An inverse image reconstruction approach that makes use of an algorithm based on the diffusion approximation of the radiative transport theory and a sequence of picosecond-duration slices of transmitted two-dimensional (2D) light intensity distribution for fast 3D image reconstruction is presented. The results of simulation and experiment for a cylindrical geometry are presented. Object localization in the lateral dimensions is better than that in the axial direction. The observed difference in axial and lateral resolutions are analyzed by introducing the concept of "longitudinal sensitivity".

Scalar diffraction theory is used to model the propagation of diffuse photon density waves in heavily scattering media. The principal difference between this and previous approaches is that the three classical boundary conditions have been considered. The model is shown to qualitatively demonstrate the correct trends for diffraction by planar objects embedded at different depths within the scattering medium. Quantitatively the three scalar diffraction boundary conditions are shown to provide different results in both amplitude and phase of the diffraction pattern at the detector plane. Preliminary results indicate that the Kirchoff boundary condition provides the best match to Monte Carlo simulations.

We present an analytical solution of the distribution function and the density distribution for the Boltzmann photon transport equation in an infinite uniform isotropic medium. This solution is quantitatively accurate up to the second order cumulant approximation. The solution shows a clear picture of the time evolution of particle migration, from near ballistic, through snake-like, and then into the diffusive regime.

Monte Carlo simulations (MCS) of light scattering concentrate on the particle property of the light and follow the path of a single photon penetrating the scattering medium. Although the Monte Carlo technique makes it possible to conduct simulations of media having various shapes and characteristics, it requires a great deal of computational time. In particular, because the calculation time is strongly affected by the distance between the incident and the detection positions, it is very difficult to simulate large media. To overcome this difficulty, we developed a new method that improves the calculation time of MCS by a dramatic reduction in the number of wasted photons. Conventional MCS are made such that the photons are traced in a fixed and well defined 3D space. This space has the exact size and shape of the simulated medium, and the incident and the detection positions are situated accordingly. In developed Transformed Monte Carlo Simulation (TMCS) the photon tracing is done inside a different imaginary simulation space, and a transformation is applied to each photon path in the imaginary simulation space in order to appropriate the photon path inside the model space simulating the scattering medium. Consequently, the calculation time was greatly reduced.

Currently available image reconstruction schemes for photon migration tomography are based on the diffusion approximation for light propagation in turbid media. We have shown in previous works that these schemes fail to describe the light propagation in very low scattering media such as the cerebrospinal fluid of the brain or the synovial fluid of finger joints. Therefore, we have developed a reconstruction algorithm that is based on the equation of radiative transfer. This equation describes the photon propagation in turbid media most accurately without any assumptions regarding the optical properties. Analytical solutions for complex geometries and heterogeneous media are not available. Thus, a numerical method is considered, which is based on a finite-difference formulation of the time-dependent transport equation. The reconstruction code consists of three major parts: (1) a forward model based that predicts detector readings based on the equation of radiative transfer , (2) an objective function that describes the differences between the measured and the predicted data, and (3) an updating scheme that uses the gradient of the objective function to perform a line minimization to get new guesses of the optical properties. Based on a new guess of the optical properties a new forward calculation is performed. The reconstruction process is completed when the minimum of the objective function is found.

It is well known that the reconstruction problem in optical tomography is ill-posed. Therefore, the choice of an appropriate regularization method is of crucial importance for any successful image reconstruction algorithm. In this work we approach the regularization problem within a gradient-based image iterative reconstruction (GIIR) scheme. The image reconstruction is considered as a minimization of an appropriately defined objective function. The objective function can be separated into a least-square-error term, which compares predicted and actual detector readings, and additional penalty terms that may contain additional a priori information about the system. For the efficient minimization of this objective function the gradient with respect to the spatial distribution of optical properties is calculated. Besides presenting the underlying concepts in our approach to the regularization problem, we will show numerical results that demonstrate how prior knowledge can improve the reconstruction results.

In this paper we consider new methods for localizing and characterizing the structure of an anomalous areas embedded in an overall region of interest given sparse observations of diffuse photon density wavefields. Unlike traditional techniques which use the scattered field measurements first to form an image of the full region and then post-process the resulting reconstruction to localize areas of interest, our approach finds anomalies directly from the data. To accomplish this, we model the unknowns as a superposition of a slowly varying perturbation on a background of unknown structure. We assume that the perturbation is delineated from the background by a smooth perimeter which is modeled as a spline curve of unknown knot sequence. A greedy-type approach is employed to deform the curve in a manner which optimizes a cost function enforcing both consistency with the data along with a regularization term designed to reflect prior information we have concerning the likely structure of anomalies. As the algorithm progresses, we adaptively determine the optimal weighting coefficients describing both the texture of the anomaly as well as that of the background. Examples of this approach are provided for a diffuse photon density wave problem arising in a bio-imaging application.

In this paper we examine the performance of a number of linear techniques for reconstructing the 3-D distribution of absorption coefficient within a highly scattering medium using the diffuse photon density wave (DPDW) approximation. The simulation consists of a coplanar array of sources and detectors at the boundary of an infinite slab medium. The primary difficulty in the linear reconstruction of the 3D volume from a 2D array of measurement is that the forward matrix is both underdetermined and ill-conditioned.

Propagation of light in a homogeneous scattering slab is conveniently modelled with a diffusion equation. This approach can be extended to a heterogeneous slab through a perturbation analysis. Within BornÕs approximation, the effect of an inclusion on the transmitted light is described by space-time integrals. Closed-form time integration is possible, which reduces the perturbation expressions to volume integrals over the inclusion. These can be useful to model small inclusions over which the integrand can be considered as constant. In the case of cubic inclusions with sides parallel and perpendicular to the boundaries of the surrounding slab, closed-form volume integration over the inclusion can be performed instead. Only time integrals are left, which reduces the numerical work. Numerical examples are presented. It is shown that inclusions with different volume and contrast with regards to the surrounding medium can produce the same effect on the transmitted light and are thus indistinguishable. The perturbation analysis has been used to assess the possibility of obtaining some longitudinal localization of an inclusion by using source beams and detectors of different sizes. Calculation results are also compared to experimental measurements to illustrate the validity of this analysis in the presence of small perturbations.

The scattering and the absorption coefficients of tissue can provide important information about the physiological condition of tissue. Living bodies have a complex structure, therefore in most cases tissue cannot be regarded as a homogeneous medium. In some cases biological tissue have layered structures (head, skin, bone, etc.), consequently in order to measure the optical properties in vivo it is necessary to formulate the photon migration in layered media. Making use of the diffusion equation a theoretical time domain model for photon migration through a highly scattering layered slab has been elaborated. A pencil pulsed light beam is normally incident upon the surface of the first diffusing layer. The effect of the refractive index mismatch with the external medium is taken into account and proper boundary conditions between the diffusing layers are considered. Analytical solutions of the fluence rate inside the layers have been obtained. A Monte Carlo program for the photon migration through a two or a three layered slab has been also developed. The results obtained by using the analytical solution and Monte Carlo simulations are compared and presented. The main goal of this work is the future application of the model on retrieving the optical properties of layered media.

We consider the scattering and inverse-scattering of scalar waves in the near field. By exploiting the similarity to the corresponding problems for diffusing waves, an analytic solution to the near-field inverse scattering problem is obtained. Applications to near-field microscopy are described.

In the process of developing optical tomographic imaging systems for diagnosis of disease and study of brain functions of human heads, we need realistic optical phantoms which anatomically and optically simulate human heads with complicated and multi-layered structures. Previously we have reported design and fabrication methods of optical head phantoms based on MRI images of a human head. The phantoms were simulated the multilayered structure with different optical properties at five layers; i.e., skin, skull, cerebrospinal fluid layer, gray matter and white matter. Also the phantoms which were mainly made of solid resin had dynamic parts to simulate the temporal variation of physiological functions in brain. The optical properties of the liquid circulating through the dynamic part can be changed to simulate the change in oxygenation states. The material of the skin layer has been changed from solid resin to soft rubber in order to have a good contact with optical fibers. The fabricated optical head phantom has been used in time-resolved spectroscopic measurement, and the usefulness of the dynamic optical head phantom is verified.

In recent years, models based on the "Extended Huygens-FresnelÓ principle has been applied in the description of light propagation in tissue. This principle was originally developed for light propagation through aerosols and clear air turbulence. An adaptation of this principle to tissue for the analysis of Optical Coherence Tomography (OCT) systems is advantageous, because of the convenient description of complex optical systems through implication of the ABCD-matrix formalism, and because -contrary to transport theory and diffusion theory- the phase of the light can be modeled. OCT is often combined with con focal microscopy by focusing the probing beam. It is therefore interesting to gauge the performance of the Extended-Huygens Fresnel theory for a geometry with a focused gaussian beam against a (for tissue) more well established model such as diffusion theory. A new model of a focused beam in a semi-infinite slab of scatterers using the diffusion theory is presented here. The model is developed using a GreenÕs function and a Hankel transform. The performance of both models is gauged against Monte Carlo simulations for a medium with low loss and highly forward scattering particles, which is typical of tissue. The model based on diffusion theory utilizes an exponential phase function, and the implication of using this phase function instead of the traditionally used Henyey-Greenstein phase function is investigated. Finally, reasonable qualitative agreement between the three models is shown and discussed.

Diffuse reflectance measurements of pigmented lesions has been shown to provide information useful for diagnosis. This technique can provide additional information that is not available to dermatologists by visual examination alone. It is desirable however, to gain an understanding of how this information originates from the histological and structural differences within lesions. If such an understanding could be obtained particularly through an accurate theoretical model for the interaction of light with pigmented lesions, it might be possible to improve the design of the reflectance measurement system and the resulting diagnostic performance. Semi-quantitative measurements of histological sections of pigmented lesions were made of the features thought most likely to affect the optical properties. A theoretical model using a layered Monte Carlo simulation was suggested to explain the observed reflectance spectra in terms of the histological features and to better understand how the optical properties are changed due to the presence of malignancy. We found that the microscopic structure such as cell morphology, layer thickness and melanin content are associated with differing optical properties and hence reflectance spectra.

The effect of lateral boundaries on time-resolved measurements of light transmitted through slabs of finite thickness is considered in the framework of a random walk model of photon transport in tissue. A model for a single lateral boundary is derived from the result obtained previously for an infinite slab by employing a standard technique known as the method of images. The predictions of the model are compared with time of flight data from Monte Carlo simulations (University of Florence) and experiments (University College London) using a homogenous phantom having tissue-like optical properties. Agreements in both cases are very good, indicating that the simple formalism of the random walk model is quite adequate to describe the influence of the side boundaries of the tissue slab on the observed characteristics of the transmitted light. The same methodology is applied to assess the influence of side boundaries on the time-dependent contrast functions observed in time-resolved transillumination experiments when an abnormally scattering and absorbing target is embedded in the slab. The potential use of suggested lateral boundary corrections in optical tomography is briefly discussed.

Some comparison between the results of 3D FEM and 3D Monte Carlo calculations of light propagation inside highly scattering cylindrical optical phantoms are presented. The temporal profiles of transmitted ? Dirac pulses through the cylindrical phantoms have been calculated for different source-detector angles. The comparisons have been carried out in time-domain and for both homogeneous and inhomogeneous cylindrical phantoms; in the latter case some absorbing cylinders were included inside a scattering cylinder. Refractive index mismatch between the phantoms and the surrounding medium was taken into account. Some comparison with experimental results are also shown.

Spatially resolved diffuse reflectance measurements of two-layer tissue simulating phantoms were performed in the frequency-domain. The measurements were fitted to a two-layer diffusion model of photon transport. The specific case of skin containing increasing amounts of exogenous absorber over homogeneous muscle tissue was simulated. The retrieved optical properties were compared with the known optical properties of the phantoms as well as with results obtained from Monte Carlo (M.C.) simulated data. It was found that in most cases, experimental factors that are yet to be accounted for rather than diffusion theory limitations determine the accuracy of retrieved optical properties. The absorption increase in the top layer could only be quantified for a layer thickness of 4mm. The errors in all other retrieved optical properties were in the 15-40% range with estimated uncertainties to the fitted parameter values being of similar magnitude.

The characteristics of the light transmitted through diffusing media have been studied by using the diffusion approximation of the radiative transfer equation. Making use of the partial-current-boundary condition an analytical expression for the angular dependence of the specific intensity emerging from a diffusing medium has been obtained. The analytical expression gives results in excellent agreement with Monte Carlo simulations for the slab geometry. The limits of validity of the theory have been investigated. Making use of the diffusion equation and of the extrapolated boundary condition, an heuristic analytical expression for the time-resolved reflectance has also been obtained assuming the photon flux as being simply proportional to the fluence rate. The proposed expression is simpler with respect to other commonly used expressions and seems to give better results when used to retrieve the optical properties from time-resolved measurements.

Perinatal brain injury, a major cause of neurological morbidity in infants, might be prevented or ameliorated if the preceding cerebrovascular disturbances known to cause brain ischemia and hemorrhage could be detected and monitored. We used an ISS tissue oximeter (model 96208, 110 MHz modulated light at two NIR wavelengths) to detect and monitor the development of experimental brain ischemia and hemorrhage by frequency domain optical spectroscopy (8-source/1-detector probe) and CW imaging (16-source/2-detector probe) in nineteen instrumented and ventilated newborn piglets. Our study demonstrates: 1) the sensitivity of the ISS system to detect <0.04 mL of blood at a depth of 1-2 cm; 2) the specificity to distinguish subcortical, subarachnoid, and intraventricular hemorrhage; 3) the possibility of using the optically detected arterial pulse signal to differentiate injured from functional brain regions. In the case of subcortical hemorrhage, we observed that the measured cerebral pa increased linearly with the sectional area of the hematoma rather than with the volume of injected blood. A 1-min movie clip demonstrates the ability of the ISS system to perform the very fast (0.19 s per image) imaging of the developing brain ischemia and hemorrhage.

We developed a three wavelength time-resolved spectroscopy system called the TRS-10 for use as a tissue oxygenation monitor. The TRS-10 achieved a higher data acquisition rate and a system miniaturization maintaining high sensitivity and time resolution. The TRS-10 consists of a three wavelength picosecond light pulser (PLP) with auto power control as a pulsed light source, a photomultiplier tube (PMT) having high speed and high sensitivity and miniaturized signal processing circuits for time-resolved measurement (CFD, TAC, A/D converter and histogram memory). The TRS-10 system was used to measure the time course of oxygen metabolism in a human forearm with a pneumatic cuff attached to occlude blood flow. Our system succeeded in accurately measuring concentrations of Oxy- (HbO2) and deoxyhemoglobin (Hb) in the human forearm based on diffusion theory by means of TRS data observation having an accumulation time of one second per point.

We investigated the influence of uncertainties in the scattering and absorption coefficients of the background, the system error, the depth position of the object on the reconstructed image values of diffuse optical tomography using cw illumination. Nth bom perturbation approach was used for generating forward data, and k-space diffraction tomography was used as inverse algorithm. We also examined the influences on reconstructing the perturbation in oxy- , deoxy-hemoglobin concentrations and their ratio.

Changes in regional brain blood flow and hemoglobin oxygen saturation occur in the human cortex in response to neural activation. These changes create a signal that can be imaged and quantitated using various methods, most of which do not allow for continuous bedside measurements. Imaging methods using near-infrared light, however, have been described. These allow for non-invasive measurements, and take advantage of the fact that hemoglobin is a strong absorber at these wavelengths and thus acts as a naturalÓ contrast agent. We have generated brain functional images of ill, premature infants during passive movement of the forearm using the Boston Diffusion Optical Tomography System (DOTS), a system which allows for near real-time bedside assessments. For these initial feasibility studies in the neonatal intensive care unit (NICU), custom-made soft flexible probes were made, and passive motor tasks were performed during imaging. We found that specific passive movements of the arm resulted in focal, reproducible changes in cerebral absorption at 830 nm, indicating an increase in regional blood flow and oxygenation. Further bedside studies have since been undertaken using 780 nm and 830 nm lasers. These studies indicate that the Boston DOTS is a safe and feasible bedside near-infrared functional imaging device, and underline the importance offurther studies in this critically ill patient group at high risk for brain injury.

A recently developed 24-channel optical topography (OT) system, that uses intensity-modulated near-infrared spectroscopy (NIRS) at wavelength of 780 nm and 830 nm can be used to visualize spatiotemporal changes of blood oxygenation states in human brain caused by cortical activity. We have used this system to estimate the hemodynamic changes during a language function task. To stimulate language function in the brain, we had a subject perform a writing task, which is more demanding than a speaking task. The subject, who was relaxed and sitting in a chair, was shown a card with a picture of an object for 3 seconds. During the 3 seconds, the subject wrote down the name of the object. The task was repeated 30 times, so the total stimulation period was 90 seconds. As a control task, the subject was shown figures with no meaning and drew each figure. The control task was performed as pre-stimulation for 60 seconds and as post-stimulation for 70 seconds, respectively. The subject rested for 30 seconds between post-stimulation and pre-stimulation. We observed a significant increase in the blood volume, which is proportional to the total change in the oxy-hemoglobin and deoxy-hemoglobin concentrations in the spatiotemporal topographic images in the area corresponding to Broca's area.

Optical topography (OT) was proposed in 1995 as a new imaging method for observing brain activity. It can bring out meaningful information such as spatio-temporal blood volume and oxygenation changes in cortices, and its facility can allow non-invasive measurement of human brain function under various conditions without restriction on the subjects. Researchers are beginning to use OT to investigate brain functions, and for clinical applications. In this paper, we introduce the evolution of the OT system and present a few results of brain functional measurement and clinical uses. The development of OT was advanced on 4 steps. The first generation system had only one probe with dual wavelengths. Therefore in order to reconstruct topographic images of cortical activity the probe had to be manually moved for 10 measurement positions. The second system was able to measure 12 positions, which were sequentially sampled by using a multiplexer and an optical switching device, in 6 seconds. The third system was developed to evaluate simultaneous measurement of 8 positions. At present, we finished the fourth OT system having 24 measurement channels with dual wavelengths. Using this system, we are performing dynamical observations of hemodynamic changes during brain activation and epileptic seizures.

Using photon-density-wave fluctuation-correlation-spectroscopy, we studied the fluctuations of the optical signal measured in the brain and skeletal muscles. We investigated the autocorrelation power spectra at specific tissue locations, and the coherence between different tissue regions in the frequency band 0-0.3 Hz. We found specific dominant frequency components that can be assigned to vasoconstriction activity. In a measurement protocol involving voluntary motor stimulation (right hand finger movements), we found that the optical fluctuations observed in the left forearm muscle and in the left cerebral motor cortex show different coherence features at rest and during stimulation. These changes are particularly significant at the frequencies associated with vasoconstriction activity. A power-spectrum analysis of the optical fluctuations revealed a resonance-like dependence of the optical signal at the motor cortex on the period of the motor stimulation sequence.

Diffuse optical tomography (DOT) can image spatial variations in highly scattering, tissue-like optical media. We have built an inexpensive and portable continuous-wave DOT system containing 32 laser sources (16 at 780nm and 16 at 830nm) and 16 detectors, which can acquire 288 independent measurements in less than 1 second. These data can then be processed using a variety of imaging algorithms. Preliminary studies have shown that this system can image brain bleeds in piglets, modulation of cerebral hemodynamics in rats, and brain function in both neonate and adult humans. The technical challenges involved in performing DOT over large optode areas is discussed. We describe the instrument and discuss a number of the technical issues which influenced its design. We then present a study of rat brain functional response to electrical forepaw stimulation measured with DOT, and compare it to functional MRI (fMRI). fMRI can separately measure blood volume, blood flow, and deoxy-hemoglobin concentration, and is thus a good benchmark for DOT. The relative performance of DOT and fMRI will be discussed. Our comparison shows similar temporal and spatial trends in blood volume and oxygen saturation following functional activation. These results clearly demonstrate the capabilities of DOT and set the stage for advancement to quantitative functional brain imaging.

We have developed a 24-channel simultaneous measurement system for optical topography that noninvasively obtains dynamic images of brain activity over a measurement area of 90 mm by 90 mm. To evaluate the system performance, we utilized a phantom containing a rotating absorber in a cylindrical scattering medium. In this system, eight incident and eight detecting optical fibers are arranged alternately at points of a square lattice on the phantom. The phantom is illuminated with light of two wavelengths (780 and 830 nm) from each incident fiber. Reflected light is received by the detecting fibers, each of which is connected to an avalanche photodiode. Multiple modulation and lock-in detection are used to enable highly sensitive measurement with negligible crosstalk (less than -50 dB). In the phantom measurement, we obtained topographic dynamic images of the absorber rotating in the medium with a temporal resolution of 0.5 s.

The possibility that a highly portable, fast, safe and affordable imaging system could obtain interpretable images of brain function in full- and pre-term neonates in a few seconds, has been explored in this paper. We have used a uniquely sensitive optical tomography system, termed phased array, which has revealed significant functional responses, particularly to parietal stimulation in normal and pathological states of neonate brain. This two-wavelength, frequency encoded phased array system can indicate the blood concentration and oxygenation change during the parietal brain activation in neonates. The preliminary clinical results are presented in this paper.

Ten near infrared spectroscopic reflectance images were collected from the forearm of a volunteer subject during the course of two ischemia-reperfusion events, one involving complete circulator arrest (full ischemia), the other a compromised venous out flow (venous occlusion). The data from this time series of spectroscopic images was then analyzed using an unsupervised classification scheme. Fuzzy C-means clustering was performed on a spectroscopic imaging cube, on a time series of 760 nm optical density images extracted from the ten timepoints, and on the time series of hemodynamic images derived from applying a two-wavelength, non-isobestic point oxygen saturation algorithm and a six-wavelength ÆThemoglobin, Æoxyhemoglobin and Æcytochrome-aa3 algorithm to the spectra from each time point. Periods of full ischemia and venous occlusion cause very difference changes in the absorption, scattering and hemodynamic parameters of the forearm. The application of various algorithms to the spectra of each timepoint was successful to varying degrees. Near infrared spectroscopic imaging is shown to be an excellent technique to monitor spatial variations in the tissue response to ischemic events.

Optical imaging technologies have been proven to be useful in brain function studies. Here we show the feasibility of this technology for studying muscle function. We have tested the functions of two important lower extremity muscle groups: gastrocnemius in the calf muscle and two muscles in the quadriceps muscle group during several common muscle exercises. Images obtained during the exercises demonstrated clearly the degree of muscle functions with respect to deoxygenation and blood volume changes. The optical imaging technologies can be applied to exercise physiology and training. It can be also useful for diagnoses, prognoses of certain muscle conditions in clinical and pathological cases.

A clinical trial has been performed on 95 human subjects to determine if Near Infrared (NIR) Frequency Domain Tissue Spectroscopy can detect and diagnose Peripheral Vascular Disease (PVD). Tests were performed using the ISS Oximeter on control subjects, patients at risk for PVD, patients with Intermittent Claudication, patients with Ischemic Rest Pain, and patients before and after hemodialysis. The tests consisted of measurements of total hemoglobin concentration (THC) and tissue oxygen saturation (StO2) before, during, and after walking on a treadmill. Ankle Brachial Index (ABI) measurements, a commonly used technique for PVD assessment, were also performed. Conclusions include; 1) NIR Oximetry does detect PVD, with baseline values, exercise induced values, and recovery rate measurements statistically different between the groups, 2) NIR Oximetry is more sensitive than ABI at detection of PVD in patients at risk, 3) NIR Oximetry detects dialysis induced drops in baseline oxygen saturation, as well as increases in baseline THC, 4) Reduced scattering coefficient t' has a widely variable baseline value in the patient population and does not remain constant during exercise, 5)In addition to the ability to assess PVD, the technique has other diagnostic capabilities.

The inhomogeneity of tissue structure greatly affects the sensitivity of tissue oxygenation measurement by reflectance NIRS. We have examined the influence of a subcutaneous fat layer on muscle oxygenation measurements. In this study, the influences of a fat layer and skin on muscle oxygenation measurement were investigated using Monte Carlo simulation and in vivo tests. Based on the experimental results, a correction curve for measurement sensitivity was determined. In the simulation, a 3-D model consisting of the epidermis, dermis, fat and muscle layers was used. In in vivo tests, measurement sensitivity was examined by measuring the falling rate of oxygenation in ischemia tests on the forearm using a newly developed multisensor type of oximeter with source-detector distances of 3-40 mm. Fat layer thickness was also measured by ultrasonography. The correction curve of measurement sensitivity against fat layer thickness was obtained from the results of simulation and in vivo tests. The measurements of oxygen consumption, calculated from the falling rates of oxygenation without correction, varied widely due to different thicknesses of fat layers. In contrast, the measurements of oxygen consumption with correction were almost the same (0.21 ±0.03 ml 100g-1 min-1). In this correction, the effect of skin on change in optical density was also taken into account using a detector with a short separation.

Besides morphological images created by optical coherence tomography (OCT), further parameters are evaluated in skin tissue under in vivo conditions. Scattering coefficients and refractive indices may support tissue characterization for research and diagnostic purposes in cosmetics / pharmacy and medicine, respectively. The known Klett algorithm15 has been applied for retrieval of scattering coefficients. To permit refractive index evaluation, the sample arm of our OCT setup has been arranged to minimize mechanical adjustment of a lens within the objective. A corresponding simple algorithm has been derived. Both parameters have been measured in layered structures in skin like stratum comeum, epidermis and dermis. Significant water content in a localized sweatgland duct has been observed by refractive index evaluation.

A subsurface imaging system is utilized to test the ability of the spectral polarization difference imaging technique for deep subsurface imaging in tissues. The illumination of the system is derived from compact class III lasers in the red and NIR spectral region and, alternatively, from a white light source and selection of the appropriate illumination wavelength using band-pass optical filters. The experimental results demonstrate detection and imaging of a high-scattering object located up to 1.5 cm underneath the surface of a host chicken tissue.

We developed Photoacoustic Tissue Scanning to tomografically image optically absorbing structures in tissue. With 532 nm light depths down to 6-9 mm were reached. As the samples we used capillaries with blood, human hairs or a variable dilution of Evans blue, embedded in real tissue (chicken breast) or a 10 % solution of Intralipid-10%. Small PVdF piezoelectric hydrophones were used for detection, in scanning mode for imaging purposes. The depth resolution is about 10 ?m, and the lateral resolution is limited by the diameter of the detector (200 ?m in our case).

We describe here the detailed signal processing technique that we have introduced to detect ultrasonic speckle modulation using a CCD array. We show that this new approach leads to a better signal to noise ratio than the single detector one not only because all the speckle grains signals are used but also because averaging over a few 104 grains smoothes this random signal. Finally a few examples of signal generated by locally absorbing volume immersed in real biological tissues are given.

A novel optical imaging technique was developed for noninvasive cross-sectional imaging of tissue-like turbid media. By use of a sonoluminescence signal generated internally in the media by continuous-wave ultrasound, two-dimensional images were produced for objects embedded in turbid media by raster scanning the media. Multiple objects of different shapes were resolved using this imaging technique. The images showed a high contrast and good spatial resolution. The spatial resolution was limited by the focal size of the ultrasonic focus.

First clinical prototype laser optoacoustic imaging system (LOIS) for breast cancer detection was designed and fabricated using a compact Nd:YAG laser, fiberoptic light delivery system, a linear array of 12 wide-band acoustic transducers, and a data acquisition card operated by computer with original signal processing and image reconstruction code. Initially images of small absorbing spheres were recorded in the milk with optical properties resembling those of the breast at the wavelength of 1064-nm. The system was optimized for contrast, sensitivity and axial (in-depth) resolution. The small number of acoustic transducers (12), which in turn was determined by the system cost and the time of image acquisition limited the lateral resolution of the images. Clinical ex-vivo studies on radical mastectomy specimens were performed and compared with x-ray radiography, MRI and ultrasound imaging. The results of our pilot clinical studies showed pronounced opto-acoustic contrast of ~300% between breast tumors and normal breast tissues. This contrast substantially exceeds any other endogenous tissue contrast currently utilized in clinical ultrasonography, MRI and x-ray mammography. Based on literature data and our gross observations of tumor cross-sections we hypothesize that the opto-acoustic contrast results primarily from increased optical absorption in the dense microvascularity of the tumors. In patients receiving radiotherapy, tumors were found to contain enhances concentration of dense highly scattering fiberotic tissue.

We introduce a novel imaging technique that combines the complementary features of ultrasound and near infrared diffusive light imaging. It may have potential for improved breast cancer detection and specificity. The combined technology is achieved experimentally by mounting an ultrasound array together with multiple laser diode sources and multiple optical detectors on a hand held scan probe. We demonstrate the technique in tissue phantoms wherein both acoustic and optical sensors image the volume underneath the probe. Co-registration of acoustic and optical images is achieved with an accuracy of 0.42 cm (std 0.28cm) which is the image pixel size of our prototype.

A novel chirped ultrasound-modulated optical tomography technique was developed to image turbid media. Frequency analysis was employed to obtain spatial resolution along the ultrasonic axis. 2D images from scattering medium were obtained. The chirped ultrasound modulated signal was detected in chicken breast tissue.

AcoustoÑphotonic imaging (API) , combines the technique of diffusiveÑphotonÑdensityÑ wave (DPDW) imaging with ultrasound modulation of light in a new mode of medical imaging which promises improved resolution, accuracy, and penetration. Analytical results reported previously showed that a focussed ultrasound beam at one frequency, mixing with a diffusive wave at another, produces signals at the sum and difference frequencies, and that these signals originate near the ultrasound focus. In heterogeneous tissue, reconstruction from DPDW images is difficult because the problem is underÑdetermined and illÑposed, because practical considerations constrain probe locations, and because probe placement and orientation are susceptible to motion. API provides additional data in several ways: First scanning the ultrasound focus throughout the sample volume and measuring the mixing signals images three-dimensional structures with contrasting optical or ultrasound properties. Secondly, the ultrasound focus acts as a "virtual" DPDW source which can be located deep inside the tissue without invasive probe placement . Surface detectors and deep virtual sources provide the equivalent of transillumination imaging. Diffusive imaging using transillumination and reflection together offers significant improvements in resolution and accuracy over that using either alone. Finally, sampling a few locations with the ultrasound focus, perhaps aided by ultrasound imagery, may to correct for probe motion. In summary, API combines DPDW and ultrasound imaging, retaining the best features of both.

We have developed a random walk model that uses time-dependent contrast functions to quantify the crosssection and the diffusion and absorption coefficients of an optically abnormal target from time-of-flight (TOF) data obtained in time-resolved transillumination experiments1. To substantiate our methodology we have used two different sets of data. The first set of data, provided by colleagues at University College London, are TOF measurements obtained using a solid phantom whose thickness (55mm), optical properties (absorption, ?a=0.006mm-1, transport corrected scattering, ?sc'=0.7mm-1), and characteristics of its abnormal target (size=5mm, optical properties twice those of the background) are close to those of a human breast. The second set of data, provided by colleagues at Politecnico di Milan, are TOF measurements on a 50mm thick phantom (?a=0.01mm-1, ?sc'= 1mm-1) in which two 10mm abnormal targets (one abnormally scattering, ?sc'=2mm-1; one abnormally absorbing and scattering, ?a =0.04mm-1, ?sc'=2mm-1), are embedded. None of these data includes very short path photons whose measurements are clinically impractical. Using our time-dependent contrast functions, we were able to estimate the size and optical properties of the targets with an error margin of 3-25%.

In this work, ultraviolet and visible emission and excitation fluorescence spectroscopy was performed on excised human gastrointestinal tissue for the purpose of determining whether fluorescence measurements can be used to detect and monitor patients with BarrettÕs esophagus. Tissue samples were acquired from eleven esophagectomy patients. For each excised esophagus, spectral measurements were taken from multiple locations of normal, malignant and, when present, suspected BarrettÕs tissue. At each site scanned, a biopsy sample was removed for histopathology and the results were correlated with spectral measurements. Differences were observed in the spectra from BarrettÕs, normal, and malignant esophageal tissue. Initial results indicate that these differences in spectra can clearly distinguish normal, BarrettÕs and malignant tissue, and may allow in vivo optical detection of dysplasia in patients with BarrettÕs esophagus during endoscopic examinations.

A comparison between time-resolved transmittance and reflectance techniques has been performed by considering data obtained by using different experimental set-ups with streak camera and photon counting system as detectors and Ti:Sa laser and solid state diode laser as sources. Many latex and intralipid solutions have been investigated. To analyze data analytical solutions to diffusion equation have been used and particular attention to boundary conditions has been paid. The joined use of transmittance and reflectance measurement could aid in separating the contributions of the scattering coefficient from the anisotropy g-factor.

The measurement of the propagation characteristics of Diffuse Photon-Density Waves (DPDW) is a viable path to determine the optical properties of tissue, i.e. the absorption and the reduced scattering coefficient. The technique allows to take measurements from bulk tissue, either by interstitial placement (infinite medium measurement) or superficial placement (semi-infinite medium measurement) of the source and detection fibers. Treatment of the tissue prior to measurement is unnecessary, such that changes to the structure of the tissue, and thus the optical properties, can be avoided. The optical properties can then be recovered by application of a diffusion model. When using a semi-infinite medium model, non-invasive in vivo measurements are feasible. While the theory is well established and verified by phantom measurements, the small amount of published data for biological tissue acquired by DPDW measurements suggests that there are still open practical problems, both concerning the measurement technique and the application of the diffusion model. In this study, we investigate the optimal choice of modulation frequencies for extracting a maximum amount of information on the scattering and absorption properties, examine different measurement setups, analyze how amplitude and phase data should be combined and compare the quality of infinite and semi-infinite medium measurements. The considerations are supported by experimental data, based on measurements from chicken and turkey breast muscle with modulation frequencies of up to 0.4 GHz.

The transition of linearly polarized light into randomly polarized light is discussed in terms of a diffusion process characterized by a diffusivity [radians2/mean free path] for the change in orientation of the polarized light. The linearly polarized transmission data of Jarry et al. (Applied Optics 37:7357-7367, 1998) and the circularly polarized transmission data of Schmitt et al. (Applied Optics 31:6535-6546, 1992) are analyzed. The results indicate that tissues (liver, blood) exhibit a lower diffusivity, 0.0015 [radians2/mean free path], relative to the diffusivity of polystyrene microspheres in the 0.2-1.0 ?m dia. range in water (2.0-0.025 [rad2/mfp]). The tissue diffusivity would equal the diffusivity expected for a 2.8-?m dia. polystyrene sphere in water if the linear polarization behavior for polystyrene microspheres is extrapolated linearly to larger diameters that match the behavior for tissues.

The non-destructive assessment of the quality of fruits is getting a crucial task. As a possible first step toward the noninvasive evaluation of quality, we propose the measurement of the internal optical properties of fruits and vegetables by means of time-resolved reflectance spectroscopy. Mode-locked dye and Titanium:Sapphire lasers are used as the illumination source. The reflectance signal is processed by an electronic chain for time-correlated single-photon counting. The time-dispersion of re-emitted photons is fitted to the solution of the diffusion equation, with absorption and transport scattering as free parameters. The whole instrumentation is computer-controlled and allows one to perform measurements from 610 to 1000 nm. Absorption and scattering spectra of several species of fruits were acquired, and particular attention was devoted to the optical characterization of apples.

We developed a dual-wavelength multichannel system for time-resolved reflectance measurements based on the time- correlated single-photon counting (TCSPC) technique. The output pulses of two laser diodes (672 and 818 nm, 80 MHz pulse repetition rate, 1 mW average power, 100 ps FWHM) are delayed and coupled into a multimode graded-index fiber (50/125 ?m) and injected into the tissue. The reflectance photons are collected by 1 mm fibers and detected by a multianode photomultiplier. The output signals of the photomultiplier are redirected by a router to different memory blocks of a TCSPC PC board, allowing the parallel acquisition of the curves. The solution of the time-dependent diffusion equation for the reflectance through a semi infinite homogeneous medium is best-fitted to the experimental curves, leading to the simultaneous estimate of the reduced scattering and absorption coefficients at two wavelengths in four different positions. The minimum acquisition time is 100 ms, fixed by the maximum count rate of the TCSPC PC board (1 MHz) and by the required S/N ratio. The performances of the system were tested on phantoms in terms of stability and accuracy in the determination of absolute values of the absorption and reduced scattering coefficients. Preliminary in vivo measurements were performed on healthy volunteers to estimate absorption and reduced scattering in different physiological conditions and to derive tissue functional parameters such as tissue oxygenation.

We present a multivariate fitting method to determine the absorption and reduced scattering coefficients of turbid biological tissue, from diffuse reflectance spatial distribution in the range of 0.4 to 2mm. This range of distance is about 0.5 to 5 times the reduced mean free path for most of the optical parameters we measure. Monte Carlo simulation provides numerical estimates of reflectance profiles for a limited set of optical parameter combinations. This forms a Òsampling gridÓ of the forward mapping from the two dimensional space of absorption and scattering coefficients to the reflectance profiles represented by N values. Through interpolation and least square fitting, in the inverse process the optical parameters can be determined with finer resolution than the sampling grid. We present a few strategies to search for the least square fit, including one step and iterative refinement. The precision of absorption and scattering coefficients is affected by the number of detection points, measurement errors, forward mapping grid points, and Monte Carlo simulation statistical errors. There is a trade-off between the resolution of optical coefficients and computation time or memory. Examples of tissue phantoms and in vivo skin optical property determination are presented.

Optical reflectance spectroscopy of turbid media is one of promising novel techniques for diagnostics and monitoring of biological tissues. A widely used approach in the realization of this technique is the application of photon-density waves. Until recently, the frequency-domain techniques have been limited to probing relatively large volumes of tissue. Main reason for this limitation was almost exclusive use of the diffusion approximation for the interpretation of the measured data. The diffusion model becomes invalid for source-detector separations less than several transport mean free paths (approximately 5-7 mm for most of the soft tissues in the near-infrared spectral range). We have developed an inverse technique that allows data processing even at small source-detector separations. This technique is based on a frequency-domain-optiinized Monte Carlo algorithm as a light propagation model. A specially designed acceleration scheme affords using the technique as a forward method in an iterative inverse algorithm. In this paper, a description of the technique is given, and results of simulations and preliminary phantom experiments are presented.

We presented two approaches for separating a diffusive component of the backscattered signal originated in deep tissue layers and a non-diffusive single backscattering component which backscattered from a thin epithelial layer. Both approaches can be effective and have their advantages and disadvantages. The modeling technique can provide important information about hemoglobin concentration, oxygenation, and average scattering properties of the mucosal tissue. On the other hand, when applied to new tissues, it has to be adjusted to take into account tissue morphology. Also, the polarization technique can be very robust and more effective in background removal. However, it lacks extracting capabilities of the modeling technique. Both techniques can be quite valuable and compliment each other in a future clinical device.

The intensity of two photon excited fluorescence (TPF) in highly scattering media generated by ultrashort laser pulses as a function of depth of focal point inside the media were measured for investigating the spatial location of TPF in scattering media. Scattering medium attenuates the intensity of incident beam. TPF signal decayed exponentially as the focal point was scanned into the medium. TPF was not confined in the focal region and had a wider distribution as the scattering strength of the medium increased resulting degradation of optical sectioning ability.

Second harmonic imaging was used to evaluate the subsurface layer structure of hamster cheek pouch mucosal tissues using femtosecond laser technology. Two dimensional images of hamster cheek pouch mucosa were obtained by scanning the second harmonic signal at various lateral and axial positions. The spatial mapping of the second harmonic signals showed depth differentiation between different tissue layers.

The ability to combine MRI and NIR information into a single imaging modality can potentially provide high-resolution images of absolute hemoglobin concentration and oxygen saturation from non-invasive simultaneous measurement. In order to take advantage of the MRI information, a finite element based imaging algorithm has been used to include both a priori structural information and spatial constraints during NIR image reconstruction. A case study of the optical changes in a rat cranium in response to variations in the inhaled oxygen supply has been used to test this imaging algorithm. The ability to introduce structure from MRI into NW image reconstruction resulted in significant improvements in the discrimination of hemoglobin concentration changes from oxygen saturation changes in the brain and muscle tissues.

Time-resolved spectroscopy of tissue is considered to be a powerful tool in the NIR that can readily measure absorption and scattering properties of diffuse media. This can be done employing curve-shape fitting of the experimental measurements to the diffusion equation. Several experimental factors however can significantly affect the accuracy of the method and should be taken under consideration when quantitative measurements are desired. In this work we examine the influence of laser time drift and jitter, limited excitation of waveguide modes in multimode collection fibers and the finite width of the impulse response. The percent error in absorption coefficient calculation grows with an average rate of 5% error per 20 ps of time-uncertainty while the percent error in determining the reduced scattering coefficient increases with a rate of 10% per 20 ps of time uncertainty. Deconvolution is found necessary to compensate for photon dispersion higher than 50 ps. Furthermore the quantification accuracy is shown to be dependent on the medium optical properties in a non-linear fashion due to intrinsic time-characteristics of the diffusion process.

Diffuse Photon Density Waves (DPDWs) characterise photon migration through diffusive media in the frequency-domain. Theoretically, we describe DPDWs propagation using the diffraction theory based on the diffusion approximation to the transport equation. Experimentally, we study the propagation of 100 MHz-modulated DPDWs through (optical)-breast-like phantoms. The great interest of probing these phantoms lies in the fact that they contain very small optical inhomogeneities, with a diameter of 5 mm only and with inhomogeneitj^ackground absorption and scattering contrasts of 1.1,1.5,2.0 and 4.0 respectively. We show that both most contrasted inhomogeneities (in absorption and scattering) can be distinguished, while the 1.5 contrasted one is visible only in scattering. This is consistent with the measured actual accuracy of our set-up which is 0.3% in amplitude and 0.15¡ in phase. Experimental results are compared to simulations, time-domain experimental results and X-ray measurements.

Two-dimensional transillumination images of in vitro human breast tissue specimens were recorded as a function of the wavelength of near-infrared laser light, as well as, the position of picosecond-duration temporal slices of transmitted light pulse. Spectroscopic imaging arrangement made use of 1225 - 1300 nm light from a chromium-doped forsterite laser for sample illumination, a Fourier space gate and a polarization gate to sort out a fraction of the image-bearing photons, and an InGaAs area camera for recording two-dimensional images. Time-sliced imaging approach used 800-nm, approximately 130-fs duration, 1 kHz repetition-rate pulses from a Ti:sapphire laser system to illuminate the sample, and a gated imaging system that provided a variable-position, ~80 ps-duration electronic gate to record time-sliced two-dimensional images. Excised female breast tissue specimens comprising normal and cancerous, as well as, adipose and fibrous tissues were imaged. Marked enhancement of image contrast between adipose and fibrous tissues in the specimen was observed when the wavelength of imaging light was near-resonant with an absorption band of the human fatty tissue centered on 1203 nm. In time-sliced images, earlier time slices highlighted cancerous tissues, while the later slices accentuated the normal tissues in most of the specimens.

Salient differences observed in the excitation spectra have been shown to reveal changes of proteins in the cellular and extracellular matrices of malignant tissues and degree of invasiveness

Numerical simulations of near-infrared diffusion tomography are used to reconstruct absorption and scattering images of continuously varying objects and simulated breast cancer lesions. The breast cancer lesions are modeled using chromophore levels derived from typical breast anatomy and an approximation for breast cancer lesions based on contemporary investigations of tissue/tumor optical properties. Images of hemoglobin concentration and oxygenation are derived using least squares regressions on multiple wavelength near-infrared absorption images. The simulations are performed using finite element based iterative reconstructions. The results are compared with in vivo measurements of breast cancer patients acquired using near-infrared diffusion tomography.

Tissue inhomogeneity is often neglected in the analysis of the spectral data collected in tissues. We have measured the absorption and reduced scattering spectra of the human forearm in the wavelength range 633-841 nm, and in the range of source-detector separations 0.7-3.3 cm. We found that the layered tissue structure due to the skin and adipose layers may have a significant effect not only on the data collected at different distances, but also on the data at 633 nm, where the optical absorbance is significantly higher than at the longer wavelengths considered. We conclude that the dependence of the optical penetration depth on the absorption and reduced scattering coefficients should be carefully considered in the quantitative analysis of the optical spectra of tissues.

We measure the optical properties of turbid media by performing a spatial scan for homogeneous liquid samples and by immersion into a known reference medium for small secondary samples. We demonstrate that both types of measurements have a precision of better than 1% (standard deviation) and an absolute accuracy of better than 3% (mean error). Furthermore we conduct an experiment measuring the absorption and scattering coefficients of a human hand immersed into a reference medium for a changing level of blood oxygenation.

The role of near infrared diffusive (NIR) light imaging as an adjunct to ultrasound in differentiating benign from malignant lesions was evaluated in 27 patients with cancers, apocrine metaplasia, fibroadenomas, radial scar and ductal hypoplasia, cysts and normal tissues. The NIR functional imaging parameters used were deoxygenation and blood volume. Three out of four biopsy confirmed benign lesions found suspicious by ultrasound did not show significant and/or consistent deoxygenation and blood volume changes. Three out of four biopsy confirmed malignant lesions found suspicious for malignancy by ultrasound showed significant and consistent deoxygenation and blood volume changes. One complicated biopsy confirmed benign radial scar and ductal hypoplasia showed architectural distortion on screening mammogram and an isolated area of prominent ductal structure in ultrasound image. Optical functional imaging showed no consistent deoxygenation and blood volume changes. Optical imaging did not produce any false positives in obvious normal tissues and cysts diagnosed by ultrasound. These preliminary results indicate that optical imaging has a great potential to aid ultrasound in differentiating benign from malignant lesions and therefore to reduce the unnecessary biopsies. Ultrasound is frequently used in conjunction with mammography to differentiate simple cysts from solid lesions. When all criteria for a simple cyst are strictly adhered to, the accuracy of ultrasound is 96%-100% [1]. However, the use of sonography to distinguish solid nodules has not yet been widely accepted [2]. The considerable overlap of benign and malignant lesion characteristics disappointed many radiologists that they recommended biopsies should be performed on all solid nodules [3-12]. This results in a large number of unnecessary biopsies (currently 70% to 80% of biopsies are normal [13]). Goldberg et al. have speculated the cause of lack acoustic contrast between benign and malignant lesions and suggested that it may be related to the fact that even when tissues are pathologically different, their ultrasound properties are quite similar [15]. More recently, the imaging capability of ultrasound equipment has been markedly improved. Stavros et al. have revisited the sonography issue and evaluated the current ability to distinguish benign from malignant solid breast lesions with ultrasound [2]. In [2], a total of 750 palpable solid breast lesions were studied. Despite the known overlap features in some lesions, ultrasound was able to correctly classify 123 of 125 malignant lesions as intermediate or malignant. In this paper, we introduce for the first time the use of optical imaging as an adjunct to ultrasound in differentiating benign from malignant lesions. We demonstrate that optical contrast between benign and malignant lesions can provide more diagnostic information to further classify the acoustic intermediate group, thereby improving ultrasound specificity. Recently, optical diagnostics based on diffusing near infrared (NIR) light, have been employed in breast cancer detection [16-22. Functional imaging with NIR light is made possible in a spectrum window that exists within tissues in the 700 - 900 nm NIR region, in which photon transport is dominated by scattering rather than absorption (see Figure 1). To a very good approximation, NIR photons diffuse through relatively thick tissues, such as several cms of a human breast. Functional imaging with NIR light offers several novel tissue parameters that differentiate tumors from normal breast tissue. For example, hemoglobin desaturation in tumors may be increased due to the high oxygen demand of cancers [22], and blood volume may be increased over that of normal background tissue due to the metabolic needs of cancers. In addition, light scattering properties of the tumor may be enhanced due to the increased mitochondria population with respect to the background of normal cells [23], water concentration may be different in normal and cancer cells [19], and concentration of exogenous contrast agents allows tumor identification by observing the delivery of contrast agents through the permeable vascular beds to the extravascular space surrounding the tumor [16]. In addition, NIR optical techniques have the great potential to provide information about other fundamental biological characteristics of tumors such as mitochondrial membrane potentials, intracellular calcium levels, and plasma membrane potentials. While NIR imaging provides novel optical contrast to improved tumor specificity, its relatively low resolution makes it unsuitable for morphological diagnosis. Its resolution is intrinsically limited by the diffusive nature of NIR light in tissue. Currently, optical imaging can distinguish simple structures of approximately 0.6 to 1 cm in size; sharp edges are typically blurred by a few millimeters [24]. By utilizing complementary features of ultrasound and NIR imaging, i.e., ultrasound imaging capability and optical contrast, the combined diagnostic can overcome the deficiencies of either imaging technique [25-26J. This study reports the clinical evidence that low resolution NIR imaging can be used as an adjunct tool to improve tumor specificity and to reduce the unnecessary biopsies.

Local and superficial optical property characterization of biological tissues can be performed by measuring spatially-resolved diffuse reflectance at small source-detector separations. Monte Carlo simulations and experiments were performed to assess the performance of a spatially-resolved reflectance probe, employing multiple detector fibers (0.3 to 1.4 mm from the source). Under these conditions, the inverse problem, i.e. calculating the absorption and reduced scattering coefficients, is necessarily sensitive to the phase function. This effect must be taken into account by considering a new parameter of the phase function, which depends on the first and second moments of the phase function. Probe performance is compared to another technique for quantitatively measuring optical coefficients, based on the analysis of photon density waves (Frequency Domain Photon Migration). The two techniques are found to be in reasonable agreement. However, the spatially resolved probe shows optimum measurement sensitivity in the volume immediately beneath the probe, while FDPM typically samples much larger regions of tissues. Measurements on human brain in vivo are reported using both methods.

A prototype for laser mammography based on a time-domain technique has been developed. The system uses a streak camera and a Titanium:sapphire laser which provides ultrashort pulses at a repetition rate of 80 MHz. A multi-port scanning head which includes optical fibers scans the breast in a point-by-point scanning procedure. Time-resolved transmission is measured at 15000 locations in 7 minutes. The breast is slightly compressed in both the cranio-caudal and the mediolateral projections. Amplitude calibration of the streak camera has been performed allowing for absolute measurement of time- resolved transmission. In addition to the shape of the time-resolved transmission, the absolute amplitude is relevant in properly evaluating the absorption and scattering coefficients. Promising results on solid phantoms and in vivo have been obtained. Both breasts of 10 volunteers have been scanned to date and a larger pilot study is planned in the near future. In addition to the usual time-gating processing, images of the scattering and absorption contributions are also extracted using an original data processing technique.

A technique for properly separating the scattering and absorption contributions in laser mammography is proposed. The technique is based on an empirical model obtained from a series of experiments performed on homogeneous scattering slabs containing a single inclusion. The scattering and absorption contributions are obtained by performing a Fit of an Inhomogeneous Diffusion Model (FIDM). The performance of this new technique is compared to that involving a curve fit of the solution of the diffusion model for a homogeneous slab. The FIDM technique allows a very good discrimination between scattering and absorption inclusions, better than that obtained with a curve fit of the homogeneous diffusion model. The mathematical expressions of the empirical model are extremely simple and allow for a fast calculation (about 1 second for computing two 441 pixel images compared to about 6 minutes with the previous technique). A perturbation analysis of the diffusion model will provide theoretical support to the FIDM technique and should allow its refinement. Although it has not been demonstrated that the separation between absorption and scattering is totally correct when applied to real breast scanning, the method associates some structures to absorption and some other to scattering, which could result in a better specificity of laser mammography.

In this work we reconstruct and compare NIR images of the intrinsic and extrinsic distribution of the absorption coefficient in the breast. The NIR protocol followed involves optical imaging before and after the administration of the NIR contrast agent Infracyanine 25 or Cardiogreen¨ and runs in parallel with the MRI examination protocol, while the patient is located in the magnet. For imaging purposes we employ a time-resolved imager using two wavelengths at 780 and 830 nm. The tomographic needs are served with a perturbation analysis that solves the heterogeneous diffusion equation and creates the forward model. Inversion is performed using the algebraic reconstruction technique (ART). Coregistration with Gd-DTPA enhanced MR images allows for direct and unequivocal comparison between optical and MR images. We focus on the comparison between the use of intrinsic or extrinsic contrast to differentiate breast tissue types. The distribution of the NIR contrast agent yields very good coregistration with the MR contrast enhanced images and that it allows accurate quantification. Our results greatly exemplify the need for NIR contrast agents in optical mammography. Conversely imaging of the intrinsic coefficient is associated with more experimental uncertainties leading to less quantification accuracy.

For X-ray mammography, we are unable to apply to younger subjects of woman under 40 years old. The under 40 years old BRCA-1 population has a serious diagnostic problem. So we have chosen to develop an optical technique for breast cancer detection. In order to provide a safe, fast and portable dual wavelength breast imager, We have used high sensitive cancellation techniques (amplitude cancellation image system) as an imaging system that can better to localize the metabolic and blood volume responses. This device consists of 8 sources (750 nm, 830 nm) and 21 silicon photodiodes. The sources are modulated by 50 us pulse and pulse encode the dual wavelength sources (750nm, 830 nm). Using programmed amplifies, the image system is first calibrated on a breast model by adjusting all outputs to be equal, hence their differential amplifier outputs are zero. The high sensitive null (zero) plane is applied to detect small objects. By averaging of the imaged data over 8 sec. The system indicated a very high signal to noise ratio (100) in detecting inhomogeneity of the breast with respect to the surrounding tissue or to the contralateral breast in voxeis containing blood volume and oxygenation changes corresponding to the presence of tumors or related tissue pathologies.

A new clinical tissue oxygenation monitor, the NIRO-300, has been developed. In addition to monitoring changes in hemoglobin concentration and redox state of cytochrome oxidase, the instrument can measure a Tissue Oxygenation Index (TOI), which is the ratio of oxygenated to total tissue hemoglobin, by utilizing NIR spatially resolved spectroscopy (SRS). In SRS, the slope of light attenuation versus distance is measured at a distant point from the light input, from which the TOI is calculated using photon diffusion theory. Four laser diodes are used as the light source and Class 1 laser light (IEC 825) is irradiated onto the skin. A high gain and low noise amplifier is used in the detector, which enables a large emitter-detector separations of around 5cm. To evaluate the TOI measured by the NIRO-300, it was compared to the data from a blood gas analyzer (S02 value) in measurements made on phantoms containing intralipid and blood. We also made simultaneous measurements by the NIRO-300 and an NIRS machine based on time resolved spectroscopy (TRS) on human arms. In the both measurements, TOI showed an excellent correlation with the data from blood gas analyzer and the TRS machine, which suggests the efficacy of the TOI data also in clinical use.

We present a preliminary assessment of the feasibility of using a frequency domain spectrometer for measuring fetal cerebral blood oxygenation in-utero at 750 and 780 nm of modulation at 70 MHz. The amplitude and phase of the diffuse photon density wave which propagates through the tissue is detected by a photomultiplier tube. The absorption coefficient ?a and the reduced scattering coefficient ?'s are calculated based on the analytic solution of the photon diffusion equation for a semi-infinite geometry. The accuracy of the instrument and the algorithm in measuring ?a is tested by an experiment performed on a homogeneous ink/intralipid tissue phantom. Also the sensitivity of the system to the change of blood volume and oxgenation is evaluated through a homogeneous blood/intralipid tissue phantom experiment. Clinical data was recorded during the elective Cesarean section of human subjects. The measurements were made on maternal abdomen in the post-epidural stage and on the fontanel region of neonate before (pre-cord) and after (post-cord) the umbilical cord was detached. The source detector separation used were 7 and 3(or 4) cm respectively. The deoxygenated and oxygenated hemoglobin concentration, blood volume and saturation were calculated based on ?a of 2 wavelength. The averaged optical properties and blood parameters for each stage are presented. The post-epidural blood saturation value correlates well with the pre-cord saturation value. A larger number of human subjects are being evaluated to provide statistically significant results.

Pulse oximetry (oxygen saturation monitoring) has markedly improved medical care in many fields, including anesthesiology, intensive care, and newborn intensive care. In obstetrics, fetal heart rate monitoring remains the standard for intrapartum assessment of fetal well being. Fetal oxygen saturation monitoring is a new technique currently under development. It is potentially superior to electronic fetal heart rate monitoring (cardiotocography) because it allows direct assessment of both fetal oxygen status and fetal tissue perfusion. Here, we present feasibility studies for trans-abdominal fetal cerebral pulse oximetry. Our experiments on more than 20 patients indicate feasibility. We will present the methodology for obtaining these data, as well as a summary of our pilot clinical study.

Most of the instruments reported for near infrared absorption measurements in highly scattering media, suitable for noninvasive hemoglobin concentration measurements in living tissues, require some type of instrumental calibration, or knowledge of the initial concentrations of the two species. A new type of probe, which does not require calibration, has been developed. These new probes can provide a rapid and ÒabsoluteÓ measurement of absorption and scattering in tissue when combined with the Òmulti-distanceÓ method developed, in part, at the Laboratory for Fluorescence Dynamics at the University of Illinois and ISS Incorporated. An initial evaluation of the performance and efficacy of these new probes has been performed using a frequency-domain ISS tissue oximeter. The new probes offer the convenience of not requiring calibration, and also provide more reliable measurements and greater long-term stability compared with the standard multidistance design. The new probes rely on a symmetrical arrangement of light sources and detectors. These geometries allow for measurements that are, in theory, entirely independent of the intensity differences between light sources, phase differences between the multiple light sources, sensitivity differences between the detectors, and any optical coupling losses, including those at the probe sample interface.

There are few published works (Marbach 1993,1995) in the field of tissue spectroscopy for glucose self-monitoring, in the wavelength range from 1.0 microns to 2.5 microns, and the error described is too large for clinical use or for home monitoring. Measurements (Arnold 1990, Marbach 1993) of aqueous solutions of glucose, with concentrations in the physiological range, have been promising, including measurements (McNulty 1998) in our laboratory.

Studies have shown that hypoxic tumor cells are relatively more resistant to radiotherapy, chemotherapy, and photodynamic therapy. Tumor oximetry, e.g., measurement of oxygen tension (PO2) of tissue and/or blood oxygenation (S02) of the vascular bed, could be valuable for optimizing treatment plans. In this study, we employed a recently developed homodyne system to measure changes in hemoglobin saturation (S02) and concentration in the vascular bed of rat prostate and breast tumors. For comparison, tissue p02 values were measured using 19F MR EPI of hexafluorobenzene, providing a map of regional tumor oxygenation tension. Both S02 and p02 measurements were taken while the inhaled gas was alternated between 33% oxygen, 100% oxygen and carbogen (95% oxygen, 5% C02). The results obtained for both techniques showed significant changes in tumor oxygenation accompanying respiratory challenge, with changes in vascular S02 preceding tissue p02 change. The combined use of these two techniques provides new insight into the dynamics of tumor oxygenation by making available a method of obtaining regional information of the state of the tissue, as well as a non-invasive, real-time method for determining changes in the vascular bed.

We present tissue phantom experimental results and theoretical simulations to study photon migration through the fetal head in-utero. A continuous-wave (CW), dual wavelength (760 & 850 nm) spectrometer was developed and employed for the experiments at a source-detector separation of 10 cm. Theoretical simulations were performed using time-independent, finite-difference, discrete-ordinate, radiative-transport and diffusion equations. Two phantom geometries viz. circular and rectangular were considered. The tissue phantom incorporates a fetal head (absorption coefficient, .?a: 0.15 cm-1 & reduced scattering coefficient, ?s': 5.0 cm-1), an amniotic fluid sac (?a=0.02 cm-1, ?s= 0. 1 cm-1) and a maternal tissue layer (?a= 0.08 cm-1, ?s'= 5.0 cm-1). Photon fluence from the tissue phantom was quantified as a function of fetal head depth and its position relative to probe placement. Experimental results obtained with spectrometer were found to be congruent with theoretical results and clinical investigations. The results indicate that photon fluence decreases with increase in fetal head depth for circular geometry, while it increases with increase in fetal head depth for rectangular geometry. This paradoxical result observed may be attributed to the effect of amniotic fluid in the light path. Photon fluence is sensitive for fetal head depths within 40 mm. This is well within the fetal head depths expected in near-term patients (~20 mm)

We present a novel technique based on tilting the bed where the subject is lying, to non-invasively measure the tissue blood flow (BF) and oxygen consumption (OC) with near-infrared (NIR) spectroscopy. We used a NIR, frequency domain spectrometer to measure the concentrations of oxy-hemoglobin ([HbO2]), deoxy-hemoglobin ([Hb]) and total hemoglobin (THC) in the calf muscle of human subjects. The subject was lying horizontally, and after a baseline acquisition, the bed was tilted by 10 degrees (feet down, head up). This position was kept for 1 min, then the subject was brought back to the horizontal position. This tilting procedure caused variations in the calf [Hb02], [Hb], and THC similar to those observed during a pneumatic-cuff-induced venous occlusion. The increasing rate of THC and [Hb] caused by tilting allowed the calculation of blood flow and oxygen consumption. We found a quantitative agreement between the values of BF (OC) measured with the tilting table and with the venous occlusion protocols. On the 26 subjects examined with the tilting table protocol, we found population average values of BF = 1.51 ml (100ml)-1-min-1 and OC = 6.10 ?mol( 100ml)-1 min-1.

We present non-invasive measurements of the calf muscle blood flow (BF) and oxygen consumption (OC) by near-infrared spectroscopy. We used a frequency domain tissue oximeter (modulation frequency: 110 MHz, wavelengths: 758 and 830 nm) to measure in real time (acquisition time: 0.64 s) the hemoglobin concentration and saturation. After 1-min of baseline acquisition, we achieved venous occlusion by inflating a pneumatic cuff on the subjectÕs thigh to a pressure of 60 mmHg. The cuff was released after 1 min. The baseline/inflation/release procedure was repeated 3 times to verify reproducibility. We calculated the BF and OC from the initial rate of increase of the total hemoglobin and deoxy-hemoglobin concentration immediately after the onset of venous occlusion. We examined 8 healthy subjects and 18 patients affected by peripheral vascular disease (PVD) in 1 or 2 legs to investigate whether muscle BF and OC at rest can be useful indicators of vascular insufficiency. In healthy legs, we obtained average values of BF=0.73 ml/(100ml)/min and OC=0.10 ml/(100g)/min. The corresponding average values found in legs affected by PVD are BF=1.39 ml/(100ml)/min and OC=0.16 ml/(100g)/min. The ranges of values of BF and OC measured in the healthy legs broadly overlap with the corresponding ranges measured in the PVD legs.

An otoacoustic transducer capable of detecting wide-band ultrasonic profiles at the site of pulsed laser irradiation was developed and applied in imaging various stages of squamous cell carcinoma produced in a hamster cheek pouch model. Imaging was performed using 12-nanosecond Nd:YAG laser pulses (FWHM) at the wavelength of 532 nm, which provided sufficient depth of monitoring (2 mm), significant tissue contrast and 15-?m in depth resolution. The opto-acoustic images depict distinct tissue layers in control (normal) animals and heterogeneous structureless tissue in pre-cancer and cancer stages. Pronounced changes in subsurface tissue structures that occur in the course of cancer development were not detected upon gross examination. Optoacoustic images of later stages in cancer development presented exophytic or irregular profiles on the surfaces, possibly associated with enhanced microvascularization. The changes associated with benign inflammation yielded optoacoustic images different from those depicting cancer progression. The experimental results suggest that the optoacoustic images displaying layered tissue structure on the basis of variations in optical absorption and scattering may be utilized for early detection and staging of superficial mucosa cancer.

We present a tomography method for fluorescence and absorption biomedical optical imaging which minimizes the computational burden of three-dimensional image reconstruction and enables data conditioning on the basis of variable and possibly spatially-correlated measurement and system noise. Specifically, we present three-dimensional images reconstructed from (i) synthetic frequency-domain measurements; (ii) finite difference solution to the diffusion equation employing partial current boundary conditions; (iii) a recursive, minimum variance, optimization algorithm employing a Bayesian approximate extended Kalman filter accounting for measurement and system noise; and (iv) a unique, data-driven zonation scheme to dynamically determine parameterization and accelerate convergence. Using a synthetic data set with 0.1¡ standard deviation Gaussian noise added to phase, we demonstrate the ability to image multiple distinct 0.5 cm diameter absorbing/fluorescing heterogeneities within a combined transillumination/reflectance geometry comprising 8 sources and 90 detectors. Reconstruction of absorption maps owing to spatial distribution of fluorophores that were discretized onto a 9x9x9 node grid required just over 4 minutes on a 350 MHz Pentium II computer.

The quantification of tissue hemoglobin oxygen saturation has great clinical significance either to predict per-term baby brain damage or other dysfunction due to abnormal hemoglobin saturation level. From our experiments we find out that the background absorption can affect the accuracy in calculating saturation of hemoglobin either the range or the value. Practically it is difficult to get the absolute value of background absorption so we try to use empirical approach to obtain this value. Our goal is to investigate the transferability of this empirical value from a small ground of population to general population and from piglets to man.

There are no reliable values for the transport scattering coefficient (?'s) of whole blood, particularly at the high hematocrits found in-vivo. Existing values are largely based on single scattering measurements of scattering crosssection and angular scattering. The accuracy of the single scattering method is limited by the accuracy with which the g-value can be measured, as it is near 0.99. Diffuse measurements, coupled to diffusion theory models, are normally hampered by the relatively high absorption coefficient of oxy-hemoglobin even in the near infrared. Blood was drawn from 9 healthy volunteers, heparinised, and the cells washed twice with phosphate buffered saline. Oxyhemoglobin was converted to carbonmonoxide-hemoglobin, which has an absorption coefficient that is approximately 5% of that of oxy-hemoglobin in the near infrared region. Picosecond pulses of light were transmitted through an 80x60x20 mm cuvette and the resulting temporal point spread functions recorded with a streak camera. The data were fitted to a time-dependent diffusion model to derive values for ?'s and ?'a which is now valid because of the low ?a of carbonmonoxide-hemoglobin. The effect of changing the hematocrit was determined by stepwise dilution of the washed red cells with further phosphate buffered saline. Mean values for ?'s and ?a at 802 nm were 3.15±0.12 mm-1 and 0.0098±0.0004 mm-1 respectively for the hematocrit in the normal range. The undiluted red cells showed a small negative wavelength dependence for ?'s over the range 740 to 860 nm with a slope of -0.0045±0.0011 mm-1nm-1 which was highly significant (p=0.0001). The negative wavelength dependence of ?'s is compatible with Mie theory predictions of scattering by large particles. A near linear relationship between ?'s versus hematocrit (H) was found in this study (up to H=0.45) as opposed to the ?'s ° H( 1 Ñ H) or ?'s ° H(1 Ñ H)(1.4 Ñ H) macroscopic scattering coefficient relationship of others..

This paper discusses the experimental validity of a frequently assumed fact: in a highly anisotropic scattering medium considered as infinite, a collimated intensity-modulated beam source can be modelled by a point-like source located on the emissive axis, at one time the mean transport-length from the source fibre tip.