is compatible with endoscopic techniques. Samples were broken into five groups: Normal-Low Risk (for developing ovarian cancer) Normal-High Risk, Benign, and Cancer. Rigorous statistical analysis was applied to the data using variance tests for direct intensity versus diagnostic group comparisons and principal component analysis (PCA) to study the variance of the whole data set. We conclude that the diagnostically most useful excitation wavelengths are located in the UV. Furthermore, our results indicate that UV B and C are most useful. A safety analysis indicates that UV-C imaging can be conducted at exposure levels below safety thresholds. We found that fluorescence excited in the UV-C and UV-B range increases from benign to normal to cancerous tissues. This is in contrast to the emission created with UV-A excitation which decreased in the same order. We hypothesize that an increase of protein production and a decrease of fluorescence contributions of the extracellular matrix could explain this behavior. Variance analysis also identified fluctuation of fluorescence at 320/380 which is associated with collagen cross link residues. Small differences were observed between the group at high risk and normal risk for ovarian cancer. High risk samples deviated towards the cancer group and low risk samples towards benign group.

The use of small animal models to investigate human diseases is an integral part of the development of new diagnostic and treatment regimens. Consequently, functional imaging modalities such as single photon emission computed tomography (SPECT) are increasingly being utilized to streamline the screening of animal phenotypes and to monitor disease states, progressions, and therapies. This paper focuses on the utilization of polarization filtering to minimize specular reflection from a glass tube used for holding live human-tumor-mice during functional imaging in a dedicated small animal SPECT system. The system presented is potentially useful for the real-time non-invasive investigation of diseases, such as cancer, and drug therapies in small animals because it utilizes optical motion-registered functional imaging that minimizes the effects of motion artifacts.

Continuous wave near-IR spectroscopy (CW-NIRS) has been increasingly applied for the noninvasive, in vivo measurement of tissue and blood chemistry. It is hypothesized that there is a quantifiable relationship between fat thickness and near infrared diffuse reflectance spectra at all wavelengths, and this relationship can be used to remove the spectral influence of the overlying fat layer from the muscle spectrum. The hypothesis was investigated at a single wavelength using Monte Carlo simulations of a two-layer structure and with phantom experiments. The influence of a range of optical coefficients (absorption and reduced scattering) for fat and muscle over the known range of human physiological values was also investigated. A polynomial relationship was established between the fat thickness and the detected diffuse reflectance. It is also shown that the optical properties of the muscle and fat layers influence this relationship under certain conditions. Subject-to-subject variation in the fat optical coefficients and thickness can be ignored if the fat thickness is less than 5 mm, such as on the forearm. If NIRS measurement is to be performed on an anatomical region with a thicker fat layer, a spectral correction for fat will be needed to account for its thickness and the variation in optical coefficients for both the fat and the muscle layers.

In the present paper the problem of protection of human skin against harmful UV solar rays using nano-sized spherical particles of titanium dioxide and sensing their concentration if embedded into skin is considered. Experimental tape-stripping method was used to reveal the in-depth distribution of the particles within the horny layer up to 20 µm. Computer simulations of optical coherence tomography (OCT) investigations of skin and, in particular, horny layer in vitro with and without titanium dioxide particles added were also performed in order to understand, if this modern non-invasive technique is applicable for skin study and revealing the distribution of nanoparticles within the horny layer. The effect of particles size (25-200 nm) and concentration on simulated OCT signals was analyzed. The increase of scattering in the sample (with increase of particles concentration or size) leads to increase of the OCT signal slope and decrease of rear border peak. We also performed simulations implementing the Monte Carlo technique to evaluate the protecting effect of titanium dioxide nanoparticles of different size. The most effective sizes were revealed. Computations were performed for the wavelength of 290.5 nm as the most harmful one. Dependencies of light intensities absorbed, backscattered, and transmitted through the whole horny layer (20 µm thick) on concentration of titanium dioxide particles (0-5%) were obtained and analyzed.

This paper presents a theoretical model of the effect of the geometry of illumination and collection in fluorescent media, which exhibit self-absorption at sufficiently high concentrations. In order to derive a relation between the incident excitation intensity and the fluorescence emission intensity, we consider the series of paths and transformations that light takes between the source and the detector. The preliminary supporting experiments were conducted on non-turbid liquid fluorescent samples using classical right-angle detection scheme, based on Time-Correlated Single Photon Counting (TCSPC). The fluorescent dyes tested in these experiments (Coumarins 1, 314 and 343) were chosen because they all are excitable at 405 nm, and exhibit varying Stokes shifts. The results suggest that the geometry of the illumination and collection, as well as the self-absorption process, should be taken into account in time-resolved and intensity fluorescence measurements.

We have previously demonstrated the correlation of continuous-wave near infrared (CW-NIR) tissue measurements, to blood and tissue metabolic parameters using Partial Least Squares (PLS) regression. The practical use of this non-invasive measurement technique depends on the transfer of PLS calibration models from a single calibration unit to multiple secondary units. Variations in the spectral characteristics of the optical components across multiple units result in marked differences in the spectral output, preventing the direct transfer of parameter models from one unit to another. Consequently, we have developed a method for standardizing the spectral output across units that utilizes physical, traceable, reference materials for aligning the wavelength and intensity axes to fixed values, followed by spectral normalization via Standard Normal Variate transformation. The approach employed in this study adjusts the slope and bias differences in the optical spectra across multiple units, without the loss of useful information needed for parameter estimation. In this study, phantoms containing Agar, intralipid and lyophilized human hemoglobin (met-hemoglobin) were used to mimic human tissue. Using PLS regression, a hemoglobin calibration model was developed on the tissue-like phantoms on a prototype of the portable NIR medical monitor. The calibration model was successfully transferred to a second, distinctly different system. The Root Mean Squared Error of Prediction of met-hemoglobin in the phantom samples measured in the second system, improved from 4.94g/dl to 1.15g/dl after the standardization procedure. This compares favorably the PLS model error on the primary instrument (0.94g/dl).

In the work, the light source model for the fluorescence emission in the biochips has been extensively studied such as to effectively design the necessary micro-optic elements for the fluorescence signal detection in biochips. With most advantaging properties, the fluorescence technology does provide the high sensitivity, response in real time, and multiple target labeling for the applications in biochips. To practical applications, the final signal detection is to measure the fluorescence emission. In fact, the fluorescence emission process can be determined through four stages of transformation; that is the excitation, the absorption, the fluorescence conversion, and the fluorescence scattering. As the total internal reflection configuration for the fluorescence excitation is utilized, the evanescent waves are introduced from different excitation sources in the viewpoints of the principle analysis and the practical applications, respectively. In such a way, the spatial intensity of the fluorescence emission is found not to be uniformly distributed, and the performance of the micro-optic detection system thus diversed deviated. Except that, the fluorescence emission is further considered to include the extinction ratio and the quantum yield of the fluorescent dyes and the scattering effect from the molecules in the reaction solution. To the end, the precise fluorescence emission model in the microstructure has been obtained through the above 4 stages by the optic ray-tracing simulation. Accordingly, one corresponding collimating lens has been designed based on the new light source model.

Myoglobin is an important intracellular oxygen transport molecule in muscle. Oxygen binding to myoglobin can be determined spectroscopically due to differences in absorption of oxymyoglobin and deoxymyoglobin. Myoglobin oxygenation can be used as a measure of intracellular oxygen tension in muscle. We sought to determine the effects of differences in temperature and pH on myoglobin absorption spectra in the near-infrared spectral region. Transmission spectra were taken of pure solutions of oxymyoglobin and deoxymyoglobin at 10°, 20°, 30°, and 40°C at pH values of 6.0, 7.0, and 8.0 (n=4). In second derivative spectra at 40°C, the deoxymyoglobin peak near 760 nm was shifted by 0.9-1.2 nm toward longer wavelengths relative to 10°C at constant pH. Differences in pH did not result in statistically significant shifts in this peak at constant temperature. Estimations of myoglobin saturation from myoglobin spectra with intermediate saturations were obtained by least squares (LS) and partial least squares (PLS) analyses. Both algorithms estimate myoglobin saturation with small root mean square errors (<1e-6) when component spectra and calibration set spectra are at the same temperature as test spectra (n=100). However, when spectra at 20°C or 40°C were used as component spectra in LS with test spectra at 30°C (all at pH 7.0), errors were 0.8% and 1.4%, respectively. PLS analysis of 30°C test spectra using 20°C or 40°C calibration set spectra yielded errors of 1.6% and 1.5%, respectively. When the PLS analysis is endpoint corrected, these errors become vanishingly small. These results demonstrate that peak shifts due to temperature are potential sources of error if calibration and test spectra differ by 10°C. These errors can be minimized by appropriate spectral analytic methods.

For the detection of molecular interaction, a novel approach of the guided-mode resonance (GMR) spectroscopy identifies molecules via specific bindings with their ligands immobilized on the grating surface is presented. The structure of GMR device generally consists of two stages -- upper grating layer and waveguide layer. When the wide-band light illuminating, the GMR device inhibits on a specific resonant narrow-band of wavelength, and allows for other wavelength to transmit. The specific resonant narrow-band of wavelength results in the diffraction of the incident wide-band wave and the selection in the waveguide layer. This is very useful in highly sensitive measurement, especially for the variations in the refractive index of bulk media, and for the monitoring of variations in the thickness of thin film. In the simulation, one Si₃N₄ (n=2) GMR device is designed. When the wavelength of the illumination ranges from 1520nm to 1620nm, the resonant peak wavelength will shift 0.03nm as per 1nm bio-layer (n_bio 1.3) has been attached on. Finally, on the basis of the theoretical analysis, the optimization of a spectral GMR sensor in terms of the operation wavelength has been carried out.

Independent component analysis (ICA) method was applied as a processing step for Raman spectra. 136 Raman spectra were acquired from urine samples from 18 subjects. Each spectrum was acquired from different sample. 785nm, 100mW (at sample) laser with 2048 element linear silicon TE cooled CCD were used. In order to separate information of glucose, creatinine, urea nitrogen, uric acid and invaluable information from the urine spectrum, ICA by Maximum Likelihood (ML) fast fixed-point estimation algorithm was applied. By looking for maximum likelihood, independent information could be separated from the urine spectra. Among separated information, high frequency noise which could be generated by ambient noise and low frequency noise which contain information of baseline shift were observed. Additionally, peak information of each component was observed. The processing time was very short because fast fixed point algorithm was added to ML estimation method. Before applying ICA, all spectra were mean centered in order to enhance the peak information. In addition, all spectra were pre-processed to have unit variance in order to shorten calculation time. This first study about applying ICA suggested that this algorithm can be used as a pattern recognition algorithm to extract information from Raman spectra. Additionally, because ICA can provide information with statistical independency sufficiently, further studies about ICA which can substitute PCA will be performed.